linux/drivers/nvme/host/pci.c

2791 lines
70 KiB
C
Raw Normal View History

/*
* NVM Express device driver
* Copyright (c) 2011-2014, Intel Corporation.
*
* This program is free software; you can redistribute it and/or modify it
* under the terms and conditions of the GNU General Public License,
* version 2, as published by the Free Software Foundation.
*
* This program is distributed in the hope it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
* more details.
*/
#include <linux/aer.h>
#include <linux/async.h>
#include <linux/blkdev.h>
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
#include <linux/blk-mq.h>
#include <linux/blk-mq-pci.h>
#include <linux/dmi.h>
#include <linux/init.h>
#include <linux/interrupt.h>
#include <linux/io.h>
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/mutex.h>
#include <linux/once.h>
#include <linux/pci.h>
NVMe: Metadata format support Adds support for NVMe metadata formats and exposes block devices for all namespaces regardless of their format. Namespace formats that are unusable will have disk capacity set to 0, but a handle to the block device is created to simplify device management. A namespace is not usable when the format requires host interleave block and metadata in single buffer, has no provisioned storage, or has better data but failed to register with blk integrity. The namespace has to be scanned in two phases to support separate metadata formats. The first establishes the sector size and capacity prior to invoking add_disk. If metadata is required, the capacity will be temporarilly set to 0 until it can be revalidated and registered with the integrity extenstions after add_disk completes. The driver relies on the integrity extensions to provide the metadata buffer. NVMe requires this be a single physically contiguous region, so only one integrity segment is allowed per command. If the metadata is used for T10 PI, the driver provides mappings to save and restore the reftag physical block translation. The driver provides no-op functions for generate and verify if metadata is not used for protection information. This way the setup is always provided by the block layer. If a request does not supply a required metadata buffer, the command is failed with bad address. This could only happen if a user manually disables verify/generate on such a disk. The only exception to where this is okay is if the controller is capable of stripping/generating the metadata, which is possible on some types of formats. The metadata scatter gather list now occupies the spot in the nvme_iod that used to be used to link retryable IOD's, but we don't do that anymore, so the field was unused. Signed-off-by: Keith Busch <keith.busch@intel.com>
2015-02-19 23:39:03 +03:00
#include <linux/t10-pi.h>
#include <linux/types.h>
#include <linux/io-64-nonatomic-lo-hi.h>
#include <linux/sed-opal.h>
asm-generic: architecture independent readq/writeq for 32bit environment This provides unified readq()/writeq() helper functions for 32-bit drivers. For some cases, readq/writeq without atomicity is harmful, and order of io access has to be specified explicitly. So in this patch, new two header files which contain non-atomic readq/writeq are added. - <asm-generic/io-64-nonatomic-lo-hi.h> provides non-atomic readq/ writeq with the order of lower address -> higher address - <asm-generic/io-64-nonatomic-hi-lo.h> provides non-atomic readq/ writeq with reversed order This allows us to remove some readq()s that were added drivers when the default non-atomic ones were removed in commit dbee8a0affd5 ("x86: remove 32-bit versions of readq()/writeq()") The drivers which need readq/writeq but can do with the non-atomic ones must add the line: #include <asm-generic/io-64-nonatomic-lo-hi.h> /* or hi-lo.h */ But this will be nop in 64-bit environments, and no other #ifdefs are required. So I believe that this patch can solve the problem of 1. driver-specific readq/writeq 2. atomicity and order of io access This patch is tested with building allyesconfig and allmodconfig as ARCH=x86 and ARCH=i386 on top of tip/master. Cc: Kashyap Desai <Kashyap.Desai@lsi.com> Cc: Len Brown <lenb@kernel.org> Cc: Ravi Anand <ravi.anand@qlogic.com> Cc: Vikas Chaudhary <vikas.chaudhary@qlogic.com> Cc: Matthew Garrett <mjg@redhat.com> Cc: Jason Uhlenkott <juhlenko@akamai.com> Cc: James Bottomley <James.Bottomley@parallels.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Roland Dreier <roland@purestorage.com> Cc: James Bottomley <jbottomley@parallels.com> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Matthew Wilcox <matthew.r.wilcox@intel.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Hitoshi Mitake <h.mitake@gmail.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-02-07 06:45:33 +04:00
#include "nvme.h"
#define SQ_SIZE(depth) (depth * sizeof(struct nvme_command))
#define CQ_SIZE(depth) (depth * sizeof(struct nvme_completion))
#define SGES_PER_PAGE (PAGE_SIZE / sizeof(struct nvme_sgl_desc))
static int use_threaded_interrupts;
module_param(use_threaded_interrupts, int, 0);
static bool use_cmb_sqes = true;
module_param(use_cmb_sqes, bool, 0644);
MODULE_PARM_DESC(use_cmb_sqes, "use controller's memory buffer for I/O SQes");
static unsigned int max_host_mem_size_mb = 128;
module_param(max_host_mem_size_mb, uint, 0444);
MODULE_PARM_DESC(max_host_mem_size_mb,
"Maximum Host Memory Buffer (HMB) size per controller (in MiB)");
static unsigned int sgl_threshold = SZ_32K;
module_param(sgl_threshold, uint, 0644);
MODULE_PARM_DESC(sgl_threshold,
"Use SGLs when average request segment size is larger or equal to "
"this size. Use 0 to disable SGLs.");
static int io_queue_depth_set(const char *val, const struct kernel_param *kp);
static const struct kernel_param_ops io_queue_depth_ops = {
.set = io_queue_depth_set,
.get = param_get_int,
};
static int io_queue_depth = 1024;
module_param_cb(io_queue_depth, &io_queue_depth_ops, &io_queue_depth, 0644);
MODULE_PARM_DESC(io_queue_depth, "set io queue depth, should >= 2");
struct nvme_dev;
struct nvme_queue;
static void nvme_dev_disable(struct nvme_dev *dev, bool shutdown);
/*
* Represents an NVM Express device. Each nvme_dev is a PCI function.
*/
struct nvme_dev {
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
struct nvme_queue *queues;
struct blk_mq_tag_set tagset;
struct blk_mq_tag_set admin_tagset;
u32 __iomem *dbs;
struct device *dev;
struct dma_pool *prp_page_pool;
struct dma_pool *prp_small_pool;
unsigned online_queues;
unsigned max_qid;
unsigned int num_vecs;
int q_depth;
u32 db_stride;
void __iomem *bar;
unsigned long bar_mapped_size;
struct work_struct remove_work;
struct mutex shutdown_lock;
bool subsystem;
void __iomem *cmb;
pci_bus_addr_t cmb_bus_addr;
u64 cmb_size;
u32 cmbsz;
u32 cmbloc;
struct nvme_ctrl ctrl;
struct completion ioq_wait;
/* shadow doorbell buffer support: */
nvme: improve performance for virtual NVMe devices This change provides a mechanism to reduce the number of MMIO doorbell writes for the NVMe driver. When running in a virtualized environment like QEMU, the cost of an MMIO is quite hefy here. The main idea for the patch is provide the device two memory location locations: 1) to store the doorbell values so they can be lookup without the doorbell MMIO write 2) to store an event index. I believe the doorbell value is obvious, the event index not so much. Similar to the virtio specification, the virtual device can tell the driver (guest OS) not to write MMIO unless you are writing past this value. FYI: doorbell values are written by the nvme driver (guest OS) and the event index is written by the virtual device (host OS). The patch implements a new admin command that will communicate where these two memory locations reside. If the command fails, the nvme driver will work as before without any optimizations. Contributions: Eric Northup <digitaleric@google.com> Frank Swiderski <fes@google.com> Ted Tso <tytso@mit.edu> Keith Busch <keith.busch@intel.com> Just to give an idea on the performance boost with the vendor extension: Running fio [1], a stock NVMe driver I get about 200K read IOPs with my vendor patch I get about 1000K read IOPs. This was running with a null device i.e. the backing device simply returned success on every read IO request. [1] Running on a 4 core machine: fio --time_based --name=benchmark --runtime=30 --filename=/dev/nvme0n1 --nrfiles=1 --ioengine=libaio --iodepth=32 --direct=1 --invalidate=1 --verify=0 --verify_fatal=0 --numjobs=4 --rw=randread --blocksize=4k --randrepeat=false Signed-off-by: Rob Nelson <rlnelson@google.com> [mlin: port for upstream] Signed-off-by: Ming Lin <mlin@kernel.org> [koike: updated for upstream] Signed-off-by: Helen Koike <helen.koike@collabora.co.uk> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <keith.busch@intel.com>
2017-04-10 18:51:07 +03:00
u32 *dbbuf_dbs;
dma_addr_t dbbuf_dbs_dma_addr;
u32 *dbbuf_eis;
dma_addr_t dbbuf_eis_dma_addr;
/* host memory buffer support: */
u64 host_mem_size;
u32 nr_host_mem_descs;
dma_addr_t host_mem_descs_dma;
struct nvme_host_mem_buf_desc *host_mem_descs;
void **host_mem_desc_bufs;
};
static int io_queue_depth_set(const char *val, const struct kernel_param *kp)
{
int n = 0, ret;
ret = kstrtoint(val, 10, &n);
if (ret != 0 || n < 2)
return -EINVAL;
return param_set_int(val, kp);
}
nvme: improve performance for virtual NVMe devices This change provides a mechanism to reduce the number of MMIO doorbell writes for the NVMe driver. When running in a virtualized environment like QEMU, the cost of an MMIO is quite hefy here. The main idea for the patch is provide the device two memory location locations: 1) to store the doorbell values so they can be lookup without the doorbell MMIO write 2) to store an event index. I believe the doorbell value is obvious, the event index not so much. Similar to the virtio specification, the virtual device can tell the driver (guest OS) not to write MMIO unless you are writing past this value. FYI: doorbell values are written by the nvme driver (guest OS) and the event index is written by the virtual device (host OS). The patch implements a new admin command that will communicate where these two memory locations reside. If the command fails, the nvme driver will work as before without any optimizations. Contributions: Eric Northup <digitaleric@google.com> Frank Swiderski <fes@google.com> Ted Tso <tytso@mit.edu> Keith Busch <keith.busch@intel.com> Just to give an idea on the performance boost with the vendor extension: Running fio [1], a stock NVMe driver I get about 200K read IOPs with my vendor patch I get about 1000K read IOPs. This was running with a null device i.e. the backing device simply returned success on every read IO request. [1] Running on a 4 core machine: fio --time_based --name=benchmark --runtime=30 --filename=/dev/nvme0n1 --nrfiles=1 --ioengine=libaio --iodepth=32 --direct=1 --invalidate=1 --verify=0 --verify_fatal=0 --numjobs=4 --rw=randread --blocksize=4k --randrepeat=false Signed-off-by: Rob Nelson <rlnelson@google.com> [mlin: port for upstream] Signed-off-by: Ming Lin <mlin@kernel.org> [koike: updated for upstream] Signed-off-by: Helen Koike <helen.koike@collabora.co.uk> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <keith.busch@intel.com>
2017-04-10 18:51:07 +03:00
static inline unsigned int sq_idx(unsigned int qid, u32 stride)
{
return qid * 2 * stride;
}
static inline unsigned int cq_idx(unsigned int qid, u32 stride)
{
return (qid * 2 + 1) * stride;
}
static inline struct nvme_dev *to_nvme_dev(struct nvme_ctrl *ctrl)
{
return container_of(ctrl, struct nvme_dev, ctrl);
}
/*
* An NVM Express queue. Each device has at least two (one for admin
* commands and one for I/O commands).
*/
struct nvme_queue {
struct device *q_dmadev;
struct nvme_dev *dev;
spinlock_t sq_lock;
struct nvme_command *sq_cmds;
struct nvme_command __iomem *sq_cmds_io;
spinlock_t cq_lock ____cacheline_aligned_in_smp;
volatile struct nvme_completion *cqes;
struct blk_mq_tags **tags;
dma_addr_t sq_dma_addr;
dma_addr_t cq_dma_addr;
u32 __iomem *q_db;
u16 q_depth;
s16 cq_vector;
u16 sq_tail;
u16 cq_head;
u16 qid;
u8 cq_phase;
nvme: improve performance for virtual NVMe devices This change provides a mechanism to reduce the number of MMIO doorbell writes for the NVMe driver. When running in a virtualized environment like QEMU, the cost of an MMIO is quite hefy here. The main idea for the patch is provide the device two memory location locations: 1) to store the doorbell values so they can be lookup without the doorbell MMIO write 2) to store an event index. I believe the doorbell value is obvious, the event index not so much. Similar to the virtio specification, the virtual device can tell the driver (guest OS) not to write MMIO unless you are writing past this value. FYI: doorbell values are written by the nvme driver (guest OS) and the event index is written by the virtual device (host OS). The patch implements a new admin command that will communicate where these two memory locations reside. If the command fails, the nvme driver will work as before without any optimizations. Contributions: Eric Northup <digitaleric@google.com> Frank Swiderski <fes@google.com> Ted Tso <tytso@mit.edu> Keith Busch <keith.busch@intel.com> Just to give an idea on the performance boost with the vendor extension: Running fio [1], a stock NVMe driver I get about 200K read IOPs with my vendor patch I get about 1000K read IOPs. This was running with a null device i.e. the backing device simply returned success on every read IO request. [1] Running on a 4 core machine: fio --time_based --name=benchmark --runtime=30 --filename=/dev/nvme0n1 --nrfiles=1 --ioengine=libaio --iodepth=32 --direct=1 --invalidate=1 --verify=0 --verify_fatal=0 --numjobs=4 --rw=randread --blocksize=4k --randrepeat=false Signed-off-by: Rob Nelson <rlnelson@google.com> [mlin: port for upstream] Signed-off-by: Ming Lin <mlin@kernel.org> [koike: updated for upstream] Signed-off-by: Helen Koike <helen.koike@collabora.co.uk> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <keith.busch@intel.com>
2017-04-10 18:51:07 +03:00
u32 *dbbuf_sq_db;
u32 *dbbuf_cq_db;
u32 *dbbuf_sq_ei;
u32 *dbbuf_cq_ei;
};
/*
* The nvme_iod describes the data in an I/O, including the list of PRP
* entries. You can't see it in this data structure because C doesn't let
* me express that. Use nvme_init_iod to ensure there's enough space
* allocated to store the PRP list.
*/
struct nvme_iod {
struct nvme_request req;
struct nvme_queue *nvmeq;
bool use_sgl;
int aborted;
int npages; /* In the PRP list. 0 means small pool in use */
int nents; /* Used in scatterlist */
int length; /* Of data, in bytes */
dma_addr_t first_dma;
struct scatterlist meta_sg; /* metadata requires single contiguous buffer */
struct scatterlist *sg;
struct scatterlist inline_sg[0];
};
/*
* Check we didin't inadvertently grow the command struct
*/
static inline void _nvme_check_size(void)
{
BUILD_BUG_ON(sizeof(struct nvme_rw_command) != 64);
BUILD_BUG_ON(sizeof(struct nvme_create_cq) != 64);
BUILD_BUG_ON(sizeof(struct nvme_create_sq) != 64);
BUILD_BUG_ON(sizeof(struct nvme_delete_queue) != 64);
BUILD_BUG_ON(sizeof(struct nvme_features) != 64);
BUILD_BUG_ON(sizeof(struct nvme_format_cmd) != 64);
BUILD_BUG_ON(sizeof(struct nvme_abort_cmd) != 64);
BUILD_BUG_ON(sizeof(struct nvme_command) != 64);
BUILD_BUG_ON(sizeof(struct nvme_id_ctrl) != NVME_IDENTIFY_DATA_SIZE);
BUILD_BUG_ON(sizeof(struct nvme_id_ns) != NVME_IDENTIFY_DATA_SIZE);
BUILD_BUG_ON(sizeof(struct nvme_lba_range_type) != 64);
BUILD_BUG_ON(sizeof(struct nvme_smart_log) != 512);
nvme: improve performance for virtual NVMe devices This change provides a mechanism to reduce the number of MMIO doorbell writes for the NVMe driver. When running in a virtualized environment like QEMU, the cost of an MMIO is quite hefy here. The main idea for the patch is provide the device two memory location locations: 1) to store the doorbell values so they can be lookup without the doorbell MMIO write 2) to store an event index. I believe the doorbell value is obvious, the event index not so much. Similar to the virtio specification, the virtual device can tell the driver (guest OS) not to write MMIO unless you are writing past this value. FYI: doorbell values are written by the nvme driver (guest OS) and the event index is written by the virtual device (host OS). The patch implements a new admin command that will communicate where these two memory locations reside. If the command fails, the nvme driver will work as before without any optimizations. Contributions: Eric Northup <digitaleric@google.com> Frank Swiderski <fes@google.com> Ted Tso <tytso@mit.edu> Keith Busch <keith.busch@intel.com> Just to give an idea on the performance boost with the vendor extension: Running fio [1], a stock NVMe driver I get about 200K read IOPs with my vendor patch I get about 1000K read IOPs. This was running with a null device i.e. the backing device simply returned success on every read IO request. [1] Running on a 4 core machine: fio --time_based --name=benchmark --runtime=30 --filename=/dev/nvme0n1 --nrfiles=1 --ioengine=libaio --iodepth=32 --direct=1 --invalidate=1 --verify=0 --verify_fatal=0 --numjobs=4 --rw=randread --blocksize=4k --randrepeat=false Signed-off-by: Rob Nelson <rlnelson@google.com> [mlin: port for upstream] Signed-off-by: Ming Lin <mlin@kernel.org> [koike: updated for upstream] Signed-off-by: Helen Koike <helen.koike@collabora.co.uk> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <keith.busch@intel.com>
2017-04-10 18:51:07 +03:00
BUILD_BUG_ON(sizeof(struct nvme_dbbuf) != 64);
}
static inline unsigned int nvme_dbbuf_size(u32 stride)
{
return ((num_possible_cpus() + 1) * 8 * stride);
}
static int nvme_dbbuf_dma_alloc(struct nvme_dev *dev)
{
unsigned int mem_size = nvme_dbbuf_size(dev->db_stride);
if (dev->dbbuf_dbs)
return 0;
dev->dbbuf_dbs = dma_alloc_coherent(dev->dev, mem_size,
&dev->dbbuf_dbs_dma_addr,
GFP_KERNEL);
if (!dev->dbbuf_dbs)
return -ENOMEM;
dev->dbbuf_eis = dma_alloc_coherent(dev->dev, mem_size,
&dev->dbbuf_eis_dma_addr,
GFP_KERNEL);
if (!dev->dbbuf_eis) {
dma_free_coherent(dev->dev, mem_size,
dev->dbbuf_dbs, dev->dbbuf_dbs_dma_addr);
dev->dbbuf_dbs = NULL;
return -ENOMEM;
}
return 0;
}
static void nvme_dbbuf_dma_free(struct nvme_dev *dev)
{
unsigned int mem_size = nvme_dbbuf_size(dev->db_stride);
if (dev->dbbuf_dbs) {
dma_free_coherent(dev->dev, mem_size,
dev->dbbuf_dbs, dev->dbbuf_dbs_dma_addr);
dev->dbbuf_dbs = NULL;
}
if (dev->dbbuf_eis) {
dma_free_coherent(dev->dev, mem_size,
dev->dbbuf_eis, dev->dbbuf_eis_dma_addr);
dev->dbbuf_eis = NULL;
}
}
static void nvme_dbbuf_init(struct nvme_dev *dev,
struct nvme_queue *nvmeq, int qid)
{
if (!dev->dbbuf_dbs || !qid)
return;
nvmeq->dbbuf_sq_db = &dev->dbbuf_dbs[sq_idx(qid, dev->db_stride)];
nvmeq->dbbuf_cq_db = &dev->dbbuf_dbs[cq_idx(qid, dev->db_stride)];
nvmeq->dbbuf_sq_ei = &dev->dbbuf_eis[sq_idx(qid, dev->db_stride)];
nvmeq->dbbuf_cq_ei = &dev->dbbuf_eis[cq_idx(qid, dev->db_stride)];
}
static void nvme_dbbuf_set(struct nvme_dev *dev)
{
struct nvme_command c;
if (!dev->dbbuf_dbs)
return;
memset(&c, 0, sizeof(c));
c.dbbuf.opcode = nvme_admin_dbbuf;
c.dbbuf.prp1 = cpu_to_le64(dev->dbbuf_dbs_dma_addr);
c.dbbuf.prp2 = cpu_to_le64(dev->dbbuf_eis_dma_addr);
if (nvme_submit_sync_cmd(dev->ctrl.admin_q, &c, NULL, 0)) {
dev_warn(dev->ctrl.device, "unable to set dbbuf\n");
nvme: improve performance for virtual NVMe devices This change provides a mechanism to reduce the number of MMIO doorbell writes for the NVMe driver. When running in a virtualized environment like QEMU, the cost of an MMIO is quite hefy here. The main idea for the patch is provide the device two memory location locations: 1) to store the doorbell values so they can be lookup without the doorbell MMIO write 2) to store an event index. I believe the doorbell value is obvious, the event index not so much. Similar to the virtio specification, the virtual device can tell the driver (guest OS) not to write MMIO unless you are writing past this value. FYI: doorbell values are written by the nvme driver (guest OS) and the event index is written by the virtual device (host OS). The patch implements a new admin command that will communicate where these two memory locations reside. If the command fails, the nvme driver will work as before without any optimizations. Contributions: Eric Northup <digitaleric@google.com> Frank Swiderski <fes@google.com> Ted Tso <tytso@mit.edu> Keith Busch <keith.busch@intel.com> Just to give an idea on the performance boost with the vendor extension: Running fio [1], a stock NVMe driver I get about 200K read IOPs with my vendor patch I get about 1000K read IOPs. This was running with a null device i.e. the backing device simply returned success on every read IO request. [1] Running on a 4 core machine: fio --time_based --name=benchmark --runtime=30 --filename=/dev/nvme0n1 --nrfiles=1 --ioengine=libaio --iodepth=32 --direct=1 --invalidate=1 --verify=0 --verify_fatal=0 --numjobs=4 --rw=randread --blocksize=4k --randrepeat=false Signed-off-by: Rob Nelson <rlnelson@google.com> [mlin: port for upstream] Signed-off-by: Ming Lin <mlin@kernel.org> [koike: updated for upstream] Signed-off-by: Helen Koike <helen.koike@collabora.co.uk> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <keith.busch@intel.com>
2017-04-10 18:51:07 +03:00
/* Free memory and continue on */
nvme_dbbuf_dma_free(dev);
}
}
static inline int nvme_dbbuf_need_event(u16 event_idx, u16 new_idx, u16 old)
{
return (u16)(new_idx - event_idx - 1) < (u16)(new_idx - old);
}
/* Update dbbuf and return true if an MMIO is required */
static bool nvme_dbbuf_update_and_check_event(u16 value, u32 *dbbuf_db,
volatile u32 *dbbuf_ei)
{
if (dbbuf_db) {
u16 old_value;
/*
* Ensure that the queue is written before updating
* the doorbell in memory
*/
wmb();
old_value = *dbbuf_db;
*dbbuf_db = value;
if (!nvme_dbbuf_need_event(*dbbuf_ei, value, old_value))
return false;
}
return true;
}
/*
* Max size of iod being embedded in the request payload
*/
#define NVME_INT_PAGES 2
#define NVME_INT_BYTES(dev) (NVME_INT_PAGES * (dev)->ctrl.page_size)
/*
* Will slightly overestimate the number of pages needed. This is OK
* as it only leads to a small amount of wasted memory for the lifetime of
* the I/O.
*/
static int nvme_npages(unsigned size, struct nvme_dev *dev)
{
unsigned nprps = DIV_ROUND_UP(size + dev->ctrl.page_size,
dev->ctrl.page_size);
return DIV_ROUND_UP(8 * nprps, PAGE_SIZE - 8);
}
/*
* Calculates the number of pages needed for the SGL segments. For example a 4k
* page can accommodate 256 SGL descriptors.
*/
static int nvme_pci_npages_sgl(unsigned int num_seg)
{
return DIV_ROUND_UP(num_seg * sizeof(struct nvme_sgl_desc), PAGE_SIZE);
}
static unsigned int nvme_pci_iod_alloc_size(struct nvme_dev *dev,
unsigned int size, unsigned int nseg, bool use_sgl)
{
size_t alloc_size;
if (use_sgl)
alloc_size = sizeof(__le64 *) * nvme_pci_npages_sgl(nseg);
else
alloc_size = sizeof(__le64 *) * nvme_npages(size, dev);
return alloc_size + sizeof(struct scatterlist) * nseg;
}
static unsigned int nvme_pci_cmd_size(struct nvme_dev *dev, bool use_sgl)
{
unsigned int alloc_size = nvme_pci_iod_alloc_size(dev,
NVME_INT_BYTES(dev), NVME_INT_PAGES,
use_sgl);
return sizeof(struct nvme_iod) + alloc_size;
}
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
static int nvme_admin_init_hctx(struct blk_mq_hw_ctx *hctx, void *data,
unsigned int hctx_idx)
{
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
struct nvme_dev *dev = data;
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
struct nvme_queue *nvmeq = &dev->queues[0];
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
WARN_ON(hctx_idx != 0);
WARN_ON(dev->admin_tagset.tags[0] != hctx->tags);
WARN_ON(nvmeq->tags);
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
hctx->driver_data = nvmeq;
nvmeq->tags = &dev->admin_tagset.tags[0];
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
return 0;
}
static void nvme_admin_exit_hctx(struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx)
{
struct nvme_queue *nvmeq = hctx->driver_data;
nvmeq->tags = NULL;
}
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
static int nvme_init_hctx(struct blk_mq_hw_ctx *hctx, void *data,
unsigned int hctx_idx)
{
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
struct nvme_dev *dev = data;
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
struct nvme_queue *nvmeq = &dev->queues[hctx_idx + 1];
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
if (!nvmeq->tags)
nvmeq->tags = &dev->tagset.tags[hctx_idx];
WARN_ON(dev->tagset.tags[hctx_idx] != hctx->tags);
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
hctx->driver_data = nvmeq;
return 0;
}
static int nvme_init_request(struct blk_mq_tag_set *set, struct request *req,
unsigned int hctx_idx, unsigned int numa_node)
{
struct nvme_dev *dev = set->driver_data;
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
int queue_idx = (set == &dev->tagset) ? hctx_idx + 1 : 0;
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
struct nvme_queue *nvmeq = &dev->queues[queue_idx];
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
BUG_ON(!nvmeq);
iod->nvmeq = nvmeq;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
return 0;
}
static int nvme_pci_map_queues(struct blk_mq_tag_set *set)
{
struct nvme_dev *dev = set->driver_data;
return blk_mq_pci_map_queues(set, to_pci_dev(dev->dev),
dev->num_vecs > 1 ? 1 /* admin queue */ : 0);
}
/**
* __nvme_submit_cmd() - Copy a command into a queue and ring the doorbell
* @nvmeq: The queue to use
* @cmd: The command to send
*
* Safe to use from interrupt context
*/
static void __nvme_submit_cmd(struct nvme_queue *nvmeq,
struct nvme_command *cmd)
{
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
u16 tail = nvmeq->sq_tail;
if (nvmeq->sq_cmds_io)
memcpy_toio(&nvmeq->sq_cmds_io[tail], cmd, sizeof(*cmd));
else
memcpy(&nvmeq->sq_cmds[tail], cmd, sizeof(*cmd));
if (++tail == nvmeq->q_depth)
tail = 0;
nvme: improve performance for virtual NVMe devices This change provides a mechanism to reduce the number of MMIO doorbell writes for the NVMe driver. When running in a virtualized environment like QEMU, the cost of an MMIO is quite hefy here. The main idea for the patch is provide the device two memory location locations: 1) to store the doorbell values so they can be lookup without the doorbell MMIO write 2) to store an event index. I believe the doorbell value is obvious, the event index not so much. Similar to the virtio specification, the virtual device can tell the driver (guest OS) not to write MMIO unless you are writing past this value. FYI: doorbell values are written by the nvme driver (guest OS) and the event index is written by the virtual device (host OS). The patch implements a new admin command that will communicate where these two memory locations reside. If the command fails, the nvme driver will work as before without any optimizations. Contributions: Eric Northup <digitaleric@google.com> Frank Swiderski <fes@google.com> Ted Tso <tytso@mit.edu> Keith Busch <keith.busch@intel.com> Just to give an idea on the performance boost with the vendor extension: Running fio [1], a stock NVMe driver I get about 200K read IOPs with my vendor patch I get about 1000K read IOPs. This was running with a null device i.e. the backing device simply returned success on every read IO request. [1] Running on a 4 core machine: fio --time_based --name=benchmark --runtime=30 --filename=/dev/nvme0n1 --nrfiles=1 --ioengine=libaio --iodepth=32 --direct=1 --invalidate=1 --verify=0 --verify_fatal=0 --numjobs=4 --rw=randread --blocksize=4k --randrepeat=false Signed-off-by: Rob Nelson <rlnelson@google.com> [mlin: port for upstream] Signed-off-by: Ming Lin <mlin@kernel.org> [koike: updated for upstream] Signed-off-by: Helen Koike <helen.koike@collabora.co.uk> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <keith.busch@intel.com>
2017-04-10 18:51:07 +03:00
if (nvme_dbbuf_update_and_check_event(tail, nvmeq->dbbuf_sq_db,
nvmeq->dbbuf_sq_ei))
writel(tail, nvmeq->q_db);
nvmeq->sq_tail = tail;
}
static void **nvme_pci_iod_list(struct request *req)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
return (void **)(iod->sg + blk_rq_nr_phys_segments(req));
}
static inline bool nvme_pci_use_sgls(struct nvme_dev *dev, struct request *req)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
int nseg = blk_rq_nr_phys_segments(req);
unsigned int avg_seg_size;
if (nseg == 0)
return false;
avg_seg_size = DIV_ROUND_UP(blk_rq_payload_bytes(req), nseg);
if (!(dev->ctrl.sgls & ((1 << 0) | (1 << 1))))
return false;
if (!iod->nvmeq->qid)
return false;
if (!sgl_threshold || avg_seg_size < sgl_threshold)
return false;
return true;
}
static blk_status_t nvme_init_iod(struct request *rq, struct nvme_dev *dev)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(rq);
int nseg = blk_rq_nr_phys_segments(rq);
unsigned int size = blk_rq_payload_bytes(rq);
iod->use_sgl = nvme_pci_use_sgls(dev, rq);
if (nseg > NVME_INT_PAGES || size > NVME_INT_BYTES(dev)) {
size_t alloc_size = nvme_pci_iod_alloc_size(dev, size, nseg,
iod->use_sgl);
iod->sg = kmalloc(alloc_size, GFP_ATOMIC);
if (!iod->sg)
return BLK_STS_RESOURCE;
} else {
iod->sg = iod->inline_sg;
}
iod->aborted = 0;
iod->npages = -1;
iod->nents = 0;
iod->length = size;
return BLK_STS_OK;
}
static void nvme_free_iod(struct nvme_dev *dev, struct request *req)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
const int last_prp = dev->ctrl.page_size / sizeof(__le64) - 1;
dma_addr_t dma_addr = iod->first_dma, next_dma_addr;
int i;
if (iod->npages == 0)
dma_pool_free(dev->prp_small_pool, nvme_pci_iod_list(req)[0],
dma_addr);
for (i = 0; i < iod->npages; i++) {
void *addr = nvme_pci_iod_list(req)[i];
if (iod->use_sgl) {
struct nvme_sgl_desc *sg_list = addr;
next_dma_addr =
le64_to_cpu((sg_list[SGES_PER_PAGE - 1]).addr);
} else {
__le64 *prp_list = addr;
next_dma_addr = le64_to_cpu(prp_list[last_prp]);
}
dma_pool_free(dev->prp_page_pool, addr, dma_addr);
dma_addr = next_dma_addr;
}
if (iod->sg != iod->inline_sg)
kfree(iod->sg);
}
#ifdef CONFIG_BLK_DEV_INTEGRITY
NVMe: Metadata format support Adds support for NVMe metadata formats and exposes block devices for all namespaces regardless of their format. Namespace formats that are unusable will have disk capacity set to 0, but a handle to the block device is created to simplify device management. A namespace is not usable when the format requires host interleave block and metadata in single buffer, has no provisioned storage, or has better data but failed to register with blk integrity. The namespace has to be scanned in two phases to support separate metadata formats. The first establishes the sector size and capacity prior to invoking add_disk. If metadata is required, the capacity will be temporarilly set to 0 until it can be revalidated and registered with the integrity extenstions after add_disk completes. The driver relies on the integrity extensions to provide the metadata buffer. NVMe requires this be a single physically contiguous region, so only one integrity segment is allowed per command. If the metadata is used for T10 PI, the driver provides mappings to save and restore the reftag physical block translation. The driver provides no-op functions for generate and verify if metadata is not used for protection information. This way the setup is always provided by the block layer. If a request does not supply a required metadata buffer, the command is failed with bad address. This could only happen if a user manually disables verify/generate on such a disk. The only exception to where this is okay is if the controller is capable of stripping/generating the metadata, which is possible on some types of formats. The metadata scatter gather list now occupies the spot in the nvme_iod that used to be used to link retryable IOD's, but we don't do that anymore, so the field was unused. Signed-off-by: Keith Busch <keith.busch@intel.com>
2015-02-19 23:39:03 +03:00
static void nvme_dif_prep(u32 p, u32 v, struct t10_pi_tuple *pi)
{
if (be32_to_cpu(pi->ref_tag) == v)
pi->ref_tag = cpu_to_be32(p);
}
static void nvme_dif_complete(u32 p, u32 v, struct t10_pi_tuple *pi)
{
if (be32_to_cpu(pi->ref_tag) == p)
pi->ref_tag = cpu_to_be32(v);
}
/**
* nvme_dif_remap - remaps ref tags to bip seed and physical lba
*
* The virtual start sector is the one that was originally submitted by the
* block layer. Due to partitioning, MD/DM cloning, etc. the actual physical
* start sector may be different. Remap protection information to match the
* physical LBA on writes, and back to the original seed on reads.
*
* Type 0 and 3 do not have a ref tag, so no remapping required.
*/
static void nvme_dif_remap(struct request *req,
void (*dif_swap)(u32 p, u32 v, struct t10_pi_tuple *pi))
{
struct nvme_ns *ns = req->rq_disk->private_data;
struct bio_integrity_payload *bip;
struct t10_pi_tuple *pi;
void *p, *pmap;
u32 i, nlb, ts, phys, virt;
if (!ns->pi_type || ns->pi_type == NVME_NS_DPS_PI_TYPE3)
return;
bip = bio_integrity(req->bio);
if (!bip)
return;
pmap = kmap_atomic(bip->bip_vec->bv_page) + bip->bip_vec->bv_offset;
p = pmap;
virt = bip_get_seed(bip);
phys = nvme_block_nr(ns, blk_rq_pos(req));
nlb = (blk_rq_bytes(req) >> ns->lba_shift);
ts = ns->disk->queue->integrity.tuple_size;
NVMe: Metadata format support Adds support for NVMe metadata formats and exposes block devices for all namespaces regardless of their format. Namespace formats that are unusable will have disk capacity set to 0, but a handle to the block device is created to simplify device management. A namespace is not usable when the format requires host interleave block and metadata in single buffer, has no provisioned storage, or has better data but failed to register with blk integrity. The namespace has to be scanned in two phases to support separate metadata formats. The first establishes the sector size and capacity prior to invoking add_disk. If metadata is required, the capacity will be temporarilly set to 0 until it can be revalidated and registered with the integrity extenstions after add_disk completes. The driver relies on the integrity extensions to provide the metadata buffer. NVMe requires this be a single physically contiguous region, so only one integrity segment is allowed per command. If the metadata is used for T10 PI, the driver provides mappings to save and restore the reftag physical block translation. The driver provides no-op functions for generate and verify if metadata is not used for protection information. This way the setup is always provided by the block layer. If a request does not supply a required metadata buffer, the command is failed with bad address. This could only happen if a user manually disables verify/generate on such a disk. The only exception to where this is okay is if the controller is capable of stripping/generating the metadata, which is possible on some types of formats. The metadata scatter gather list now occupies the spot in the nvme_iod that used to be used to link retryable IOD's, but we don't do that anymore, so the field was unused. Signed-off-by: Keith Busch <keith.busch@intel.com>
2015-02-19 23:39:03 +03:00
for (i = 0; i < nlb; i++, virt++, phys++) {
pi = (struct t10_pi_tuple *)p;
dif_swap(phys, virt, pi);
p += ts;
}
kunmap_atomic(pmap);
}
#else /* CONFIG_BLK_DEV_INTEGRITY */
static void nvme_dif_remap(struct request *req,
void (*dif_swap)(u32 p, u32 v, struct t10_pi_tuple *pi))
{
}
static void nvme_dif_prep(u32 p, u32 v, struct t10_pi_tuple *pi)
{
}
static void nvme_dif_complete(u32 p, u32 v, struct t10_pi_tuple *pi)
{
}
#endif
static void nvme_print_sgl(struct scatterlist *sgl, int nents)
{
int i;
struct scatterlist *sg;
for_each_sg(sgl, sg, nents, i) {
dma_addr_t phys = sg_phys(sg);
pr_warn("sg[%d] phys_addr:%pad offset:%d length:%d "
"dma_address:%pad dma_length:%d\n",
i, &phys, sg->offset, sg->length, &sg_dma_address(sg),
sg_dma_len(sg));
}
}
static blk_status_t nvme_pci_setup_prps(struct nvme_dev *dev,
struct request *req, struct nvme_rw_command *cmnd)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
struct dma_pool *pool;
int length = blk_rq_payload_bytes(req);
struct scatterlist *sg = iod->sg;
int dma_len = sg_dma_len(sg);
u64 dma_addr = sg_dma_address(sg);
u32 page_size = dev->ctrl.page_size;
int offset = dma_addr & (page_size - 1);
__le64 *prp_list;
void **list = nvme_pci_iod_list(req);
dma_addr_t prp_dma;
int nprps, i;
length -= (page_size - offset);
if (length <= 0) {
iod->first_dma = 0;
goto done;
}
dma_len -= (page_size - offset);
if (dma_len) {
dma_addr += (page_size - offset);
} else {
sg = sg_next(sg);
dma_addr = sg_dma_address(sg);
dma_len = sg_dma_len(sg);
}
if (length <= page_size) {
iod->first_dma = dma_addr;
goto done;
}
nprps = DIV_ROUND_UP(length, page_size);
if (nprps <= (256 / 8)) {
pool = dev->prp_small_pool;
iod->npages = 0;
} else {
pool = dev->prp_page_pool;
iod->npages = 1;
}
prp_list = dma_pool_alloc(pool, GFP_ATOMIC, &prp_dma);
if (!prp_list) {
iod->first_dma = dma_addr;
iod->npages = -1;
return BLK_STS_RESOURCE;
}
list[0] = prp_list;
iod->first_dma = prp_dma;
i = 0;
for (;;) {
if (i == page_size >> 3) {
__le64 *old_prp_list = prp_list;
prp_list = dma_pool_alloc(pool, GFP_ATOMIC, &prp_dma);
if (!prp_list)
return BLK_STS_RESOURCE;
list[iod->npages++] = prp_list;
prp_list[0] = old_prp_list[i - 1];
old_prp_list[i - 1] = cpu_to_le64(prp_dma);
i = 1;
}
prp_list[i++] = cpu_to_le64(dma_addr);
dma_len -= page_size;
dma_addr += page_size;
length -= page_size;
if (length <= 0)
break;
if (dma_len > 0)
continue;
if (unlikely(dma_len < 0))
goto bad_sgl;
sg = sg_next(sg);
dma_addr = sg_dma_address(sg);
dma_len = sg_dma_len(sg);
}
done:
cmnd->dptr.prp1 = cpu_to_le64(sg_dma_address(iod->sg));
cmnd->dptr.prp2 = cpu_to_le64(iod->first_dma);
return BLK_STS_OK;
bad_sgl:
WARN(DO_ONCE(nvme_print_sgl, iod->sg, iod->nents),
"Invalid SGL for payload:%d nents:%d\n",
blk_rq_payload_bytes(req), iod->nents);
return BLK_STS_IOERR;
}
static void nvme_pci_sgl_set_data(struct nvme_sgl_desc *sge,
struct scatterlist *sg)
{
sge->addr = cpu_to_le64(sg_dma_address(sg));
sge->length = cpu_to_le32(sg_dma_len(sg));
sge->type = NVME_SGL_FMT_DATA_DESC << 4;
}
static void nvme_pci_sgl_set_seg(struct nvme_sgl_desc *sge,
dma_addr_t dma_addr, int entries)
{
sge->addr = cpu_to_le64(dma_addr);
if (entries < SGES_PER_PAGE) {
sge->length = cpu_to_le32(entries * sizeof(*sge));
sge->type = NVME_SGL_FMT_LAST_SEG_DESC << 4;
} else {
sge->length = cpu_to_le32(PAGE_SIZE);
sge->type = NVME_SGL_FMT_SEG_DESC << 4;
}
}
static blk_status_t nvme_pci_setup_sgls(struct nvme_dev *dev,
struct request *req, struct nvme_rw_command *cmd, int entries)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
struct dma_pool *pool;
struct nvme_sgl_desc *sg_list;
struct scatterlist *sg = iod->sg;
dma_addr_t sgl_dma;
int i = 0;
/* setting the transfer type as SGL */
cmd->flags = NVME_CMD_SGL_METABUF;
if (entries == 1) {
nvme_pci_sgl_set_data(&cmd->dptr.sgl, sg);
return BLK_STS_OK;
}
if (entries <= (256 / sizeof(struct nvme_sgl_desc))) {
pool = dev->prp_small_pool;
iod->npages = 0;
} else {
pool = dev->prp_page_pool;
iod->npages = 1;
}
sg_list = dma_pool_alloc(pool, GFP_ATOMIC, &sgl_dma);
if (!sg_list) {
iod->npages = -1;
return BLK_STS_RESOURCE;
}
nvme_pci_iod_list(req)[0] = sg_list;
iod->first_dma = sgl_dma;
nvme_pci_sgl_set_seg(&cmd->dptr.sgl, sgl_dma, entries);
do {
if (i == SGES_PER_PAGE) {
struct nvme_sgl_desc *old_sg_desc = sg_list;
struct nvme_sgl_desc *link = &old_sg_desc[i - 1];
sg_list = dma_pool_alloc(pool, GFP_ATOMIC, &sgl_dma);
if (!sg_list)
return BLK_STS_RESOURCE;
i = 0;
nvme_pci_iod_list(req)[iod->npages++] = sg_list;
sg_list[i++] = *link;
nvme_pci_sgl_set_seg(link, sgl_dma, entries);
}
nvme_pci_sgl_set_data(&sg_list[i++], sg);
sg = sg_next(sg);
} while (--entries > 0);
return BLK_STS_OK;
}
static blk_status_t nvme_map_data(struct nvme_dev *dev, struct request *req,
struct nvme_command *cmnd)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
struct request_queue *q = req->q;
enum dma_data_direction dma_dir = rq_data_dir(req) ?
DMA_TO_DEVICE : DMA_FROM_DEVICE;
blk_status_t ret = BLK_STS_IOERR;
int nr_mapped;
sg_init_table(iod->sg, blk_rq_nr_phys_segments(req));
iod->nents = blk_rq_map_sg(q, req, iod->sg);
if (!iod->nents)
goto out;
ret = BLK_STS_RESOURCE;
nr_mapped = dma_map_sg_attrs(dev->dev, iod->sg, iod->nents, dma_dir,
DMA_ATTR_NO_WARN);
if (!nr_mapped)
goto out;
if (iod->use_sgl)
ret = nvme_pci_setup_sgls(dev, req, &cmnd->rw, nr_mapped);
else
ret = nvme_pci_setup_prps(dev, req, &cmnd->rw);
if (ret != BLK_STS_OK)
goto out_unmap;
ret = BLK_STS_IOERR;
if (blk_integrity_rq(req)) {
if (blk_rq_count_integrity_sg(q, req->bio) != 1)
goto out_unmap;
sg_init_table(&iod->meta_sg, 1);
if (blk_rq_map_integrity_sg(q, req->bio, &iod->meta_sg) != 1)
goto out_unmap;
if (req_op(req) == REQ_OP_WRITE)
nvme_dif_remap(req, nvme_dif_prep);
if (!dma_map_sg(dev->dev, &iod->meta_sg, 1, dma_dir))
goto out_unmap;
}
if (blk_integrity_rq(req))
cmnd->rw.metadata = cpu_to_le64(sg_dma_address(&iod->meta_sg));
return BLK_STS_OK;
out_unmap:
dma_unmap_sg(dev->dev, iod->sg, iod->nents, dma_dir);
out:
return ret;
}
static void nvme_unmap_data(struct nvme_dev *dev, struct request *req)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
enum dma_data_direction dma_dir = rq_data_dir(req) ?
DMA_TO_DEVICE : DMA_FROM_DEVICE;
if (iod->nents) {
dma_unmap_sg(dev->dev, iod->sg, iod->nents, dma_dir);
if (blk_integrity_rq(req)) {
if (req_op(req) == REQ_OP_READ)
nvme_dif_remap(req, nvme_dif_complete);
dma_unmap_sg(dev->dev, &iod->meta_sg, 1, dma_dir);
NVMe: Metadata format support Adds support for NVMe metadata formats and exposes block devices for all namespaces regardless of their format. Namespace formats that are unusable will have disk capacity set to 0, but a handle to the block device is created to simplify device management. A namespace is not usable when the format requires host interleave block and metadata in single buffer, has no provisioned storage, or has better data but failed to register with blk integrity. The namespace has to be scanned in two phases to support separate metadata formats. The first establishes the sector size and capacity prior to invoking add_disk. If metadata is required, the capacity will be temporarilly set to 0 until it can be revalidated and registered with the integrity extenstions after add_disk completes. The driver relies on the integrity extensions to provide the metadata buffer. NVMe requires this be a single physically contiguous region, so only one integrity segment is allowed per command. If the metadata is used for T10 PI, the driver provides mappings to save and restore the reftag physical block translation. The driver provides no-op functions for generate and verify if metadata is not used for protection information. This way the setup is always provided by the block layer. If a request does not supply a required metadata buffer, the command is failed with bad address. This could only happen if a user manually disables verify/generate on such a disk. The only exception to where this is okay is if the controller is capable of stripping/generating the metadata, which is possible on some types of formats. The metadata scatter gather list now occupies the spot in the nvme_iod that used to be used to link retryable IOD's, but we don't do that anymore, so the field was unused. Signed-off-by: Keith Busch <keith.busch@intel.com>
2015-02-19 23:39:03 +03:00
}
}
NVMe: Metadata format support Adds support for NVMe metadata formats and exposes block devices for all namespaces regardless of their format. Namespace formats that are unusable will have disk capacity set to 0, but a handle to the block device is created to simplify device management. A namespace is not usable when the format requires host interleave block and metadata in single buffer, has no provisioned storage, or has better data but failed to register with blk integrity. The namespace has to be scanned in two phases to support separate metadata formats. The first establishes the sector size and capacity prior to invoking add_disk. If metadata is required, the capacity will be temporarilly set to 0 until it can be revalidated and registered with the integrity extenstions after add_disk completes. The driver relies on the integrity extensions to provide the metadata buffer. NVMe requires this be a single physically contiguous region, so only one integrity segment is allowed per command. If the metadata is used for T10 PI, the driver provides mappings to save and restore the reftag physical block translation. The driver provides no-op functions for generate and verify if metadata is not used for protection information. This way the setup is always provided by the block layer. If a request does not supply a required metadata buffer, the command is failed with bad address. This could only happen if a user manually disables verify/generate on such a disk. The only exception to where this is okay is if the controller is capable of stripping/generating the metadata, which is possible on some types of formats. The metadata scatter gather list now occupies the spot in the nvme_iod that used to be used to link retryable IOD's, but we don't do that anymore, so the field was unused. Signed-off-by: Keith Busch <keith.busch@intel.com>
2015-02-19 23:39:03 +03:00
nvme_cleanup_cmd(req);
nvme_free_iod(dev, req);
}
/*
* NOTE: ns is NULL when called on the admin queue.
*/
static blk_status_t nvme_queue_rq(struct blk_mq_hw_ctx *hctx,
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
const struct blk_mq_queue_data *bd)
{
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
struct nvme_ns *ns = hctx->queue->queuedata;
struct nvme_queue *nvmeq = hctx->driver_data;
struct nvme_dev *dev = nvmeq->dev;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
struct request *req = bd->rq;
struct nvme_command cmnd;
blk_status_t ret;
NVMe: Metadata format support Adds support for NVMe metadata formats and exposes block devices for all namespaces regardless of their format. Namespace formats that are unusable will have disk capacity set to 0, but a handle to the block device is created to simplify device management. A namespace is not usable when the format requires host interleave block and metadata in single buffer, has no provisioned storage, or has better data but failed to register with blk integrity. The namespace has to be scanned in two phases to support separate metadata formats. The first establishes the sector size and capacity prior to invoking add_disk. If metadata is required, the capacity will be temporarilly set to 0 until it can be revalidated and registered with the integrity extenstions after add_disk completes. The driver relies on the integrity extensions to provide the metadata buffer. NVMe requires this be a single physically contiguous region, so only one integrity segment is allowed per command. If the metadata is used for T10 PI, the driver provides mappings to save and restore the reftag physical block translation. The driver provides no-op functions for generate and verify if metadata is not used for protection information. This way the setup is always provided by the block layer. If a request does not supply a required metadata buffer, the command is failed with bad address. This could only happen if a user manually disables verify/generate on such a disk. The only exception to where this is okay is if the controller is capable of stripping/generating the metadata, which is possible on some types of formats. The metadata scatter gather list now occupies the spot in the nvme_iod that used to be used to link retryable IOD's, but we don't do that anymore, so the field was unused. Signed-off-by: Keith Busch <keith.busch@intel.com>
2015-02-19 23:39:03 +03:00
/*
* We should not need to do this, but we're still using this to
* ensure we can drain requests on a dying queue.
*/
if (unlikely(nvmeq->cq_vector < 0))
return BLK_STS_IOERR;
ret = nvme_setup_cmd(ns, req, &cmnd);
if (ret)
return ret;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
ret = nvme_init_iod(req, dev);
if (ret)
goto out_free_cmd;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
if (blk_rq_nr_phys_segments(req)) {
ret = nvme_map_data(dev, req, &cmnd);
if (ret)
goto out_cleanup_iod;
}
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
blk_mq_start_request(req);
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
spin_lock(&nvmeq->sq_lock);
__nvme_submit_cmd(nvmeq, &cmnd);
spin_unlock(&nvmeq->sq_lock);
return BLK_STS_OK;
out_cleanup_iod:
nvme_free_iod(dev, req);
out_free_cmd:
nvme_cleanup_cmd(req);
return ret;
}
NVMe: Metadata format support Adds support for NVMe metadata formats and exposes block devices for all namespaces regardless of their format. Namespace formats that are unusable will have disk capacity set to 0, but a handle to the block device is created to simplify device management. A namespace is not usable when the format requires host interleave block and metadata in single buffer, has no provisioned storage, or has better data but failed to register with blk integrity. The namespace has to be scanned in two phases to support separate metadata formats. The first establishes the sector size and capacity prior to invoking add_disk. If metadata is required, the capacity will be temporarilly set to 0 until it can be revalidated and registered with the integrity extenstions after add_disk completes. The driver relies on the integrity extensions to provide the metadata buffer. NVMe requires this be a single physically contiguous region, so only one integrity segment is allowed per command. If the metadata is used for T10 PI, the driver provides mappings to save and restore the reftag physical block translation. The driver provides no-op functions for generate and verify if metadata is not used for protection information. This way the setup is always provided by the block layer. If a request does not supply a required metadata buffer, the command is failed with bad address. This could only happen if a user manually disables verify/generate on such a disk. The only exception to where this is okay is if the controller is capable of stripping/generating the metadata, which is possible on some types of formats. The metadata scatter gather list now occupies the spot in the nvme_iod that used to be used to link retryable IOD's, but we don't do that anymore, so the field was unused. Signed-off-by: Keith Busch <keith.busch@intel.com>
2015-02-19 23:39:03 +03:00
static void nvme_pci_complete_rq(struct request *req)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
nvme_unmap_data(iod->nvmeq->dev, req);
nvme_complete_rq(req);
}
/* We read the CQE phase first to check if the rest of the entry is valid */
static inline bool nvme_cqe_pending(struct nvme_queue *nvmeq)
{
return (le16_to_cpu(nvmeq->cqes[nvmeq->cq_head].status) & 1) ==
nvmeq->cq_phase;
}
static inline void nvme_ring_cq_doorbell(struct nvme_queue *nvmeq)
{
u16 head = nvmeq->cq_head;
if (likely(nvmeq->cq_vector >= 0)) {
if (nvme_dbbuf_update_and_check_event(head, nvmeq->dbbuf_cq_db,
nvmeq->dbbuf_cq_ei))
writel(head, nvmeq->q_db + nvmeq->dev->db_stride);
}
}
static inline void nvme_handle_cqe(struct nvme_queue *nvmeq, u16 idx)
{
volatile struct nvme_completion *cqe = &nvmeq->cqes[idx];
struct request *req;
if (unlikely(cqe->command_id >= nvmeq->q_depth)) {
dev_warn(nvmeq->dev->ctrl.device,
"invalid id %d completed on queue %d\n",
cqe->command_id, le16_to_cpu(cqe->sq_id));
return;
}
/*
* AEN requests are special as they don't time out and can
* survive any kind of queue freeze and often don't respond to
* aborts. We don't even bother to allocate a struct request
* for them but rather special case them here.
*/
if (unlikely(nvmeq->qid == 0 &&
cqe->command_id >= NVME_AQ_BLK_MQ_DEPTH)) {
nvme_complete_async_event(&nvmeq->dev->ctrl,
cqe->status, &cqe->result);
return;
}
req = blk_mq_tag_to_rq(*nvmeq->tags, cqe->command_id);
nvme_end_request(req, cqe->status, cqe->result);
}
static void nvme_complete_cqes(struct nvme_queue *nvmeq, u16 start, u16 end)
{
while (start != end) {
nvme_handle_cqe(nvmeq, start);
if (++start == nvmeq->q_depth)
start = 0;
}
}
static inline void nvme_update_cq_head(struct nvme_queue *nvmeq)
{
if (++nvmeq->cq_head == nvmeq->q_depth) {
nvmeq->cq_head = 0;
nvmeq->cq_phase = !nvmeq->cq_phase;
}
}
static inline bool nvme_process_cq(struct nvme_queue *nvmeq, u16 *start,
u16 *end, int tag)
{
bool found = false;
*start = nvmeq->cq_head;
while (!found && nvme_cqe_pending(nvmeq)) {
if (nvmeq->cqes[nvmeq->cq_head].command_id == tag)
found = true;
nvme_update_cq_head(nvmeq);
}
*end = nvmeq->cq_head;
if (*start != *end)
nvme_ring_cq_doorbell(nvmeq);
return found;
}
static irqreturn_t nvme_irq(int irq, void *data)
{
struct nvme_queue *nvmeq = data;
u16 start, end;
spin_lock(&nvmeq->cq_lock);
nvme_process_cq(nvmeq, &start, &end, -1);
spin_unlock(&nvmeq->cq_lock);
if (start == end)
return IRQ_NONE;
nvme_complete_cqes(nvmeq, start, end);
return IRQ_HANDLED;
}
static irqreturn_t nvme_irq_check(int irq, void *data)
{
struct nvme_queue *nvmeq = data;
if (nvme_cqe_pending(nvmeq))
return IRQ_WAKE_THREAD;
return IRQ_NONE;
}
static int __nvme_poll(struct nvme_queue *nvmeq, unsigned int tag)
{
u16 start, end;
bool found;
if (!nvme_cqe_pending(nvmeq))
return 0;
spin_lock_irq(&nvmeq->cq_lock);
found = nvme_process_cq(nvmeq, &start, &end, tag);
spin_unlock_irq(&nvmeq->cq_lock);
nvme_complete_cqes(nvmeq, start, end);
return found;
}
static int nvme_poll(struct blk_mq_hw_ctx *hctx, unsigned int tag)
{
struct nvme_queue *nvmeq = hctx->driver_data;
return __nvme_poll(nvmeq, tag);
}
static void nvme_pci_submit_async_event(struct nvme_ctrl *ctrl)
{
struct nvme_dev *dev = to_nvme_dev(ctrl);
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
struct nvme_queue *nvmeq = &dev->queues[0];
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
struct nvme_command c;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
memset(&c, 0, sizeof(c));
c.common.opcode = nvme_admin_async_event;
c.common.command_id = NVME_AQ_BLK_MQ_DEPTH;
spin_lock(&nvmeq->sq_lock);
__nvme_submit_cmd(nvmeq, &c);
spin_unlock(&nvmeq->sq_lock);
}
static int adapter_delete_queue(struct nvme_dev *dev, u8 opcode, u16 id)
{
struct nvme_command c;
memset(&c, 0, sizeof(c));
c.delete_queue.opcode = opcode;
c.delete_queue.qid = cpu_to_le16(id);
return nvme_submit_sync_cmd(dev->ctrl.admin_q, &c, NULL, 0);
}
static int adapter_alloc_cq(struct nvme_dev *dev, u16 qid,
struct nvme_queue *nvmeq)
{
struct nvme_command c;
int flags = NVME_QUEUE_PHYS_CONTIG | NVME_CQ_IRQ_ENABLED;
/*
* Note: we (ab)use the fact that the prp fields survive if no data
* is attached to the request.
*/
memset(&c, 0, sizeof(c));
c.create_cq.opcode = nvme_admin_create_cq;
c.create_cq.prp1 = cpu_to_le64(nvmeq->cq_dma_addr);
c.create_cq.cqid = cpu_to_le16(qid);
c.create_cq.qsize = cpu_to_le16(nvmeq->q_depth - 1);
c.create_cq.cq_flags = cpu_to_le16(flags);
c.create_cq.irq_vector = cpu_to_le16(nvmeq->cq_vector);
return nvme_submit_sync_cmd(dev->ctrl.admin_q, &c, NULL, 0);
}
static int adapter_alloc_sq(struct nvme_dev *dev, u16 qid,
struct nvme_queue *nvmeq)
{
struct nvme_command c;
int flags = NVME_QUEUE_PHYS_CONTIG;
/*
* Note: we (ab)use the fact that the prp fields survive if no data
* is attached to the request.
*/
memset(&c, 0, sizeof(c));
c.create_sq.opcode = nvme_admin_create_sq;
c.create_sq.prp1 = cpu_to_le64(nvmeq->sq_dma_addr);
c.create_sq.sqid = cpu_to_le16(qid);
c.create_sq.qsize = cpu_to_le16(nvmeq->q_depth - 1);
c.create_sq.sq_flags = cpu_to_le16(flags);
c.create_sq.cqid = cpu_to_le16(qid);
return nvme_submit_sync_cmd(dev->ctrl.admin_q, &c, NULL, 0);
}
static int adapter_delete_cq(struct nvme_dev *dev, u16 cqid)
{
return adapter_delete_queue(dev, nvme_admin_delete_cq, cqid);
}
static int adapter_delete_sq(struct nvme_dev *dev, u16 sqid)
{
return adapter_delete_queue(dev, nvme_admin_delete_sq, sqid);
}
static void abort_endio(struct request *req, blk_status_t error)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
struct nvme_queue *nvmeq = iod->nvmeq;
dev_warn(nvmeq->dev->ctrl.device,
"Abort status: 0x%x", nvme_req(req)->status);
atomic_inc(&nvmeq->dev->ctrl.abort_limit);
blk_mq_free_request(req);
}
static bool nvme_should_reset(struct nvme_dev *dev, u32 csts)
{
/* If true, indicates loss of adapter communication, possibly by a
* NVMe Subsystem reset.
*/
bool nssro = dev->subsystem && (csts & NVME_CSTS_NSSRO);
/* If there is a reset/reinit ongoing, we shouldn't reset again. */
switch (dev->ctrl.state) {
case NVME_CTRL_RESETTING:
case NVME_CTRL_CONNECTING:
return false;
default:
break;
}
/* We shouldn't reset unless the controller is on fatal error state
* _or_ if we lost the communication with it.
*/
if (!(csts & NVME_CSTS_CFS) && !nssro)
return false;
return true;
}
static void nvme_warn_reset(struct nvme_dev *dev, u32 csts)
{
/* Read a config register to help see what died. */
u16 pci_status;
int result;
result = pci_read_config_word(to_pci_dev(dev->dev), PCI_STATUS,
&pci_status);
if (result == PCIBIOS_SUCCESSFUL)
dev_warn(dev->ctrl.device,
"controller is down; will reset: CSTS=0x%x, PCI_STATUS=0x%hx\n",
csts, pci_status);
else
dev_warn(dev->ctrl.device,
"controller is down; will reset: CSTS=0x%x, PCI_STATUS read failed (%d)\n",
csts, result);
}
static enum blk_eh_timer_return nvme_timeout(struct request *req, bool reserved)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
struct nvme_queue *nvmeq = iod->nvmeq;
struct nvme_dev *dev = nvmeq->dev;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
struct request *abort_req;
struct nvme_command cmd;
u32 csts = readl(dev->bar + NVME_REG_CSTS);
/* If PCI error recovery process is happening, we cannot reset or
* the recovery mechanism will surely fail.
*/
mb();
if (pci_channel_offline(to_pci_dev(dev->dev)))
return BLK_EH_RESET_TIMER;
/*
* Reset immediately if the controller is failed
*/
if (nvme_should_reset(dev, csts)) {
nvme_warn_reset(dev, csts);
nvme_dev_disable(dev, false);
nvme_reset_ctrl(&dev->ctrl);
return BLK_EH_HANDLED;
}
/*
* Did we miss an interrupt?
*/
if (__nvme_poll(nvmeq, req->tag)) {
dev_warn(dev->ctrl.device,
"I/O %d QID %d timeout, completion polled\n",
req->tag, nvmeq->qid);
return BLK_EH_HANDLED;
}
/*
* Shutdown immediately if controller times out while starting. The
* reset work will see the pci device disabled when it gets the forced
* cancellation error. All outstanding requests are completed on
* shutdown, so we return BLK_EH_HANDLED.
*/
switch (dev->ctrl.state) {
case NVME_CTRL_CONNECTING:
case NVME_CTRL_RESETTING:
dev_warn(dev->ctrl.device,
"I/O %d QID %d timeout, disable controller\n",
req->tag, nvmeq->qid);
nvme_dev_disable(dev, false);
nvme_req(req)->flags |= NVME_REQ_CANCELLED;
return BLK_EH_HANDLED;
default:
break;
}
/*
* Shutdown the controller immediately and schedule a reset if the
* command was already aborted once before and still hasn't been
* returned to the driver, or if this is the admin queue.
*/
if (!nvmeq->qid || iod->aborted) {
dev_warn(dev->ctrl.device,
"I/O %d QID %d timeout, reset controller\n",
req->tag, nvmeq->qid);
nvme_dev_disable(dev, false);
nvme_reset_ctrl(&dev->ctrl);
/*
* Mark the request as handled, since the inline shutdown
* forces all outstanding requests to complete.
*/
nvme_req(req)->flags |= NVME_REQ_CANCELLED;
return BLK_EH_HANDLED;
}
if (atomic_dec_return(&dev->ctrl.abort_limit) < 0) {
atomic_inc(&dev->ctrl.abort_limit);
return BLK_EH_RESET_TIMER;
}
iod->aborted = 1;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
memset(&cmd, 0, sizeof(cmd));
cmd.abort.opcode = nvme_admin_abort_cmd;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
cmd.abort.cid = req->tag;
cmd.abort.sqid = cpu_to_le16(nvmeq->qid);
dev_warn(nvmeq->dev->ctrl.device,
"I/O %d QID %d timeout, aborting\n",
req->tag, nvmeq->qid);
abort_req = nvme_alloc_request(dev->ctrl.admin_q, &cmd,
BLK_MQ_REQ_NOWAIT, NVME_QID_ANY);
if (IS_ERR(abort_req)) {
atomic_inc(&dev->ctrl.abort_limit);
return BLK_EH_RESET_TIMER;
}
abort_req->timeout = ADMIN_TIMEOUT;
abort_req->end_io_data = NULL;
blk_execute_rq_nowait(abort_req->q, NULL, abort_req, 0, abort_endio);
/*
* The aborted req will be completed on receiving the abort req.
* We enable the timer again. If hit twice, it'll cause a device reset,
* as the device then is in a faulty state.
*/
return BLK_EH_RESET_TIMER;
}
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
static void nvme_free_queue(struct nvme_queue *nvmeq)
{
dma_free_coherent(nvmeq->q_dmadev, CQ_SIZE(nvmeq->q_depth),
(void *)nvmeq->cqes, nvmeq->cq_dma_addr);
if (nvmeq->sq_cmds)
dma_free_coherent(nvmeq->q_dmadev, SQ_SIZE(nvmeq->q_depth),
nvmeq->sq_cmds, nvmeq->sq_dma_addr);
}
static void nvme_free_queues(struct nvme_dev *dev, int lowest)
{
int i;
for (i = dev->ctrl.queue_count - 1; i >= lowest; i--) {
dev->ctrl.queue_count--;
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
nvme_free_queue(&dev->queues[i]);
}
}
/**
* nvme_suspend_queue - put queue into suspended state
* @nvmeq - queue to suspend
*/
static int nvme_suspend_queue(struct nvme_queue *nvmeq)
{
int vector;
spin_lock_irq(&nvmeq->cq_lock);
if (nvmeq->cq_vector == -1) {
spin_unlock_irq(&nvmeq->cq_lock);
return 1;
}
vector = nvmeq->cq_vector;
nvmeq->dev->online_queues--;
nvmeq->cq_vector = -1;
spin_unlock_irq(&nvmeq->cq_lock);
/*
* Ensure that nvme_queue_rq() sees it ->cq_vector == -1 without
* having to grab the lock.
*/
mb();
if (!nvmeq->qid && nvmeq->dev->ctrl.admin_q)
blk_mq_quiesce_queue(nvmeq->dev->ctrl.admin_q);
pci_free_irq(to_pci_dev(nvmeq->dev->dev), vector, nvmeq);
return 0;
}
static void nvme_disable_admin_queue(struct nvme_dev *dev, bool shutdown)
{
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
struct nvme_queue *nvmeq = &dev->queues[0];
u16 start, end;
if (shutdown)
nvme_shutdown_ctrl(&dev->ctrl);
else
nvme_disable_ctrl(&dev->ctrl, dev->ctrl.cap);
spin_lock_irq(&nvmeq->cq_lock);
nvme_process_cq(nvmeq, &start, &end, -1);
spin_unlock_irq(&nvmeq->cq_lock);
nvme_complete_cqes(nvmeq, start, end);
}
static int nvme_cmb_qdepth(struct nvme_dev *dev, int nr_io_queues,
int entry_size)
{
int q_depth = dev->q_depth;
unsigned q_size_aligned = roundup(q_depth * entry_size,
dev->ctrl.page_size);
if (q_size_aligned * nr_io_queues > dev->cmb_size) {
u64 mem_per_q = div_u64(dev->cmb_size, nr_io_queues);
mem_per_q = round_down(mem_per_q, dev->ctrl.page_size);
q_depth = div_u64(mem_per_q, entry_size);
/*
* Ensure the reduced q_depth is above some threshold where it
* would be better to map queues in system memory with the
* original depth
*/
if (q_depth < 64)
return -ENOMEM;
}
return q_depth;
}
static int nvme_alloc_sq_cmds(struct nvme_dev *dev, struct nvme_queue *nvmeq,
int qid, int depth)
{
/* CMB SQEs will be mapped before creation */
if (qid && dev->cmb && use_cmb_sqes && (dev->cmbsz & NVME_CMBSZ_SQS))
return 0;
nvmeq->sq_cmds = dma_alloc_coherent(dev->dev, SQ_SIZE(depth),
&nvmeq->sq_dma_addr, GFP_KERNEL);
if (!nvmeq->sq_cmds)
return -ENOMEM;
return 0;
}
static int nvme_alloc_queue(struct nvme_dev *dev, int qid, int depth)
{
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
struct nvme_queue *nvmeq = &dev->queues[qid];
if (dev->ctrl.queue_count > qid)
return 0;
nvmeq->cqes = dma_zalloc_coherent(dev->dev, CQ_SIZE(depth),
&nvmeq->cq_dma_addr, GFP_KERNEL);
if (!nvmeq->cqes)
goto free_nvmeq;
if (nvme_alloc_sq_cmds(dev, nvmeq, qid, depth))
goto free_cqdma;
nvmeq->q_dmadev = dev->dev;
nvmeq->dev = dev;
spin_lock_init(&nvmeq->sq_lock);
spin_lock_init(&nvmeq->cq_lock);
nvmeq->cq_head = 0;
nvmeq->cq_phase = 1;
nvmeq->q_db = &dev->dbs[qid * 2 * dev->db_stride];
nvmeq->q_depth = depth;
nvmeq->qid = qid;
nvmeq->cq_vector = -1;
dev->ctrl.queue_count++;
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
return 0;
free_cqdma:
dma_free_coherent(dev->dev, CQ_SIZE(depth), (void *)nvmeq->cqes,
nvmeq->cq_dma_addr);
free_nvmeq:
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
return -ENOMEM;
}
static int queue_request_irq(struct nvme_queue *nvmeq)
{
struct pci_dev *pdev = to_pci_dev(nvmeq->dev->dev);
int nr = nvmeq->dev->ctrl.instance;
if (use_threaded_interrupts) {
return pci_request_irq(pdev, nvmeq->cq_vector, nvme_irq_check,
nvme_irq, nvmeq, "nvme%dq%d", nr, nvmeq->qid);
} else {
return pci_request_irq(pdev, nvmeq->cq_vector, nvme_irq,
NULL, nvmeq, "nvme%dq%d", nr, nvmeq->qid);
}
}
static void nvme_init_queue(struct nvme_queue *nvmeq, u16 qid)
{
struct nvme_dev *dev = nvmeq->dev;
spin_lock_irq(&nvmeq->cq_lock);
nvmeq->sq_tail = 0;
nvmeq->cq_head = 0;
nvmeq->cq_phase = 1;
nvmeq->q_db = &dev->dbs[qid * 2 * dev->db_stride];
memset((void *)nvmeq->cqes, 0, CQ_SIZE(nvmeq->q_depth));
nvme: improve performance for virtual NVMe devices This change provides a mechanism to reduce the number of MMIO doorbell writes for the NVMe driver. When running in a virtualized environment like QEMU, the cost of an MMIO is quite hefy here. The main idea for the patch is provide the device two memory location locations: 1) to store the doorbell values so they can be lookup without the doorbell MMIO write 2) to store an event index. I believe the doorbell value is obvious, the event index not so much. Similar to the virtio specification, the virtual device can tell the driver (guest OS) not to write MMIO unless you are writing past this value. FYI: doorbell values are written by the nvme driver (guest OS) and the event index is written by the virtual device (host OS). The patch implements a new admin command that will communicate where these two memory locations reside. If the command fails, the nvme driver will work as before without any optimizations. Contributions: Eric Northup <digitaleric@google.com> Frank Swiderski <fes@google.com> Ted Tso <tytso@mit.edu> Keith Busch <keith.busch@intel.com> Just to give an idea on the performance boost with the vendor extension: Running fio [1], a stock NVMe driver I get about 200K read IOPs with my vendor patch I get about 1000K read IOPs. This was running with a null device i.e. the backing device simply returned success on every read IO request. [1] Running on a 4 core machine: fio --time_based --name=benchmark --runtime=30 --filename=/dev/nvme0n1 --nrfiles=1 --ioengine=libaio --iodepth=32 --direct=1 --invalidate=1 --verify=0 --verify_fatal=0 --numjobs=4 --rw=randread --blocksize=4k --randrepeat=false Signed-off-by: Rob Nelson <rlnelson@google.com> [mlin: port for upstream] Signed-off-by: Ming Lin <mlin@kernel.org> [koike: updated for upstream] Signed-off-by: Helen Koike <helen.koike@collabora.co.uk> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <keith.busch@intel.com>
2017-04-10 18:51:07 +03:00
nvme_dbbuf_init(dev, nvmeq, qid);
dev->online_queues++;
spin_unlock_irq(&nvmeq->cq_lock);
}
static int nvme_create_queue(struct nvme_queue *nvmeq, int qid)
{
struct nvme_dev *dev = nvmeq->dev;
int result;
if (dev->cmb && use_cmb_sqes && (dev->cmbsz & NVME_CMBSZ_SQS)) {
unsigned offset = (qid - 1) * roundup(SQ_SIZE(nvmeq->q_depth),
dev->ctrl.page_size);
nvmeq->sq_dma_addr = dev->cmb_bus_addr + offset;
nvmeq->sq_cmds_io = dev->cmb + offset;
}
/*
* A queue's vector matches the queue identifier unless the controller
* has only one vector available.
*/
nvmeq->cq_vector = dev->num_vecs == 1 ? 0 : qid;
result = adapter_alloc_cq(dev, qid, nvmeq);
if (result < 0)
goto release_vector;
result = adapter_alloc_sq(dev, qid, nvmeq);
if (result < 0)
goto release_cq;
nvme_init_queue(nvmeq, qid);
result = queue_request_irq(nvmeq);
if (result < 0)
goto release_sq;
return result;
release_sq:
dev->online_queues--;
adapter_delete_sq(dev, qid);
release_cq:
adapter_delete_cq(dev, qid);
release_vector:
nvmeq->cq_vector = -1;
return result;
}
static const struct blk_mq_ops nvme_mq_admin_ops = {
.queue_rq = nvme_queue_rq,
.complete = nvme_pci_complete_rq,
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
.init_hctx = nvme_admin_init_hctx,
.exit_hctx = nvme_admin_exit_hctx,
.init_request = nvme_init_request,
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
.timeout = nvme_timeout,
};
static const struct blk_mq_ops nvme_mq_ops = {
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
.queue_rq = nvme_queue_rq,
.complete = nvme_pci_complete_rq,
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
.init_hctx = nvme_init_hctx,
.init_request = nvme_init_request,
.map_queues = nvme_pci_map_queues,
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
.timeout = nvme_timeout,
.poll = nvme_poll,
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
};
static void nvme_dev_remove_admin(struct nvme_dev *dev)
{
if (dev->ctrl.admin_q && !blk_queue_dying(dev->ctrl.admin_q)) {
/*
* If the controller was reset during removal, it's possible
* user requests may be waiting on a stopped queue. Start the
* queue to flush these to completion.
*/
blk_mq_unquiesce_queue(dev->ctrl.admin_q);
blk_cleanup_queue(dev->ctrl.admin_q);
blk_mq_free_tag_set(&dev->admin_tagset);
}
}
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
static int nvme_alloc_admin_tags(struct nvme_dev *dev)
{
if (!dev->ctrl.admin_q) {
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
dev->admin_tagset.ops = &nvme_mq_admin_ops;
dev->admin_tagset.nr_hw_queues = 1;
dev->admin_tagset.queue_depth = NVME_AQ_MQ_TAG_DEPTH;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
dev->admin_tagset.timeout = ADMIN_TIMEOUT;
dev->admin_tagset.numa_node = dev_to_node(dev->dev);
dev->admin_tagset.cmd_size = nvme_pci_cmd_size(dev, false);
dev->admin_tagset.flags = BLK_MQ_F_NO_SCHED;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
dev->admin_tagset.driver_data = dev;
if (blk_mq_alloc_tag_set(&dev->admin_tagset))
return -ENOMEM;
dev->ctrl.admin_tagset = &dev->admin_tagset;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
dev->ctrl.admin_q = blk_mq_init_queue(&dev->admin_tagset);
if (IS_ERR(dev->ctrl.admin_q)) {
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
blk_mq_free_tag_set(&dev->admin_tagset);
return -ENOMEM;
}
if (!blk_get_queue(dev->ctrl.admin_q)) {
nvme_dev_remove_admin(dev);
dev->ctrl.admin_q = NULL;
return -ENODEV;
}
} else
blk_mq_unquiesce_queue(dev->ctrl.admin_q);
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
return 0;
}
static unsigned long db_bar_size(struct nvme_dev *dev, unsigned nr_io_queues)
{
return NVME_REG_DBS + ((nr_io_queues + 1) * 8 * dev->db_stride);
}
static int nvme_remap_bar(struct nvme_dev *dev, unsigned long size)
{
struct pci_dev *pdev = to_pci_dev(dev->dev);
if (size <= dev->bar_mapped_size)
return 0;
if (size > pci_resource_len(pdev, 0))
return -ENOMEM;
if (dev->bar)
iounmap(dev->bar);
dev->bar = ioremap(pci_resource_start(pdev, 0), size);
if (!dev->bar) {
dev->bar_mapped_size = 0;
return -ENOMEM;
}
dev->bar_mapped_size = size;
dev->dbs = dev->bar + NVME_REG_DBS;
return 0;
}
static int nvme_pci_configure_admin_queue(struct nvme_dev *dev)
{
int result;
u32 aqa;
struct nvme_queue *nvmeq;
result = nvme_remap_bar(dev, db_bar_size(dev, 0));
if (result < 0)
return result;
dev->subsystem = readl(dev->bar + NVME_REG_VS) >= NVME_VS(1, 1, 0) ?
NVME_CAP_NSSRC(dev->ctrl.cap) : 0;
if (dev->subsystem &&
(readl(dev->bar + NVME_REG_CSTS) & NVME_CSTS_NSSRO))
writel(NVME_CSTS_NSSRO, dev->bar + NVME_REG_CSTS);
result = nvme_disable_ctrl(&dev->ctrl, dev->ctrl.cap);
if (result < 0)
return result;
result = nvme_alloc_queue(dev, 0, NVME_AQ_DEPTH);
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
if (result)
return result;
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
nvmeq = &dev->queues[0];
aqa = nvmeq->q_depth - 1;
aqa |= aqa << 16;
writel(aqa, dev->bar + NVME_REG_AQA);
lo_hi_writeq(nvmeq->sq_dma_addr, dev->bar + NVME_REG_ASQ);
lo_hi_writeq(nvmeq->cq_dma_addr, dev->bar + NVME_REG_ACQ);
result = nvme_enable_ctrl(&dev->ctrl, dev->ctrl.cap);
if (result)
return result;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
nvmeq->cq_vector = 0;
nvme_init_queue(nvmeq, 0);
result = queue_request_irq(nvmeq);
if (result) {
nvmeq->cq_vector = -1;
return result;
}
return result;
}
static int nvme_create_io_queues(struct nvme_dev *dev)
{
unsigned i, max;
int ret = 0;
for (i = dev->ctrl.queue_count; i <= dev->max_qid; i++) {
if (nvme_alloc_queue(dev, i, dev->q_depth)) {
ret = -ENOMEM;
break;
}
}
max = min(dev->max_qid, dev->ctrl.queue_count - 1);
for (i = dev->online_queues; i <= max; i++) {
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
ret = nvme_create_queue(&dev->queues[i], i);
if (ret)
break;
}
/*
* Ignore failing Create SQ/CQ commands, we can continue with less
* than the desired amount of queues, and even a controller without
* I/O queues can still be used to issue admin commands. This might
* be useful to upgrade a buggy firmware for example.
*/
return ret >= 0 ? 0 : ret;
}
static ssize_t nvme_cmb_show(struct device *dev,
struct device_attribute *attr,
char *buf)
{
struct nvme_dev *ndev = to_nvme_dev(dev_get_drvdata(dev));
return scnprintf(buf, PAGE_SIZE, "cmbloc : x%08x\ncmbsz : x%08x\n",
ndev->cmbloc, ndev->cmbsz);
}
static DEVICE_ATTR(cmb, S_IRUGO, nvme_cmb_show, NULL);
static u64 nvme_cmb_size_unit(struct nvme_dev *dev)
{
u8 szu = (dev->cmbsz >> NVME_CMBSZ_SZU_SHIFT) & NVME_CMBSZ_SZU_MASK;
return 1ULL << (12 + 4 * szu);
}
static u32 nvme_cmb_size(struct nvme_dev *dev)
{
return (dev->cmbsz >> NVME_CMBSZ_SZ_SHIFT) & NVME_CMBSZ_SZ_MASK;
}
static void nvme_map_cmb(struct nvme_dev *dev)
{
u64 size, offset;
resource_size_t bar_size;
struct pci_dev *pdev = to_pci_dev(dev->dev);
int bar;
dev->cmbsz = readl(dev->bar + NVME_REG_CMBSZ);
if (!dev->cmbsz)
return;
dev->cmbloc = readl(dev->bar + NVME_REG_CMBLOC);
if (!use_cmb_sqes)
return;
size = nvme_cmb_size_unit(dev) * nvme_cmb_size(dev);
offset = nvme_cmb_size_unit(dev) * NVME_CMB_OFST(dev->cmbloc);
bar = NVME_CMB_BIR(dev->cmbloc);
bar_size = pci_resource_len(pdev, bar);
if (offset > bar_size)
return;
/*
* Controllers may support a CMB size larger than their BAR,
* for example, due to being behind a bridge. Reduce the CMB to
* the reported size of the BAR
*/
if (size > bar_size - offset)
size = bar_size - offset;
dev->cmb = ioremap_wc(pci_resource_start(pdev, bar) + offset, size);
if (!dev->cmb)
return;
dev->cmb_bus_addr = pci_bus_address(pdev, bar) + offset;
dev->cmb_size = size;
if (sysfs_add_file_to_group(&dev->ctrl.device->kobj,
&dev_attr_cmb.attr, NULL))
dev_warn(dev->ctrl.device,
"failed to add sysfs attribute for CMB\n");
}
static inline void nvme_release_cmb(struct nvme_dev *dev)
{
if (dev->cmb) {
iounmap(dev->cmb);
dev->cmb = NULL;
sysfs_remove_file_from_group(&dev->ctrl.device->kobj,
&dev_attr_cmb.attr, NULL);
dev->cmbsz = 0;
}
}
static int nvme_set_host_mem(struct nvme_dev *dev, u32 bits)
{
u64 dma_addr = dev->host_mem_descs_dma;
struct nvme_command c;
int ret;
memset(&c, 0, sizeof(c));
c.features.opcode = nvme_admin_set_features;
c.features.fid = cpu_to_le32(NVME_FEAT_HOST_MEM_BUF);
c.features.dword11 = cpu_to_le32(bits);
c.features.dword12 = cpu_to_le32(dev->host_mem_size >>
ilog2(dev->ctrl.page_size));
c.features.dword13 = cpu_to_le32(lower_32_bits(dma_addr));
c.features.dword14 = cpu_to_le32(upper_32_bits(dma_addr));
c.features.dword15 = cpu_to_le32(dev->nr_host_mem_descs);
ret = nvme_submit_sync_cmd(dev->ctrl.admin_q, &c, NULL, 0);
if (ret) {
dev_warn(dev->ctrl.device,
"failed to set host mem (err %d, flags %#x).\n",
ret, bits);
}
return ret;
}
static void nvme_free_host_mem(struct nvme_dev *dev)
{
int i;
for (i = 0; i < dev->nr_host_mem_descs; i++) {
struct nvme_host_mem_buf_desc *desc = &dev->host_mem_descs[i];
size_t size = le32_to_cpu(desc->size) * dev->ctrl.page_size;
dma_free_coherent(dev->dev, size, dev->host_mem_desc_bufs[i],
le64_to_cpu(desc->addr));
}
kfree(dev->host_mem_desc_bufs);
dev->host_mem_desc_bufs = NULL;
dma_free_coherent(dev->dev,
dev->nr_host_mem_descs * sizeof(*dev->host_mem_descs),
dev->host_mem_descs, dev->host_mem_descs_dma);
dev->host_mem_descs = NULL;
dev->nr_host_mem_descs = 0;
}
static int __nvme_alloc_host_mem(struct nvme_dev *dev, u64 preferred,
u32 chunk_size)
{
struct nvme_host_mem_buf_desc *descs;
u32 max_entries, len;
dma_addr_t descs_dma;
int i = 0;
void **bufs;
u64 size, tmp;
tmp = (preferred + chunk_size - 1);
do_div(tmp, chunk_size);
max_entries = tmp;
if (dev->ctrl.hmmaxd && dev->ctrl.hmmaxd < max_entries)
max_entries = dev->ctrl.hmmaxd;
descs = dma_zalloc_coherent(dev->dev, max_entries * sizeof(*descs),
&descs_dma, GFP_KERNEL);
if (!descs)
goto out;
bufs = kcalloc(max_entries, sizeof(*bufs), GFP_KERNEL);
if (!bufs)
goto out_free_descs;
for (size = 0; size < preferred && i < max_entries; size += len) {
dma_addr_t dma_addr;
len = min_t(u64, chunk_size, preferred - size);
bufs[i] = dma_alloc_attrs(dev->dev, len, &dma_addr, GFP_KERNEL,
DMA_ATTR_NO_KERNEL_MAPPING | DMA_ATTR_NO_WARN);
if (!bufs[i])
break;
descs[i].addr = cpu_to_le64(dma_addr);
descs[i].size = cpu_to_le32(len / dev->ctrl.page_size);
i++;
}
if (!size)
goto out_free_bufs;
dev->nr_host_mem_descs = i;
dev->host_mem_size = size;
dev->host_mem_descs = descs;
dev->host_mem_descs_dma = descs_dma;
dev->host_mem_desc_bufs = bufs;
return 0;
out_free_bufs:
while (--i >= 0) {
size_t size = le32_to_cpu(descs[i].size) * dev->ctrl.page_size;
dma_free_coherent(dev->dev, size, bufs[i],
le64_to_cpu(descs[i].addr));
}
kfree(bufs);
out_free_descs:
dma_free_coherent(dev->dev, max_entries * sizeof(*descs), descs,
descs_dma);
out:
dev->host_mem_descs = NULL;
return -ENOMEM;
}
static int nvme_alloc_host_mem(struct nvme_dev *dev, u64 min, u64 preferred)
{
u32 chunk_size;
/* start big and work our way down */
for (chunk_size = min_t(u64, preferred, PAGE_SIZE * MAX_ORDER_NR_PAGES);
chunk_size >= max_t(u32, dev->ctrl.hmminds * 4096, PAGE_SIZE * 2);
chunk_size /= 2) {
if (!__nvme_alloc_host_mem(dev, preferred, chunk_size)) {
if (!min || dev->host_mem_size >= min)
return 0;
nvme_free_host_mem(dev);
}
}
return -ENOMEM;
}
static int nvme_setup_host_mem(struct nvme_dev *dev)
{
u64 max = (u64)max_host_mem_size_mb * SZ_1M;
u64 preferred = (u64)dev->ctrl.hmpre * 4096;
u64 min = (u64)dev->ctrl.hmmin * 4096;
u32 enable_bits = NVME_HOST_MEM_ENABLE;
int ret;
preferred = min(preferred, max);
if (min > max) {
dev_warn(dev->ctrl.device,
"min host memory (%lld MiB) above limit (%d MiB).\n",
min >> ilog2(SZ_1M), max_host_mem_size_mb);
nvme_free_host_mem(dev);
return 0;
}
/*
* If we already have a buffer allocated check if we can reuse it.
*/
if (dev->host_mem_descs) {
if (dev->host_mem_size >= min)
enable_bits |= NVME_HOST_MEM_RETURN;
else
nvme_free_host_mem(dev);
}
if (!dev->host_mem_descs) {
if (nvme_alloc_host_mem(dev, min, preferred)) {
dev_warn(dev->ctrl.device,
"failed to allocate host memory buffer.\n");
return 0; /* controller must work without HMB */
}
dev_info(dev->ctrl.device,
"allocated %lld MiB host memory buffer.\n",
dev->host_mem_size >> ilog2(SZ_1M));
}
ret = nvme_set_host_mem(dev, enable_bits);
if (ret)
nvme_free_host_mem(dev);
return ret;
}
static int nvme_setup_io_queues(struct nvme_dev *dev)
{
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
struct nvme_queue *adminq = &dev->queues[0];
struct pci_dev *pdev = to_pci_dev(dev->dev);
int result, nr_io_queues;
unsigned long size;
struct irq_affinity affd = {
.pre_vectors = 1
};
nr_io_queues = num_possible_cpus();
result = nvme_set_queue_count(&dev->ctrl, &nr_io_queues);
if (result < 0)
return result;
if (nr_io_queues == 0)
return 0;
if (dev->cmb && (dev->cmbsz & NVME_CMBSZ_SQS)) {
result = nvme_cmb_qdepth(dev, nr_io_queues,
sizeof(struct nvme_command));
if (result > 0)
dev->q_depth = result;
else
nvme_release_cmb(dev);
}
do {
size = db_bar_size(dev, nr_io_queues);
result = nvme_remap_bar(dev, size);
if (!result)
break;
if (!--nr_io_queues)
return -ENOMEM;
} while (1);
adminq->q_db = dev->dbs;
/* Deregister the admin queue's interrupt */
pci_free_irq(pdev, 0, adminq);
/*
* If we enable msix early due to not intx, disable it again before
* setting up the full range we need.
*/
pci_free_irq_vectors(pdev);
result = pci_alloc_irq_vectors_affinity(pdev, 1, nr_io_queues + 1,
PCI_IRQ_ALL_TYPES | PCI_IRQ_AFFINITY, &affd);
if (result <= 0)
return -EIO;
dev->num_vecs = result;
dev->max_qid = max(result - 1, 1);
/*
* Should investigate if there's a performance win from allocating
* more queues than interrupt vectors; it might allow the submission
* path to scale better, even if the receive path is limited by the
* number of interrupts.
*/
result = queue_request_irq(adminq);
if (result) {
adminq->cq_vector = -1;
return result;
}
return nvme_create_io_queues(dev);
}
static void nvme_del_queue_end(struct request *req, blk_status_t error)
{
struct nvme_queue *nvmeq = req->end_io_data;
blk_mq_free_request(req);
complete(&nvmeq->dev->ioq_wait);
}
static void nvme_del_cq_end(struct request *req, blk_status_t error)
{
struct nvme_queue *nvmeq = req->end_io_data;
u16 start, end;
if (!error) {
unsigned long flags;
nvme: add missing lock nesting notation When unloading driver, nvme_disable_io_queues() calls nvme_delete_queue() that sends nvme_admin_delete_cq command to admin sq. So when the command completed, the lock acquired by nvme_irq() actually belongs to admin queue. While the lock that nvme_del_cq_end() trying to acquire belongs to io queue. So it will not deadlock. This patch adds lock nesting notation to fix following report. [ 109.840952] ============================================= [ 109.846379] [ INFO: possible recursive locking detected ] [ 109.851806] 4.5.0+ #180 Tainted: G E [ 109.856533] --------------------------------------------- [ 109.861958] swapper/0/0 is trying to acquire lock: [ 109.866771] (&(&nvmeq->q_lock)->rlock){-.....}, at: [<ffffffffc0820bc6>] nvme_del_cq_end+0x26/0x70 [nvme] [ 109.876535] [ 109.876535] but task is already holding lock: [ 109.882398] (&(&nvmeq->q_lock)->rlock){-.....}, at: [<ffffffffc0820c2b>] nvme_irq+0x1b/0x50 [nvme] [ 109.891547] [ 109.891547] other info that might help us debug this: [ 109.898107] Possible unsafe locking scenario: [ 109.898107] [ 109.904056] CPU0 [ 109.906515] ---- [ 109.908974] lock(&(&nvmeq->q_lock)->rlock); [ 109.913381] lock(&(&nvmeq->q_lock)->rlock); [ 109.917787] [ 109.917787] *** DEADLOCK *** [ 109.917787] [ 109.923738] May be due to missing lock nesting notation [ 109.923738] [ 109.930558] 1 lock held by swapper/0/0: [ 109.934413] #0: (&(&nvmeq->q_lock)->rlock){-.....}, at: [<ffffffffc0820c2b>] nvme_irq+0x1b/0x50 [nvme] [ 109.944010] [ 109.944010] stack backtrace: [ 109.948389] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G E 4.5.0+ #180 [ 109.955734] Hardware name: Dell Inc. OptiPlex 7010/0YXT71, BIOS A15 08/12/2013 [ 109.962989] 0000000000000000 ffff88011e203c38 ffffffff81383d9c ffffffff81c13540 [ 109.970478] ffffffff826711d0 ffff88011e203ce8 ffffffff810bb429 0000000000000046 [ 109.977964] 0000000000000046 0000000000000000 0000000000b2e597 ffffffff81f4cb00 [ 109.985453] Call Trace: [ 109.987911] <IRQ> [<ffffffff81383d9c>] dump_stack+0x85/0xc9 [ 109.993711] [<ffffffff810bb429>] __lock_acquire+0x19b9/0x1c60 [ 109.999575] [<ffffffff810b6d1d>] ? trace_hardirqs_off+0xd/0x10 [ 110.005524] [<ffffffff810b386d>] ? complete+0x3d/0x50 [ 110.010688] [<ffffffff810bb760>] lock_acquire+0x90/0xf0 [ 110.016029] [<ffffffffc0820bc6>] ? nvme_del_cq_end+0x26/0x70 [nvme] [ 110.022418] [<ffffffff81772afb>] _raw_spin_lock_irqsave+0x4b/0x60 [ 110.028632] [<ffffffffc0820bc6>] ? nvme_del_cq_end+0x26/0x70 [nvme] [ 110.035019] [<ffffffffc0820bc6>] nvme_del_cq_end+0x26/0x70 [nvme] [ 110.041232] [<ffffffff8135b485>] blk_mq_end_request+0x35/0x60 [ 110.047095] [<ffffffffc0821ad8>] nvme_complete_rq+0x68/0x190 [nvme] [ 110.053481] [<ffffffff8135b53f>] __blk_mq_complete_request+0x8f/0x130 [ 110.060043] [<ffffffff8135b611>] blk_mq_complete_request+0x31/0x40 [ 110.066343] [<ffffffffc08209e3>] __nvme_process_cq+0x83/0x240 [nvme] [ 110.072818] [<ffffffffc0820c35>] nvme_irq+0x25/0x50 [nvme] [ 110.078419] [<ffffffff810cdb66>] handle_irq_event_percpu+0x36/0x110 [ 110.084804] [<ffffffff810cdc77>] handle_irq_event+0x37/0x60 [ 110.090491] [<ffffffff810d0ea3>] handle_edge_irq+0x93/0x150 [ 110.096180] [<ffffffff81012306>] handle_irq+0xa6/0x130 [ 110.101431] [<ffffffff81011abe>] do_IRQ+0x5e/0x120 [ 110.106333] [<ffffffff8177384c>] common_interrupt+0x8c/0x8c Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de> Signed-off-by: Ming Lin <ming.l@ssi.samsung.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Jens Axboe <axboe@fb.com>
2016-04-05 20:32:04 +03:00
/*
* We might be called with the AQ cq_lock held
* and the I/O queue cq_lock should always
nvme: add missing lock nesting notation When unloading driver, nvme_disable_io_queues() calls nvme_delete_queue() that sends nvme_admin_delete_cq command to admin sq. So when the command completed, the lock acquired by nvme_irq() actually belongs to admin queue. While the lock that nvme_del_cq_end() trying to acquire belongs to io queue. So it will not deadlock. This patch adds lock nesting notation to fix following report. [ 109.840952] ============================================= [ 109.846379] [ INFO: possible recursive locking detected ] [ 109.851806] 4.5.0+ #180 Tainted: G E [ 109.856533] --------------------------------------------- [ 109.861958] swapper/0/0 is trying to acquire lock: [ 109.866771] (&(&nvmeq->q_lock)->rlock){-.....}, at: [<ffffffffc0820bc6>] nvme_del_cq_end+0x26/0x70 [nvme] [ 109.876535] [ 109.876535] but task is already holding lock: [ 109.882398] (&(&nvmeq->q_lock)->rlock){-.....}, at: [<ffffffffc0820c2b>] nvme_irq+0x1b/0x50 [nvme] [ 109.891547] [ 109.891547] other info that might help us debug this: [ 109.898107] Possible unsafe locking scenario: [ 109.898107] [ 109.904056] CPU0 [ 109.906515] ---- [ 109.908974] lock(&(&nvmeq->q_lock)->rlock); [ 109.913381] lock(&(&nvmeq->q_lock)->rlock); [ 109.917787] [ 109.917787] *** DEADLOCK *** [ 109.917787] [ 109.923738] May be due to missing lock nesting notation [ 109.923738] [ 109.930558] 1 lock held by swapper/0/0: [ 109.934413] #0: (&(&nvmeq->q_lock)->rlock){-.....}, at: [<ffffffffc0820c2b>] nvme_irq+0x1b/0x50 [nvme] [ 109.944010] [ 109.944010] stack backtrace: [ 109.948389] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G E 4.5.0+ #180 [ 109.955734] Hardware name: Dell Inc. OptiPlex 7010/0YXT71, BIOS A15 08/12/2013 [ 109.962989] 0000000000000000 ffff88011e203c38 ffffffff81383d9c ffffffff81c13540 [ 109.970478] ffffffff826711d0 ffff88011e203ce8 ffffffff810bb429 0000000000000046 [ 109.977964] 0000000000000046 0000000000000000 0000000000b2e597 ffffffff81f4cb00 [ 109.985453] Call Trace: [ 109.987911] <IRQ> [<ffffffff81383d9c>] dump_stack+0x85/0xc9 [ 109.993711] [<ffffffff810bb429>] __lock_acquire+0x19b9/0x1c60 [ 109.999575] [<ffffffff810b6d1d>] ? trace_hardirqs_off+0xd/0x10 [ 110.005524] [<ffffffff810b386d>] ? complete+0x3d/0x50 [ 110.010688] [<ffffffff810bb760>] lock_acquire+0x90/0xf0 [ 110.016029] [<ffffffffc0820bc6>] ? nvme_del_cq_end+0x26/0x70 [nvme] [ 110.022418] [<ffffffff81772afb>] _raw_spin_lock_irqsave+0x4b/0x60 [ 110.028632] [<ffffffffc0820bc6>] ? nvme_del_cq_end+0x26/0x70 [nvme] [ 110.035019] [<ffffffffc0820bc6>] nvme_del_cq_end+0x26/0x70 [nvme] [ 110.041232] [<ffffffff8135b485>] blk_mq_end_request+0x35/0x60 [ 110.047095] [<ffffffffc0821ad8>] nvme_complete_rq+0x68/0x190 [nvme] [ 110.053481] [<ffffffff8135b53f>] __blk_mq_complete_request+0x8f/0x130 [ 110.060043] [<ffffffff8135b611>] blk_mq_complete_request+0x31/0x40 [ 110.066343] [<ffffffffc08209e3>] __nvme_process_cq+0x83/0x240 [nvme] [ 110.072818] [<ffffffffc0820c35>] nvme_irq+0x25/0x50 [nvme] [ 110.078419] [<ffffffff810cdb66>] handle_irq_event_percpu+0x36/0x110 [ 110.084804] [<ffffffff810cdc77>] handle_irq_event+0x37/0x60 [ 110.090491] [<ffffffff810d0ea3>] handle_edge_irq+0x93/0x150 [ 110.096180] [<ffffffff81012306>] handle_irq+0xa6/0x130 [ 110.101431] [<ffffffff81011abe>] do_IRQ+0x5e/0x120 [ 110.106333] [<ffffffff8177384c>] common_interrupt+0x8c/0x8c Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de> Signed-off-by: Ming Lin <ming.l@ssi.samsung.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Jens Axboe <axboe@fb.com>
2016-04-05 20:32:04 +03:00
* nest inside the AQ one.
*/
spin_lock_irqsave_nested(&nvmeq->cq_lock, flags,
nvme: add missing lock nesting notation When unloading driver, nvme_disable_io_queues() calls nvme_delete_queue() that sends nvme_admin_delete_cq command to admin sq. So when the command completed, the lock acquired by nvme_irq() actually belongs to admin queue. While the lock that nvme_del_cq_end() trying to acquire belongs to io queue. So it will not deadlock. This patch adds lock nesting notation to fix following report. [ 109.840952] ============================================= [ 109.846379] [ INFO: possible recursive locking detected ] [ 109.851806] 4.5.0+ #180 Tainted: G E [ 109.856533] --------------------------------------------- [ 109.861958] swapper/0/0 is trying to acquire lock: [ 109.866771] (&(&nvmeq->q_lock)->rlock){-.....}, at: [<ffffffffc0820bc6>] nvme_del_cq_end+0x26/0x70 [nvme] [ 109.876535] [ 109.876535] but task is already holding lock: [ 109.882398] (&(&nvmeq->q_lock)->rlock){-.....}, at: [<ffffffffc0820c2b>] nvme_irq+0x1b/0x50 [nvme] [ 109.891547] [ 109.891547] other info that might help us debug this: [ 109.898107] Possible unsafe locking scenario: [ 109.898107] [ 109.904056] CPU0 [ 109.906515] ---- [ 109.908974] lock(&(&nvmeq->q_lock)->rlock); [ 109.913381] lock(&(&nvmeq->q_lock)->rlock); [ 109.917787] [ 109.917787] *** DEADLOCK *** [ 109.917787] [ 109.923738] May be due to missing lock nesting notation [ 109.923738] [ 109.930558] 1 lock held by swapper/0/0: [ 109.934413] #0: (&(&nvmeq->q_lock)->rlock){-.....}, at: [<ffffffffc0820c2b>] nvme_irq+0x1b/0x50 [nvme] [ 109.944010] [ 109.944010] stack backtrace: [ 109.948389] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G E 4.5.0+ #180 [ 109.955734] Hardware name: Dell Inc. OptiPlex 7010/0YXT71, BIOS A15 08/12/2013 [ 109.962989] 0000000000000000 ffff88011e203c38 ffffffff81383d9c ffffffff81c13540 [ 109.970478] ffffffff826711d0 ffff88011e203ce8 ffffffff810bb429 0000000000000046 [ 109.977964] 0000000000000046 0000000000000000 0000000000b2e597 ffffffff81f4cb00 [ 109.985453] Call Trace: [ 109.987911] <IRQ> [<ffffffff81383d9c>] dump_stack+0x85/0xc9 [ 109.993711] [<ffffffff810bb429>] __lock_acquire+0x19b9/0x1c60 [ 109.999575] [<ffffffff810b6d1d>] ? trace_hardirqs_off+0xd/0x10 [ 110.005524] [<ffffffff810b386d>] ? complete+0x3d/0x50 [ 110.010688] [<ffffffff810bb760>] lock_acquire+0x90/0xf0 [ 110.016029] [<ffffffffc0820bc6>] ? nvme_del_cq_end+0x26/0x70 [nvme] [ 110.022418] [<ffffffff81772afb>] _raw_spin_lock_irqsave+0x4b/0x60 [ 110.028632] [<ffffffffc0820bc6>] ? nvme_del_cq_end+0x26/0x70 [nvme] [ 110.035019] [<ffffffffc0820bc6>] nvme_del_cq_end+0x26/0x70 [nvme] [ 110.041232] [<ffffffff8135b485>] blk_mq_end_request+0x35/0x60 [ 110.047095] [<ffffffffc0821ad8>] nvme_complete_rq+0x68/0x190 [nvme] [ 110.053481] [<ffffffff8135b53f>] __blk_mq_complete_request+0x8f/0x130 [ 110.060043] [<ffffffff8135b611>] blk_mq_complete_request+0x31/0x40 [ 110.066343] [<ffffffffc08209e3>] __nvme_process_cq+0x83/0x240 [nvme] [ 110.072818] [<ffffffffc0820c35>] nvme_irq+0x25/0x50 [nvme] [ 110.078419] [<ffffffff810cdb66>] handle_irq_event_percpu+0x36/0x110 [ 110.084804] [<ffffffff810cdc77>] handle_irq_event+0x37/0x60 [ 110.090491] [<ffffffff810d0ea3>] handle_edge_irq+0x93/0x150 [ 110.096180] [<ffffffff81012306>] handle_irq+0xa6/0x130 [ 110.101431] [<ffffffff81011abe>] do_IRQ+0x5e/0x120 [ 110.106333] [<ffffffff8177384c>] common_interrupt+0x8c/0x8c Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de> Signed-off-by: Ming Lin <ming.l@ssi.samsung.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Jens Axboe <axboe@fb.com>
2016-04-05 20:32:04 +03:00
SINGLE_DEPTH_NESTING);
nvme_process_cq(nvmeq, &start, &end, -1);
spin_unlock_irqrestore(&nvmeq->cq_lock, flags);
nvme_complete_cqes(nvmeq, start, end);
}
nvme_del_queue_end(req, error);
}
static int nvme_delete_queue(struct nvme_queue *nvmeq, u8 opcode)
{
struct request_queue *q = nvmeq->dev->ctrl.admin_q;
struct request *req;
struct nvme_command cmd;
memset(&cmd, 0, sizeof(cmd));
cmd.delete_queue.opcode = opcode;
cmd.delete_queue.qid = cpu_to_le16(nvmeq->qid);
req = nvme_alloc_request(q, &cmd, BLK_MQ_REQ_NOWAIT, NVME_QID_ANY);
if (IS_ERR(req))
return PTR_ERR(req);
req->timeout = ADMIN_TIMEOUT;
req->end_io_data = nvmeq;
blk_execute_rq_nowait(q, NULL, req, false,
opcode == nvme_admin_delete_cq ?
nvme_del_cq_end : nvme_del_queue_end);
return 0;
}
static void nvme_disable_io_queues(struct nvme_dev *dev)
{
int pass, queues = dev->online_queues - 1;
unsigned long timeout;
u8 opcode = nvme_admin_delete_sq;
for (pass = 0; pass < 2; pass++) {
int sent = 0, i = queues;
reinit_completion(&dev->ioq_wait);
retry:
timeout = ADMIN_TIMEOUT;
nvme: Suspend all queues before deletion When nvme_delete_queue fails in the first pass of the nvme_disable_io_queues() loop, we return early, failing to suspend all of the IO queues. Later, on the nvme_pci_disable path, this causes us to disable MSI without actually having freed all the IRQs, which triggers the BUG_ON in free_msi_irqs(), as show below. This patch refactors nvme_disable_io_queues to suspend all queues before start submitting delete queue commands. This way, we ensure that we have at least returned every IRQ before continuing with the removal path. [ 487.529200] kernel BUG at ../drivers/pci/msi.c:368! cpu 0x46: Vector: 700 (Program Check) at [c0000078c5b83650] pc: c000000000627a50: free_msi_irqs+0x90/0x200 lr: c000000000627a40: free_msi_irqs+0x80/0x200 sp: c0000078c5b838d0 msr: 9000000100029033 current = 0xc0000078c5b40000 paca = 0xc000000002bd7600 softe: 0 irq_happened: 0x01 pid = 1376, comm = kworker/70:1H kernel BUG at ../drivers/pci/msi.c:368! Linux version 4.7.0.mainline+ (root@iod76) (gcc version 5.3.1 20160413 (Ubuntu/IBM 5.3.1-14ubuntu2.1) ) #104 SMP Fri Jul 29 09:20:17 CDT 2016 enter ? for help [c0000078c5b83920] d0000000363b0cd8 nvme_dev_disable+0x208/0x4f0 [nvme] [c0000078c5b83a10] d0000000363b12a4 nvme_timeout+0xe4/0x250 [nvme] [c0000078c5b83ad0] c0000000005690e4 blk_mq_rq_timed_out+0x64/0x110 [c0000078c5b83b40] c00000000056c930 bt_for_each+0x160/0x170 [c0000078c5b83bb0] c00000000056d928 blk_mq_queue_tag_busy_iter+0x78/0x110 [c0000078c5b83c00] c0000000005675d8 blk_mq_timeout_work+0xd8/0x1b0 [c0000078c5b83c50] c0000000000e8cf0 process_one_work+0x1e0/0x590 [c0000078c5b83ce0] c0000000000e9148 worker_thread+0xa8/0x660 [c0000078c5b83d80] c0000000000f2090 kthread+0x110/0x130 [c0000078c5b83e30] c0000000000095f0 ret_from_kernel_thread+0x5c/0x6c Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com> Cc: Brian King <brking@linux.vnet.ibm.com> Cc: Keith Busch <keith.busch@intel.com> Cc: linux-nvme@lists.infradead.org Signed-off-by: Jens Axboe <axboe@fb.com>
2016-08-11 18:35:57 +03:00
for (; i > 0; i--, sent++)
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
if (nvme_delete_queue(&dev->queues[i], opcode))
break;
nvme: Suspend all queues before deletion When nvme_delete_queue fails in the first pass of the nvme_disable_io_queues() loop, we return early, failing to suspend all of the IO queues. Later, on the nvme_pci_disable path, this causes us to disable MSI without actually having freed all the IRQs, which triggers the BUG_ON in free_msi_irqs(), as show below. This patch refactors nvme_disable_io_queues to suspend all queues before start submitting delete queue commands. This way, we ensure that we have at least returned every IRQ before continuing with the removal path. [ 487.529200] kernel BUG at ../drivers/pci/msi.c:368! cpu 0x46: Vector: 700 (Program Check) at [c0000078c5b83650] pc: c000000000627a50: free_msi_irqs+0x90/0x200 lr: c000000000627a40: free_msi_irqs+0x80/0x200 sp: c0000078c5b838d0 msr: 9000000100029033 current = 0xc0000078c5b40000 paca = 0xc000000002bd7600 softe: 0 irq_happened: 0x01 pid = 1376, comm = kworker/70:1H kernel BUG at ../drivers/pci/msi.c:368! Linux version 4.7.0.mainline+ (root@iod76) (gcc version 5.3.1 20160413 (Ubuntu/IBM 5.3.1-14ubuntu2.1) ) #104 SMP Fri Jul 29 09:20:17 CDT 2016 enter ? for help [c0000078c5b83920] d0000000363b0cd8 nvme_dev_disable+0x208/0x4f0 [nvme] [c0000078c5b83a10] d0000000363b12a4 nvme_timeout+0xe4/0x250 [nvme] [c0000078c5b83ad0] c0000000005690e4 blk_mq_rq_timed_out+0x64/0x110 [c0000078c5b83b40] c00000000056c930 bt_for_each+0x160/0x170 [c0000078c5b83bb0] c00000000056d928 blk_mq_queue_tag_busy_iter+0x78/0x110 [c0000078c5b83c00] c0000000005675d8 blk_mq_timeout_work+0xd8/0x1b0 [c0000078c5b83c50] c0000000000e8cf0 process_one_work+0x1e0/0x590 [c0000078c5b83ce0] c0000000000e9148 worker_thread+0xa8/0x660 [c0000078c5b83d80] c0000000000f2090 kthread+0x110/0x130 [c0000078c5b83e30] c0000000000095f0 ret_from_kernel_thread+0x5c/0x6c Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com> Cc: Brian King <brking@linux.vnet.ibm.com> Cc: Keith Busch <keith.busch@intel.com> Cc: linux-nvme@lists.infradead.org Signed-off-by: Jens Axboe <axboe@fb.com>
2016-08-11 18:35:57 +03:00
while (sent--) {
timeout = wait_for_completion_io_timeout(&dev->ioq_wait, timeout);
if (timeout == 0)
return;
if (i)
goto retry;
}
opcode = nvme_admin_delete_cq;
}
}
/*
* return error value only when tagset allocation failed
*/
static int nvme_dev_add(struct nvme_dev *dev)
{
int ret;
if (!dev->ctrl.tagset) {
dev->tagset.ops = &nvme_mq_ops;
dev->tagset.nr_hw_queues = dev->online_queues - 1;
dev->tagset.timeout = NVME_IO_TIMEOUT;
dev->tagset.numa_node = dev_to_node(dev->dev);
dev->tagset.queue_depth =
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
min_t(int, dev->q_depth, BLK_MQ_MAX_DEPTH) - 1;
dev->tagset.cmd_size = nvme_pci_cmd_size(dev, false);
if ((dev->ctrl.sgls & ((1 << 0) | (1 << 1))) && sgl_threshold) {
dev->tagset.cmd_size = max(dev->tagset.cmd_size,
nvme_pci_cmd_size(dev, true));
}
dev->tagset.flags = BLK_MQ_F_SHOULD_MERGE;
dev->tagset.driver_data = dev;
ret = blk_mq_alloc_tag_set(&dev->tagset);
if (ret) {
dev_warn(dev->ctrl.device,
"IO queues tagset allocation failed %d\n", ret);
return ret;
}
dev->ctrl.tagset = &dev->tagset;
nvme: improve performance for virtual NVMe devices This change provides a mechanism to reduce the number of MMIO doorbell writes for the NVMe driver. When running in a virtualized environment like QEMU, the cost of an MMIO is quite hefy here. The main idea for the patch is provide the device two memory location locations: 1) to store the doorbell values so they can be lookup without the doorbell MMIO write 2) to store an event index. I believe the doorbell value is obvious, the event index not so much. Similar to the virtio specification, the virtual device can tell the driver (guest OS) not to write MMIO unless you are writing past this value. FYI: doorbell values are written by the nvme driver (guest OS) and the event index is written by the virtual device (host OS). The patch implements a new admin command that will communicate where these two memory locations reside. If the command fails, the nvme driver will work as before without any optimizations. Contributions: Eric Northup <digitaleric@google.com> Frank Swiderski <fes@google.com> Ted Tso <tytso@mit.edu> Keith Busch <keith.busch@intel.com> Just to give an idea on the performance boost with the vendor extension: Running fio [1], a stock NVMe driver I get about 200K read IOPs with my vendor patch I get about 1000K read IOPs. This was running with a null device i.e. the backing device simply returned success on every read IO request. [1] Running on a 4 core machine: fio --time_based --name=benchmark --runtime=30 --filename=/dev/nvme0n1 --nrfiles=1 --ioengine=libaio --iodepth=32 --direct=1 --invalidate=1 --verify=0 --verify_fatal=0 --numjobs=4 --rw=randread --blocksize=4k --randrepeat=false Signed-off-by: Rob Nelson <rlnelson@google.com> [mlin: port for upstream] Signed-off-by: Ming Lin <mlin@kernel.org> [koike: updated for upstream] Signed-off-by: Helen Koike <helen.koike@collabora.co.uk> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <keith.busch@intel.com>
2017-04-10 18:51:07 +03:00
nvme_dbbuf_set(dev);
} else {
blk_mq_update_nr_hw_queues(&dev->tagset, dev->online_queues - 1);
/* Free previously allocated queues that are no longer usable */
nvme_free_queues(dev, dev->online_queues);
}
NVMe: Metadata format support Adds support for NVMe metadata formats and exposes block devices for all namespaces regardless of their format. Namespace formats that are unusable will have disk capacity set to 0, but a handle to the block device is created to simplify device management. A namespace is not usable when the format requires host interleave block and metadata in single buffer, has no provisioned storage, or has better data but failed to register with blk integrity. The namespace has to be scanned in two phases to support separate metadata formats. The first establishes the sector size and capacity prior to invoking add_disk. If metadata is required, the capacity will be temporarilly set to 0 until it can be revalidated and registered with the integrity extenstions after add_disk completes. The driver relies on the integrity extensions to provide the metadata buffer. NVMe requires this be a single physically contiguous region, so only one integrity segment is allowed per command. If the metadata is used for T10 PI, the driver provides mappings to save and restore the reftag physical block translation. The driver provides no-op functions for generate and verify if metadata is not used for protection information. This way the setup is always provided by the block layer. If a request does not supply a required metadata buffer, the command is failed with bad address. This could only happen if a user manually disables verify/generate on such a disk. The only exception to where this is okay is if the controller is capable of stripping/generating the metadata, which is possible on some types of formats. The metadata scatter gather list now occupies the spot in the nvme_iod that used to be used to link retryable IOD's, but we don't do that anymore, so the field was unused. Signed-off-by: Keith Busch <keith.busch@intel.com>
2015-02-19 23:39:03 +03:00
return 0;
}
static int nvme_pci_enable(struct nvme_dev *dev)
{
int result = -ENOMEM;
struct pci_dev *pdev = to_pci_dev(dev->dev);
if (pci_enable_device_mem(pdev))
return result;
pci_set_master(pdev);
if (dma_set_mask_and_coherent(dev->dev, DMA_BIT_MASK(64)) &&
dma_set_mask_and_coherent(dev->dev, DMA_BIT_MASK(32)))
goto disable;
if (readl(dev->bar + NVME_REG_CSTS) == -1) {
result = -ENODEV;
goto disable;
}
/*
* Some devices and/or platforms don't advertise or work with INTx
* interrupts. Pre-enable a single MSIX or MSI vec for setup. We'll
* adjust this later.
*/
result = pci_alloc_irq_vectors(pdev, 1, 1, PCI_IRQ_ALL_TYPES);
if (result < 0)
return result;
dev->ctrl.cap = lo_hi_readq(dev->bar + NVME_REG_CAP);
dev->q_depth = min_t(int, NVME_CAP_MQES(dev->ctrl.cap) + 1,
io_queue_depth);
dev->db_stride = 1 << NVME_CAP_STRIDE(dev->ctrl.cap);
dev->dbs = dev->bar + 4096;
/*
* Temporary fix for the Apple controller found in the MacBook8,1 and
* some MacBook7,1 to avoid controller resets and data loss.
*/
if (pdev->vendor == PCI_VENDOR_ID_APPLE && pdev->device == 0x2001) {
dev->q_depth = 2;
dev_warn(dev->ctrl.device, "detected Apple NVMe controller, "
"set queue depth=%u to work around controller resets\n",
dev->q_depth);
} else if (pdev->vendor == PCI_VENDOR_ID_SAMSUNG &&
(pdev->device == 0xa821 || pdev->device == 0xa822) &&
NVME_CAP_MQES(dev->ctrl.cap) == 0) {
dev->q_depth = 64;
dev_err(dev->ctrl.device, "detected PM1725 NVMe controller, "
"set queue depth=%u\n", dev->q_depth);
}
nvme_map_cmb(dev);
pci_enable_pcie_error_reporting(pdev);
pci_save_state(pdev);
return 0;
disable:
pci_disable_device(pdev);
return result;
}
static void nvme_dev_unmap(struct nvme_dev *dev)
{
if (dev->bar)
iounmap(dev->bar);
pci_release_mem_regions(to_pci_dev(dev->dev));
}
static void nvme_pci_disable(struct nvme_dev *dev)
{
struct pci_dev *pdev = to_pci_dev(dev->dev);
nvme_release_cmb(dev);
pci_free_irq_vectors(pdev);
if (pci_is_enabled(pdev)) {
pci_disable_pcie_error_reporting(pdev);
pci_disable_device(pdev);
}
}
static void nvme_dev_disable(struct nvme_dev *dev, bool shutdown)
{
int i;
bool dead = true;
struct pci_dev *pdev = to_pci_dev(dev->dev);
mutex_lock(&dev->shutdown_lock);
if (pci_is_enabled(pdev)) {
u32 csts = readl(dev->bar + NVME_REG_CSTS);
if (dev->ctrl.state == NVME_CTRL_LIVE ||
dev->ctrl.state == NVME_CTRL_RESETTING)
nvme_start_freeze(&dev->ctrl);
dead = !!((csts & NVME_CSTS_CFS) || !(csts & NVME_CSTS_RDY) ||
pdev->error_state != pci_channel_io_normal);
}
nvme: Suspend all queues before deletion When nvme_delete_queue fails in the first pass of the nvme_disable_io_queues() loop, we return early, failing to suspend all of the IO queues. Later, on the nvme_pci_disable path, this causes us to disable MSI without actually having freed all the IRQs, which triggers the BUG_ON in free_msi_irqs(), as show below. This patch refactors nvme_disable_io_queues to suspend all queues before start submitting delete queue commands. This way, we ensure that we have at least returned every IRQ before continuing with the removal path. [ 487.529200] kernel BUG at ../drivers/pci/msi.c:368! cpu 0x46: Vector: 700 (Program Check) at [c0000078c5b83650] pc: c000000000627a50: free_msi_irqs+0x90/0x200 lr: c000000000627a40: free_msi_irqs+0x80/0x200 sp: c0000078c5b838d0 msr: 9000000100029033 current = 0xc0000078c5b40000 paca = 0xc000000002bd7600 softe: 0 irq_happened: 0x01 pid = 1376, comm = kworker/70:1H kernel BUG at ../drivers/pci/msi.c:368! Linux version 4.7.0.mainline+ (root@iod76) (gcc version 5.3.1 20160413 (Ubuntu/IBM 5.3.1-14ubuntu2.1) ) #104 SMP Fri Jul 29 09:20:17 CDT 2016 enter ? for help [c0000078c5b83920] d0000000363b0cd8 nvme_dev_disable+0x208/0x4f0 [nvme] [c0000078c5b83a10] d0000000363b12a4 nvme_timeout+0xe4/0x250 [nvme] [c0000078c5b83ad0] c0000000005690e4 blk_mq_rq_timed_out+0x64/0x110 [c0000078c5b83b40] c00000000056c930 bt_for_each+0x160/0x170 [c0000078c5b83bb0] c00000000056d928 blk_mq_queue_tag_busy_iter+0x78/0x110 [c0000078c5b83c00] c0000000005675d8 blk_mq_timeout_work+0xd8/0x1b0 [c0000078c5b83c50] c0000000000e8cf0 process_one_work+0x1e0/0x590 [c0000078c5b83ce0] c0000000000e9148 worker_thread+0xa8/0x660 [c0000078c5b83d80] c0000000000f2090 kthread+0x110/0x130 [c0000078c5b83e30] c0000000000095f0 ret_from_kernel_thread+0x5c/0x6c Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com> Cc: Brian King <brking@linux.vnet.ibm.com> Cc: Keith Busch <keith.busch@intel.com> Cc: linux-nvme@lists.infradead.org Signed-off-by: Jens Axboe <axboe@fb.com>
2016-08-11 18:35:57 +03:00
/*
* Give the controller a chance to complete all entered requests if
* doing a safe shutdown.
*/
if (!dead) {
if (shutdown)
nvme_wait_freeze_timeout(&dev->ctrl, NVME_IO_TIMEOUT);
}
nvme_stop_queues(&dev->ctrl);
if (!dead && dev->ctrl.queue_count > 0) {
/*
* If the controller is still alive tell it to stop using the
* host memory buffer. In theory the shutdown / reset should
* make sure that it doesn't access the host memoery anymore,
* but I'd rather be safe than sorry..
*/
if (dev->host_mem_descs)
nvme_set_host_mem(dev, 0);
nvme_disable_io_queues(dev);
nvme_disable_admin_queue(dev, shutdown);
}
for (i = dev->ctrl.queue_count - 1; i >= 0; i--)
nvme_suspend_queue(&dev->queues[i]);
nvme_pci_disable(dev);
blk_mq_tagset_busy_iter(&dev->tagset, nvme_cancel_request, &dev->ctrl);
blk_mq_tagset_busy_iter(&dev->admin_tagset, nvme_cancel_request, &dev->ctrl);
/*
* The driver will not be starting up queues again if shutting down so
* must flush all entered requests to their failed completion to avoid
* deadlocking blk-mq hot-cpu notifier.
*/
if (shutdown)
nvme_start_queues(&dev->ctrl);
mutex_unlock(&dev->shutdown_lock);
}
static int nvme_setup_prp_pools(struct nvme_dev *dev)
{
dev->prp_page_pool = dma_pool_create("prp list page", dev->dev,
PAGE_SIZE, PAGE_SIZE, 0);
if (!dev->prp_page_pool)
return -ENOMEM;
/* Optimisation for I/Os between 4k and 128k */
dev->prp_small_pool = dma_pool_create("prp list 256", dev->dev,
256, 256, 0);
if (!dev->prp_small_pool) {
dma_pool_destroy(dev->prp_page_pool);
return -ENOMEM;
}
return 0;
}
static void nvme_release_prp_pools(struct nvme_dev *dev)
{
dma_pool_destroy(dev->prp_page_pool);
dma_pool_destroy(dev->prp_small_pool);
}
static void nvme_pci_free_ctrl(struct nvme_ctrl *ctrl)
{
struct nvme_dev *dev = to_nvme_dev(ctrl);
nvme: improve performance for virtual NVMe devices This change provides a mechanism to reduce the number of MMIO doorbell writes for the NVMe driver. When running in a virtualized environment like QEMU, the cost of an MMIO is quite hefy here. The main idea for the patch is provide the device two memory location locations: 1) to store the doorbell values so they can be lookup without the doorbell MMIO write 2) to store an event index. I believe the doorbell value is obvious, the event index not so much. Similar to the virtio specification, the virtual device can tell the driver (guest OS) not to write MMIO unless you are writing past this value. FYI: doorbell values are written by the nvme driver (guest OS) and the event index is written by the virtual device (host OS). The patch implements a new admin command that will communicate where these two memory locations reside. If the command fails, the nvme driver will work as before without any optimizations. Contributions: Eric Northup <digitaleric@google.com> Frank Swiderski <fes@google.com> Ted Tso <tytso@mit.edu> Keith Busch <keith.busch@intel.com> Just to give an idea on the performance boost with the vendor extension: Running fio [1], a stock NVMe driver I get about 200K read IOPs with my vendor patch I get about 1000K read IOPs. This was running with a null device i.e. the backing device simply returned success on every read IO request. [1] Running on a 4 core machine: fio --time_based --name=benchmark --runtime=30 --filename=/dev/nvme0n1 --nrfiles=1 --ioengine=libaio --iodepth=32 --direct=1 --invalidate=1 --verify=0 --verify_fatal=0 --numjobs=4 --rw=randread --blocksize=4k --randrepeat=false Signed-off-by: Rob Nelson <rlnelson@google.com> [mlin: port for upstream] Signed-off-by: Ming Lin <mlin@kernel.org> [koike: updated for upstream] Signed-off-by: Helen Koike <helen.koike@collabora.co.uk> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <keith.busch@intel.com>
2017-04-10 18:51:07 +03:00
nvme_dbbuf_dma_free(dev);
put_device(dev->dev);
if (dev->tagset.tags)
blk_mq_free_tag_set(&dev->tagset);
if (dev->ctrl.admin_q)
blk_put_queue(dev->ctrl.admin_q);
kfree(dev->queues);
free_opal_dev(dev->ctrl.opal_dev);
kfree(dev);
}
static void nvme_remove_dead_ctrl(struct nvme_dev *dev, int status)
{
Merge branch 'for-4.6/drivers' of git://git.kernel.dk/linux-block Pull block driver updates from Jens Axboe: "This is the block driver pull request for this merge window. It sits on top of for-4.6/core, that was just sent out. This contains: - A set of fixes for lightnvm. One from Alan, fixing an overflow, and the rest from the usual suspects, Javier and Matias. - A set of fixes for nbd from Markus and Dan, and a fixup from Arnd for correct usage of the signed 64-bit divider. - A set of bug fixes for the Micron mtip32xx, from Asai. - A fix for the brd discard handling from Bart. - Update the maintainers entry for cciss, since that hardware has transferred ownership. - Three bug fixes for bcache from Eric Wheeler. - Set of fixes for xen-blk{back,front} from Jan and Konrad. - Removal of the cpqarray driver. It has been disabled in Kconfig since 2013, and we were initially scheduled to remove it in 3.15. - Various updates and fixes for NVMe, with the most important being: - Removal of the per-device NVMe thread, replacing that with a watchdog timer instead. From Christoph. - Exposing the namespace WWID through sysfs, from Keith. - Set of cleanups from Ming Lin. - Logging the controller device name instead of the underlying PCI device name, from Sagi. - And a bunch of fixes and optimizations from the usual suspects in this area" * 'for-4.6/drivers' of git://git.kernel.dk/linux-block: (49 commits) NVMe: Expose ns wwid through single sysfs entry drivers:block: cpqarray clean up brd: Fix discard request processing cpqarray: remove it from the kernel cciss: update MAINTAINERS NVMe: Remove unused sq_head read in completion path bcache: fix cache_set_flush() NULL pointer dereference on OOM bcache: cleaned up error handling around register_cache() bcache: fix race of writeback thread starting before complete initialization NVMe: Create discard zero quirk white list nbd: use correct div_s64 helper mtip32xx: remove unneeded variable in mtip_cmd_timeout() lightnvm: generalize rrpc ppa calculations lightnvm: remove struct nvm_dev->total_blocks lightnvm: rename ->nr_pages to ->nr_sects lightnvm: update closed list outside of intr context xen/blback: Fit the important information of the thread in 17 characters lightnvm: fold get bb tbl when using dual/quad plane mode lightnvm: fix up nonsensical configure overrun checking xen-blkback: advertise indirect segment support earlier ...
2016-03-19 03:13:31 +03:00
dev_warn(dev->ctrl.device, "Removing after probe failure status: %d\n", status);
nvme_get_ctrl(&dev->ctrl);
nvme_dev_disable(dev, false);
if (!queue_work(nvme_wq, &dev->remove_work))
nvme_put_ctrl(&dev->ctrl);
}
static void nvme_reset_work(struct work_struct *work)
{
struct nvme_dev *dev =
container_of(work, struct nvme_dev, ctrl.reset_work);
bool was_suspend = !!(dev->ctrl.ctrl_config & NVME_CC_SHN_NORMAL);
int result = -ENODEV;
enum nvme_ctrl_state new_state = NVME_CTRL_LIVE;
nvme-pci: fix multiple ctrl removal scheduling Commit c5f6ce97c1210 tries to address multiple resets but fails as work_busy doesn't involve any synchronization and can fail. This is reproducible easily as can be seen by WARNING below which is triggered with line: WARN_ON(dev->ctrl.state == NVME_CTRL_RESETTING) Allowing multiple resets can result in multiple controller removal as well if different conditions inside nvme_reset_work fail and which might deadlock on device_release_driver. [ 480.327007] WARNING: CPU: 3 PID: 150 at drivers/nvme/host/pci.c:1900 nvme_reset_work+0x36c/0xec0 [ 480.327008] Modules linked in: rfcomm fuse nf_conntrack_netbios_ns nf_conntrack_broadcast... [ 480.327044] btusb videobuf2_core ghash_clmulni_intel snd_hwdep cfg80211 acer_wmi hci_uart.. [ 480.327065] CPU: 3 PID: 150 Comm: kworker/u16:2 Not tainted 4.12.0-rc1+ #13 [ 480.327065] Hardware name: Acer Predator G9-591/Mustang_SLS, BIOS V1.10 03/03/2016 [ 480.327066] Workqueue: nvme nvme_reset_work [ 480.327067] task: ffff880498ad8000 task.stack: ffffc90002218000 [ 480.327068] RIP: 0010:nvme_reset_work+0x36c/0xec0 [ 480.327069] RSP: 0018:ffffc9000221bdb8 EFLAGS: 00010246 [ 480.327070] RAX: 0000000000460000 RBX: ffff880498a98128 RCX: dead000000000200 [ 480.327070] RDX: 0000000000000001 RSI: ffff8804b1028020 RDI: ffff880498a98128 [ 480.327071] RBP: ffffc9000221be50 R08: 0000000000000000 R09: 0000000000000000 [ 480.327071] R10: ffffc90001963ce8 R11: 000000000000020d R12: ffff880498a98000 [ 480.327072] R13: ffff880498a53500 R14: ffff880498a98130 R15: ffff880498a98128 [ 480.327072] FS: 0000000000000000(0000) GS:ffff8804c1cc0000(0000) knlGS:0000000000000000 [ 480.327073] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 480.327074] CR2: 00007ffcf3c37f78 CR3: 0000000001e09000 CR4: 00000000003406e0 [ 480.327074] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 480.327075] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 [ 480.327075] Call Trace: [ 480.327079] ? __switch_to+0x227/0x400 [ 480.327081] process_one_work+0x18c/0x3a0 [ 480.327082] worker_thread+0x4e/0x3b0 [ 480.327084] kthread+0x109/0x140 [ 480.327085] ? process_one_work+0x3a0/0x3a0 [ 480.327087] ? kthread_park+0x60/0x60 [ 480.327102] ret_from_fork+0x2c/0x40 [ 480.327103] Code: e8 5a dc ff ff 85 c0 41 89 c1 0f..... This patch addresses the problem by using state of controller to decide whether reset should be queued or not as state change is synchronizated using controller spinlock. Also cancel_work_sync is used to make sure remove cancels the reset_work and waits for it to finish. This patch also changes return value from -ENODEV to more appropriate -EBUSY if nvme_reset fails to change state. Fixes: c5f6ce97c1210 ("nvme: don't schedule multiple resets") Signed-off-by: Rakesh Pandit <rakesh@tuxera.com> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2017-06-05 14:43:11 +03:00
if (WARN_ON(dev->ctrl.state != NVME_CTRL_RESETTING))
goto out;
/*
* If we're called to reset a live controller first shut it down before
* moving on.
*/
if (dev->ctrl.ctrl_config & NVME_CC_ENABLE)
nvme_dev_disable(dev, false);
/*
* Introduce CONNECTING state from nvme-fc/rdma transports to mark the
* initializing procedure here.
*/
if (!nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_CONNECTING)) {
dev_warn(dev->ctrl.device,
"failed to mark controller CONNECTING\n");
goto out;
}
result = nvme_pci_enable(dev);
if (result)
goto out;
result = nvme_pci_configure_admin_queue(dev);
if (result)
goto out;
result = nvme_alloc_admin_tags(dev);
if (result)
goto out;
result = nvme_init_identify(&dev->ctrl);
if (result)
goto out;
if (dev->ctrl.oacs & NVME_CTRL_OACS_SEC_SUPP) {
if (!dev->ctrl.opal_dev)
dev->ctrl.opal_dev =
init_opal_dev(&dev->ctrl, &nvme_sec_submit);
else if (was_suspend)
opal_unlock_from_suspend(dev->ctrl.opal_dev);
} else {
free_opal_dev(dev->ctrl.opal_dev);
dev->ctrl.opal_dev = NULL;
}
nvme: improve performance for virtual NVMe devices This change provides a mechanism to reduce the number of MMIO doorbell writes for the NVMe driver. When running in a virtualized environment like QEMU, the cost of an MMIO is quite hefy here. The main idea for the patch is provide the device two memory location locations: 1) to store the doorbell values so they can be lookup without the doorbell MMIO write 2) to store an event index. I believe the doorbell value is obvious, the event index not so much. Similar to the virtio specification, the virtual device can tell the driver (guest OS) not to write MMIO unless you are writing past this value. FYI: doorbell values are written by the nvme driver (guest OS) and the event index is written by the virtual device (host OS). The patch implements a new admin command that will communicate where these two memory locations reside. If the command fails, the nvme driver will work as before without any optimizations. Contributions: Eric Northup <digitaleric@google.com> Frank Swiderski <fes@google.com> Ted Tso <tytso@mit.edu> Keith Busch <keith.busch@intel.com> Just to give an idea on the performance boost with the vendor extension: Running fio [1], a stock NVMe driver I get about 200K read IOPs with my vendor patch I get about 1000K read IOPs. This was running with a null device i.e. the backing device simply returned success on every read IO request. [1] Running on a 4 core machine: fio --time_based --name=benchmark --runtime=30 --filename=/dev/nvme0n1 --nrfiles=1 --ioengine=libaio --iodepth=32 --direct=1 --invalidate=1 --verify=0 --verify_fatal=0 --numjobs=4 --rw=randread --blocksize=4k --randrepeat=false Signed-off-by: Rob Nelson <rlnelson@google.com> [mlin: port for upstream] Signed-off-by: Ming Lin <mlin@kernel.org> [koike: updated for upstream] Signed-off-by: Helen Koike <helen.koike@collabora.co.uk> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <keith.busch@intel.com>
2017-04-10 18:51:07 +03:00
if (dev->ctrl.oacs & NVME_CTRL_OACS_DBBUF_SUPP) {
result = nvme_dbbuf_dma_alloc(dev);
if (result)
dev_warn(dev->dev,
"unable to allocate dma for dbbuf\n");
}
if (dev->ctrl.hmpre) {
result = nvme_setup_host_mem(dev);
if (result < 0)
goto out;
}
result = nvme_setup_io_queues(dev);
if (result)
goto out;
/*
* Keep the controller around but remove all namespaces if we don't have
* any working I/O queue.
*/
if (dev->online_queues < 2) {
dev_warn(dev->ctrl.device, "IO queues not created\n");
nvme_kill_queues(&dev->ctrl);
nvme_remove_namespaces(&dev->ctrl);
new_state = NVME_CTRL_ADMIN_ONLY;
} else {
nvme_start_queues(&dev->ctrl);
nvme_wait_freeze(&dev->ctrl);
/* hit this only when allocate tagset fails */
if (nvme_dev_add(dev))
new_state = NVME_CTRL_ADMIN_ONLY;
nvme_unfreeze(&dev->ctrl);
}
/*
* If only admin queue live, keep it to do further investigation or
* recovery.
*/
if (!nvme_change_ctrl_state(&dev->ctrl, new_state)) {
dev_warn(dev->ctrl.device,
"failed to mark controller state %d\n", new_state);
goto out;
}
nvme_start_ctrl(&dev->ctrl);
return;
out:
nvme_remove_dead_ctrl(dev, result);
}
static void nvme_remove_dead_ctrl_work(struct work_struct *work)
{
struct nvme_dev *dev = container_of(work, struct nvme_dev, remove_work);
struct pci_dev *pdev = to_pci_dev(dev->dev);
nvme_kill_queues(&dev->ctrl);
if (pci_get_drvdata(pdev))
device_release_driver(&pdev->dev);
nvme_put_ctrl(&dev->ctrl);
}
static int nvme_pci_reg_read32(struct nvme_ctrl *ctrl, u32 off, u32 *val)
{
*val = readl(to_nvme_dev(ctrl)->bar + off);
return 0;
}
static int nvme_pci_reg_write32(struct nvme_ctrl *ctrl, u32 off, u32 val)
{
writel(val, to_nvme_dev(ctrl)->bar + off);
return 0;
}
static int nvme_pci_reg_read64(struct nvme_ctrl *ctrl, u32 off, u64 *val)
{
*val = readq(to_nvme_dev(ctrl)->bar + off);
return 0;
}
static int nvme_pci_get_address(struct nvme_ctrl *ctrl, char *buf, int size)
{
struct pci_dev *pdev = to_pci_dev(to_nvme_dev(ctrl)->dev);
return snprintf(buf, size, "%s", dev_name(&pdev->dev));
}
static const struct nvme_ctrl_ops nvme_pci_ctrl_ops = {
.name = "pcie",
.module = THIS_MODULE,
.flags = NVME_F_METADATA_SUPPORTED,
.reg_read32 = nvme_pci_reg_read32,
.reg_write32 = nvme_pci_reg_write32,
.reg_read64 = nvme_pci_reg_read64,
.free_ctrl = nvme_pci_free_ctrl,
.submit_async_event = nvme_pci_submit_async_event,
.get_address = nvme_pci_get_address,
};
static int nvme_dev_map(struct nvme_dev *dev)
{
struct pci_dev *pdev = to_pci_dev(dev->dev);
if (pci_request_mem_regions(pdev, "nvme"))
return -ENODEV;
if (nvme_remap_bar(dev, NVME_REG_DBS + 4096))
goto release;
return 0;
release:
pci_release_mem_regions(pdev);
return -ENODEV;
}
static unsigned long check_vendor_combination_bug(struct pci_dev *pdev)
{
if (pdev->vendor == 0x144d && pdev->device == 0xa802) {
/*
* Several Samsung devices seem to drop off the PCIe bus
* randomly when APST is on and uses the deepest sleep state.
* This has been observed on a Samsung "SM951 NVMe SAMSUNG
* 256GB", a "PM951 NVMe SAMSUNG 512GB", and a "Samsung SSD
* 950 PRO 256GB", but it seems to be restricted to two Dell
* laptops.
*/
if (dmi_match(DMI_SYS_VENDOR, "Dell Inc.") &&
(dmi_match(DMI_PRODUCT_NAME, "XPS 15 9550") ||
dmi_match(DMI_PRODUCT_NAME, "Precision 5510")))
return NVME_QUIRK_NO_DEEPEST_PS;
} else if (pdev->vendor == 0x144d && pdev->device == 0xa804) {
/*
* Samsung SSD 960 EVO drops off the PCIe bus after system
* suspend on a Ryzen board, ASUS PRIME B350M-A, as well as
* within few minutes after bootup on a Coffee Lake board -
* ASUS PRIME Z370-A
*/
if (dmi_match(DMI_BOARD_VENDOR, "ASUSTeK COMPUTER INC.") &&
(dmi_match(DMI_BOARD_NAME, "PRIME B350M-A") ||
dmi_match(DMI_BOARD_NAME, "PRIME Z370-A")))
return NVME_QUIRK_NO_APST;
}
return 0;
}
static void nvme_async_probe(void *data, async_cookie_t cookie)
{
struct nvme_dev *dev = data;
nvme_reset_ctrl_sync(&dev->ctrl);
flush_work(&dev->ctrl.scan_work);
nvme_put_ctrl(&dev->ctrl);
}
static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id)
{
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
int node, result = -ENOMEM;
struct nvme_dev *dev;
unsigned long quirks = id->driver_data;
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
node = dev_to_node(&pdev->dev);
if (node == NUMA_NO_NODE)
set_dev_node(&pdev->dev, first_memory_node);
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
dev = kzalloc_node(sizeof(*dev), GFP_KERNEL, node);
if (!dev)
return -ENOMEM;
nvme-pci: allocate device queues storage space at probe It may cause race by setting 'nvmeq' in nvme_init_request() because .init_request is called inside switching io scheduler, which may happen when the NVMe device is being resetted and its nvme queues are being freed and created. We don't have any sync between the two pathes. This patch changes the nvmeq allocation to occur at probe time so there is no way we can dereference it at init_request. [ 93.268391] kernel BUG at drivers/nvme/host/pci.c:408! [ 93.274146] invalid opcode: 0000 [#1] SMP [ 93.278618] Modules linked in: nfsv3 nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc ipmi_ssif vfat fat intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel iTCO_wdt intel_cstate ipmi_si iTCO_vendor_support intel_uncore mxm_wmi mei_me ipmi_devintf intel_rapl_perf pcspkr sg ipmi_msghandler lpc_ich dcdbas mei shpchp acpi_power_meter wmi dm_multipath ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ahci libahci nvme libata crc32c_intel nvme_core tg3 megaraid_sas ptp i2c_core pps_core dm_mirror dm_region_hash dm_log dm_mod [ 93.349071] CPU: 5 PID: 1842 Comm: sh Not tainted 4.15.0-rc2.ming+ #4 [ 93.356256] Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.5.5 08/16/2017 [ 93.364801] task: 00000000fb8abf2a task.stack: 0000000028bd82d1 [ 93.371408] RIP: 0010:nvme_init_request+0x36/0x40 [nvme] [ 93.377333] RSP: 0018:ffffc90002537ca8 EFLAGS: 00010246 [ 93.383161] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008 [ 93.391122] RDX: 0000000000000000 RSI: ffff880276ae0000 RDI: ffff88047bae9008 [ 93.399084] RBP: ffff88047bae9008 R08: ffff88047bae9008 R09: 0000000009dabc00 [ 93.407045] R10: 0000000000000004 R11: 000000000000299c R12: ffff880186bc1f00 [ 93.415007] R13: ffff880276ae0000 R14: 0000000000000000 R15: 0000000000000071 [ 93.422969] FS: 00007f33cf288740(0000) GS:ffff88047ba80000(0000) knlGS:0000000000000000 [ 93.431996] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 93.438407] CR2: 00007f33cf28e000 CR3: 000000047e5bb006 CR4: 00000000001606e0 [ 93.446368] Call Trace: [ 93.449103] blk_mq_alloc_rqs+0x231/0x2a0 [ 93.453579] blk_mq_sched_alloc_tags.isra.8+0x42/0x80 [ 93.459214] blk_mq_init_sched+0x7e/0x140 [ 93.463687] elevator_switch+0x5a/0x1f0 [ 93.467966] ? elevator_get.isra.17+0x52/0xc0 [ 93.472826] elv_iosched_store+0xde/0x150 [ 93.477299] queue_attr_store+0x4e/0x90 [ 93.481580] kernfs_fop_write+0xfa/0x180 [ 93.485958] __vfs_write+0x33/0x170 [ 93.489851] ? __inode_security_revalidate+0x4c/0x60 [ 93.495390] ? selinux_file_permission+0xda/0x130 [ 93.500641] ? _cond_resched+0x15/0x30 [ 93.504815] vfs_write+0xad/0x1a0 [ 93.508512] SyS_write+0x52/0xc0 [ 93.512113] do_syscall_64+0x61/0x1a0 [ 93.516199] entry_SYSCALL64_slow_path+0x25/0x25 [ 93.521351] RIP: 0033:0x7f33ce96aab0 [ 93.525337] RSP: 002b:00007ffe57570238 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 93.533785] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f33ce96aab0 [ 93.541746] RDX: 0000000000000006 RSI: 00007f33cf28e000 RDI: 0000000000000001 [ 93.549707] RBP: 00007f33cf28e000 R08: 000000000000000a R09: 00007f33cf288740 [ 93.557669] R10: 00007f33cf288740 R11: 0000000000000246 R12: 00007f33cec42400 [ 93.565630] R13: 0000000000000006 R14: 0000000000000001 R15: 0000000000000000 [ 93.573592] Code: 4c 8d 40 08 4c 39 c7 74 16 48 8b 00 48 8b 04 08 48 85 c0 74 16 48 89 86 78 01 00 00 31 c0 c3 8d 4a 01 48 63 c9 48 c1 e1 03 eb de <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 85 f6 53 48 89 [ 93.594676] RIP: nvme_init_request+0x36/0x40 [nvme] RSP: ffffc90002537ca8 [ 93.602273] ---[ end trace 810dde3993e5f14e ]--- Reported-by: Yi Zhang <yi.zhang@redhat.com> Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
2018-01-14 13:39:01 +03:00
dev->queues = kcalloc_node(num_possible_cpus() + 1,
sizeof(struct nvme_queue), GFP_KERNEL, node);
if (!dev->queues)
goto free;
dev->dev = get_device(&pdev->dev);
pci_set_drvdata(pdev, dev);
result = nvme_dev_map(dev);
if (result)
goto put_pci;
INIT_WORK(&dev->ctrl.reset_work, nvme_reset_work);
INIT_WORK(&dev->remove_work, nvme_remove_dead_ctrl_work);
mutex_init(&dev->shutdown_lock);
init_completion(&dev->ioq_wait);
result = nvme_setup_prp_pools(dev);
if (result)
goto unmap;
quirks |= check_vendor_combination_bug(pdev);
result = nvme_init_ctrl(&dev->ctrl, &pdev->dev, &nvme_pci_ctrl_ops,
quirks);
if (result)
goto release_pools;
dev_info(dev->ctrl.device, "pci function %s\n", dev_name(&pdev->dev));
nvme_get_ctrl(&dev->ctrl);
async_schedule(nvme_async_probe, dev);
return 0;
release_pools:
nvme_release_prp_pools(dev);
unmap:
nvme_dev_unmap(dev);
put_pci:
put_device(dev->dev);
free:
kfree(dev->queues);
kfree(dev);
return result;
}
static void nvme_reset_prepare(struct pci_dev *pdev)
{
struct nvme_dev *dev = pci_get_drvdata(pdev);
pci-v4.13-changes -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAABAgAGBQJZYAFUAAoJEFmIoMA60/r8cFQP/A4fpdjhd42WRNQXGTpZieop i40lBQtGdBn/UY97U6BoutcS1ygDi9OiSzg+IR6I90iMgidqyUHFhe4hGWgVHD2g Tg0KLzd+lKKfQ6Gqt1P6t4dLGLvyEj5NUbCeFE4XYODAUkkiBaOndax6DK1GvU54 Vjuj63rHtMKFR/tG/4iFTigObqyI8QE6O9JVxwuvIyEX6RXKbJe+wkulv5taSnWt Ne94950i10MrELtNreVdi8UbCbXiqjg0r5sKI/WTJ7Bc7WsC7X5PhWlhcNrbHyBT Ivhoypkui3Ky8gvwWqL0KBG+cRp8prBXAdabrD9wRbz0TKnfGI6pQzseCGRnkE6T mhlSJpsSNIHaejoCjk93yPn5oRiTNtPMdVhMpEQL9V/crVRGRRmbd7v2TYvpMHVR JaPZ8bv+C2aBTY8uL3/v/rgrjsMKOYFeaxeNklpErxrknsbgb6BgubmeZXDvTBVv YUIbAkvveonUKisv+kbD8L7tp1+jdbRUT0AikS0NVgAJQhfArOmBcDpTL9YC51vE feFhkVx4A32vvOm7Zcg9A7IMXNjeSfccKGw3dJOAvzgDODuJiaCG6S0o7B5Yngze axMi87ixGT4QM98z/I4MC8E9rDrJdIitlpvb6ZBgiLzoO3kmvsIZZKt8UxWqf5r8 w3U2HoyKH13Qbkn1xkum =mkyb -----END PGP SIGNATURE----- Merge tag 'pci-v4.13-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci Pull PCI updates from Bjorn Helgaas: - add sysfs max_link_speed/width, current_link_speed/width (Wong Vee Khee) - make host bridge IRQ mapping much more generic (Matthew Minter, Lorenzo Pieralisi) - convert most drivers to pci_scan_root_bus_bridge() (Lorenzo Pieralisi) - mutex sriov_configure() (Jakub Kicinski) - mutex pci_error_handlers callbacks (Christoph Hellwig) - split ->reset_notify() into ->reset_prepare()/reset_done() (Christoph Hellwig) - support multiple PCIe portdrv interrupts for MSI as well as MSI-X (Gabriele Paoloni) - allocate MSI/MSI-X vector for Downstream Port Containment (Gabriele Paoloni) - fix MSI IRQ affinity pre/post/min_vecs issue (Michael Hernandez) - test INTx masking during enumeration, not at run-time (Piotr Gregor) - avoid using device_may_wakeup() for runtime PM (Rafael J. Wysocki) - restore the status of PCI devices across hibernation (Chen Yu) - keep parent resources that start at 0x0 (Ard Biesheuvel) - enable ECRC only if device supports it (Bjorn Helgaas) - restore PRI and PASID state after Function-Level Reset (CQ Tang) - skip DPC event if device is not present (Keith Busch) - check domain when matching SMBIOS info (Sujith Pandel) - mark Intel XXV710 NIC INTx masking as broken (Alex Williamson) - avoid AMD SB7xx EHCI USB wakeup defect (Kai-Heng Feng) - work around long-standing Macbook Pro poweroff issue (Bjorn Helgaas) - add Switchtec "running" status flag (Logan Gunthorpe) - fix dra7xx incorrect RW1C IRQ register usage (Arvind Yadav) - modify xilinx-nwl IRQ chip for legacy interrupts (Bharat Kumar Gogada) - move VMD SRCU cleanup after bus, child device removal (Jon Derrick) - add Faraday clock handling (Linus Walleij) - configure Rockchip MPS and reorganize (Shawn Lin) - limit Qualcomm TLP size to 2K (hardware issue) (Srinivas Kandagatla) - support Tegra MSI 64-bit addressing (Thierry Reding) - use Rockchip normal (not privileged) register bank (Shawn Lin) - add HiSilicon Kirin SoC PCIe controller driver (Xiaowei Song) - add Sigma Designs Tango SMP8759 PCIe controller driver (Marc Gonzalez) - add MediaTek PCIe host controller support (Ryder Lee) - add Qualcomm IPQ4019 support (John Crispin) - add HyperV vPCI protocol v1.2 support (Jork Loeser) - add i.MX6 regulator support (Quentin Schulz) * tag 'pci-v4.13-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci: (113 commits) PCI: tango: Add Sigma Designs Tango SMP8759 PCIe host bridge support PCI: Add DT binding for Sigma Designs Tango PCIe controller PCI: rockchip: Use normal register bank for config accessors dt-bindings: PCI: Add documentation for MediaTek PCIe PCI: Remove __pci_dev_reset() and pci_dev_reset() PCI: Split ->reset_notify() method into ->reset_prepare() and ->reset_done() PCI: xilinx: Make of_device_ids const PCI: xilinx-nwl: Modify IRQ chip for legacy interrupts PCI: vmd: Move SRCU cleanup after bus, child device removal PCI: vmd: Correct comment: VMD domains start at 0x10000, not 0x1000 PCI: versatile: Add local struct device pointers PCI: tegra: Do not allocate MSI target memory PCI: tegra: Support MSI 64-bit addressing PCI: rockchip: Use local struct device pointer consistently PCI: rockchip: Check for clk_prepare_enable() errors during resume MAINTAINERS: Remove Wenrui Li as Rockchip PCIe driver maintainer PCI: rockchip: Configure RC's MPS setting PCI: rockchip: Reconfigure configuration space header type PCI: rockchip: Split out rockchip_pcie_cfg_configuration_accesses() PCI: rockchip: Move configuration accesses into rockchip_pcie_cfg_atu() ...
2017-07-09 01:51:57 +03:00
nvme_dev_disable(dev, false);
}
static void nvme_reset_done(struct pci_dev *pdev)
{
pci-v4.13-changes -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAABAgAGBQJZYAFUAAoJEFmIoMA60/r8cFQP/A4fpdjhd42WRNQXGTpZieop i40lBQtGdBn/UY97U6BoutcS1ygDi9OiSzg+IR6I90iMgidqyUHFhe4hGWgVHD2g Tg0KLzd+lKKfQ6Gqt1P6t4dLGLvyEj5NUbCeFE4XYODAUkkiBaOndax6DK1GvU54 Vjuj63rHtMKFR/tG/4iFTigObqyI8QE6O9JVxwuvIyEX6RXKbJe+wkulv5taSnWt Ne94950i10MrELtNreVdi8UbCbXiqjg0r5sKI/WTJ7Bc7WsC7X5PhWlhcNrbHyBT Ivhoypkui3Ky8gvwWqL0KBG+cRp8prBXAdabrD9wRbz0TKnfGI6pQzseCGRnkE6T mhlSJpsSNIHaejoCjk93yPn5oRiTNtPMdVhMpEQL9V/crVRGRRmbd7v2TYvpMHVR JaPZ8bv+C2aBTY8uL3/v/rgrjsMKOYFeaxeNklpErxrknsbgb6BgubmeZXDvTBVv YUIbAkvveonUKisv+kbD8L7tp1+jdbRUT0AikS0NVgAJQhfArOmBcDpTL9YC51vE feFhkVx4A32vvOm7Zcg9A7IMXNjeSfccKGw3dJOAvzgDODuJiaCG6S0o7B5Yngze axMi87ixGT4QM98z/I4MC8E9rDrJdIitlpvb6ZBgiLzoO3kmvsIZZKt8UxWqf5r8 w3U2HoyKH13Qbkn1xkum =mkyb -----END PGP SIGNATURE----- Merge tag 'pci-v4.13-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci Pull PCI updates from Bjorn Helgaas: - add sysfs max_link_speed/width, current_link_speed/width (Wong Vee Khee) - make host bridge IRQ mapping much more generic (Matthew Minter, Lorenzo Pieralisi) - convert most drivers to pci_scan_root_bus_bridge() (Lorenzo Pieralisi) - mutex sriov_configure() (Jakub Kicinski) - mutex pci_error_handlers callbacks (Christoph Hellwig) - split ->reset_notify() into ->reset_prepare()/reset_done() (Christoph Hellwig) - support multiple PCIe portdrv interrupts for MSI as well as MSI-X (Gabriele Paoloni) - allocate MSI/MSI-X vector for Downstream Port Containment (Gabriele Paoloni) - fix MSI IRQ affinity pre/post/min_vecs issue (Michael Hernandez) - test INTx masking during enumeration, not at run-time (Piotr Gregor) - avoid using device_may_wakeup() for runtime PM (Rafael J. Wysocki) - restore the status of PCI devices across hibernation (Chen Yu) - keep parent resources that start at 0x0 (Ard Biesheuvel) - enable ECRC only if device supports it (Bjorn Helgaas) - restore PRI and PASID state after Function-Level Reset (CQ Tang) - skip DPC event if device is not present (Keith Busch) - check domain when matching SMBIOS info (Sujith Pandel) - mark Intel XXV710 NIC INTx masking as broken (Alex Williamson) - avoid AMD SB7xx EHCI USB wakeup defect (Kai-Heng Feng) - work around long-standing Macbook Pro poweroff issue (Bjorn Helgaas) - add Switchtec "running" status flag (Logan Gunthorpe) - fix dra7xx incorrect RW1C IRQ register usage (Arvind Yadav) - modify xilinx-nwl IRQ chip for legacy interrupts (Bharat Kumar Gogada) - move VMD SRCU cleanup after bus, child device removal (Jon Derrick) - add Faraday clock handling (Linus Walleij) - configure Rockchip MPS and reorganize (Shawn Lin) - limit Qualcomm TLP size to 2K (hardware issue) (Srinivas Kandagatla) - support Tegra MSI 64-bit addressing (Thierry Reding) - use Rockchip normal (not privileged) register bank (Shawn Lin) - add HiSilicon Kirin SoC PCIe controller driver (Xiaowei Song) - add Sigma Designs Tango SMP8759 PCIe controller driver (Marc Gonzalez) - add MediaTek PCIe host controller support (Ryder Lee) - add Qualcomm IPQ4019 support (John Crispin) - add HyperV vPCI protocol v1.2 support (Jork Loeser) - add i.MX6 regulator support (Quentin Schulz) * tag 'pci-v4.13-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci: (113 commits) PCI: tango: Add Sigma Designs Tango SMP8759 PCIe host bridge support PCI: Add DT binding for Sigma Designs Tango PCIe controller PCI: rockchip: Use normal register bank for config accessors dt-bindings: PCI: Add documentation for MediaTek PCIe PCI: Remove __pci_dev_reset() and pci_dev_reset() PCI: Split ->reset_notify() method into ->reset_prepare() and ->reset_done() PCI: xilinx: Make of_device_ids const PCI: xilinx-nwl: Modify IRQ chip for legacy interrupts PCI: vmd: Move SRCU cleanup after bus, child device removal PCI: vmd: Correct comment: VMD domains start at 0x10000, not 0x1000 PCI: versatile: Add local struct device pointers PCI: tegra: Do not allocate MSI target memory PCI: tegra: Support MSI 64-bit addressing PCI: rockchip: Use local struct device pointer consistently PCI: rockchip: Check for clk_prepare_enable() errors during resume MAINTAINERS: Remove Wenrui Li as Rockchip PCIe driver maintainer PCI: rockchip: Configure RC's MPS setting PCI: rockchip: Reconfigure configuration space header type PCI: rockchip: Split out rockchip_pcie_cfg_configuration_accesses() PCI: rockchip: Move configuration accesses into rockchip_pcie_cfg_atu() ...
2017-07-09 01:51:57 +03:00
struct nvme_dev *dev = pci_get_drvdata(pdev);
nvme_reset_ctrl_sync(&dev->ctrl);
}
static void nvme_shutdown(struct pci_dev *pdev)
{
struct nvme_dev *dev = pci_get_drvdata(pdev);
nvme_dev_disable(dev, true);
}
/*
* The driver's remove may be called on a device in a partially initialized
* state. This function must not have any dependencies on the device state in
* order to proceed.
*/
static void nvme_remove(struct pci_dev *pdev)
{
struct nvme_dev *dev = pci_get_drvdata(pdev);
nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_DELETING);
cancel_work_sync(&dev->ctrl.reset_work);
pci_set_drvdata(pdev, NULL);
if (!pci_device_is_present(pdev)) {
nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_DEAD);
nvme_dev_disable(dev, false);
}
flush_work(&dev->ctrl.reset_work);
nvme_stop_ctrl(&dev->ctrl);
nvme_remove_namespaces(&dev->ctrl);
nvme_dev_disable(dev, true);
nvme_free_host_mem(dev);
NVMe: Convert to blk-mq This converts the NVMe driver to a blk-mq request-based driver. The NVMe driver is currently bio-based and implements queue logic within itself. By using blk-mq, a lot of these responsibilities can be moved and simplified. The patch is divided into the following blocks: * Per-command data and cmdid have been moved into the struct request field. The cmdid_data can be retrieved using blk_mq_rq_to_pdu() and id maintenance are now handled by blk-mq through the rq->tag field. * The logic for splitting bio's has been moved into the blk-mq layer. The driver instead notifies the block layer about limited gap support in SG lists. * blk-mq handles timeouts and is reimplemented within nvme_timeout(). This both includes abort handling and command cancelation. * Assignment of nvme queues to CPUs are replaced with the blk-mq version. The current blk-mq strategy is to assign the number of mapped queues and CPUs to provide synergy, while the nvme driver assign as many nvme hw queues as possible. This can be implemented in blk-mq if needed. * NVMe queues are merged with the tags structure of blk-mq. * blk-mq takes care of setup/teardown of nvme queues and guards invalid accesses. Therefore, RCU-usage for nvme queues can be removed. * IO tracing and accounting are handled by blk-mq and therefore removed. * Queue suspension logic is replaced with the logic from the block layer. Contributions in this patch from: Sam Bradshaw <sbradshaw@micron.com> Jens Axboe <axboe@fb.com> Keith Busch <keith.busch@intel.com> Robert Nelson <rlnelson@google.com> Acked-by: Keith Busch <keith.busch@intel.com> Acked-by: Jens Axboe <axboe@fb.com> Updated for new ->queue_rq() prototype. Signed-off-by: Jens Axboe <axboe@fb.com>
2014-11-04 18:20:14 +03:00
nvme_dev_remove_admin(dev);
nvme_free_queues(dev, 0);
nvme_uninit_ctrl(&dev->ctrl);
nvme_release_prp_pools(dev);
nvme_dev_unmap(dev);
nvme_put_ctrl(&dev->ctrl);
}
static int nvme_pci_sriov_configure(struct pci_dev *pdev, int numvfs)
{
int ret = 0;
if (numvfs == 0) {
if (pci_vfs_assigned(pdev)) {
dev_warn(&pdev->dev,
"Cannot disable SR-IOV VFs while assigned\n");
return -EPERM;
}
pci_disable_sriov(pdev);
return 0;
}
ret = pci_enable_sriov(pdev, numvfs);
return ret ? ret : numvfs;
}
#ifdef CONFIG_PM_SLEEP
static int nvme_suspend(struct device *dev)
{
struct pci_dev *pdev = to_pci_dev(dev);
struct nvme_dev *ndev = pci_get_drvdata(pdev);
nvme_dev_disable(ndev, true);
return 0;
}
static int nvme_resume(struct device *dev)
{
struct pci_dev *pdev = to_pci_dev(dev);
struct nvme_dev *ndev = pci_get_drvdata(pdev);
nvme_reset_ctrl(&ndev->ctrl);
return 0;
}
#endif
static SIMPLE_DEV_PM_OPS(nvme_dev_pm_ops, nvme_suspend, nvme_resume);
static pci_ers_result_t nvme_error_detected(struct pci_dev *pdev,
pci_channel_state_t state)
{
struct nvme_dev *dev = pci_get_drvdata(pdev);
/*
* A frozen channel requires a reset. When detected, this method will
* shutdown the controller to quiesce. The controller will be restarted
* after the slot reset through driver's slot_reset callback.
*/
switch (state) {
case pci_channel_io_normal:
return PCI_ERS_RESULT_CAN_RECOVER;
case pci_channel_io_frozen:
dev_warn(dev->ctrl.device,
"frozen state error detected, reset controller\n");
nvme_dev_disable(dev, false);
return PCI_ERS_RESULT_NEED_RESET;
case pci_channel_io_perm_failure:
dev_warn(dev->ctrl.device,
"failure state error detected, request disconnect\n");
return PCI_ERS_RESULT_DISCONNECT;
}
return PCI_ERS_RESULT_NEED_RESET;
}
static pci_ers_result_t nvme_slot_reset(struct pci_dev *pdev)
{
struct nvme_dev *dev = pci_get_drvdata(pdev);
dev_info(dev->ctrl.device, "restart after slot reset\n");
pci_restore_state(pdev);
nvme_reset_ctrl_sync(&dev->ctrl);
switch (dev->ctrl.state) {
case NVME_CTRL_LIVE:
case NVME_CTRL_ADMIN_ONLY:
return PCI_ERS_RESULT_RECOVERED;
default:
return PCI_ERS_RESULT_DISCONNECT;
}
}
static void nvme_error_resume(struct pci_dev *pdev)
{
pci_cleanup_aer_uncorrect_error_status(pdev);
}
static const struct pci_error_handlers nvme_err_handler = {
.error_detected = nvme_error_detected,
.slot_reset = nvme_slot_reset,
.resume = nvme_error_resume,
.reset_prepare = nvme_reset_prepare,
.reset_done = nvme_reset_done,
};
static const struct pci_device_id nvme_id_table[] = {
{ PCI_VDEVICE(INTEL, 0x0953),
.driver_data = NVME_QUIRK_STRIPE_SIZE |
NVME_QUIRK_DEALLOCATE_ZEROES, },
{ PCI_VDEVICE(INTEL, 0x0a53),
.driver_data = NVME_QUIRK_STRIPE_SIZE |
NVME_QUIRK_DEALLOCATE_ZEROES, },
{ PCI_VDEVICE(INTEL, 0x0a54),
.driver_data = NVME_QUIRK_STRIPE_SIZE |
NVME_QUIRK_DEALLOCATE_ZEROES, },
{ PCI_VDEVICE(INTEL, 0x0a55),
.driver_data = NVME_QUIRK_STRIPE_SIZE |
NVME_QUIRK_DEALLOCATE_ZEROES, },
{ PCI_VDEVICE(INTEL, 0xf1a5), /* Intel 600P/P3100 */
.driver_data = NVME_QUIRK_NO_DEEPEST_PS },
{ PCI_VDEVICE(INTEL, 0x5845), /* Qemu emulated controller */
.driver_data = NVME_QUIRK_IDENTIFY_CNS, },
{ PCI_DEVICE(0x1bb1, 0x0100), /* Seagate Nytro Flash Storage */
.driver_data = NVME_QUIRK_DELAY_BEFORE_CHK_RDY, },
{ PCI_DEVICE(0x1c58, 0x0003), /* HGST adapter */
.driver_data = NVME_QUIRK_DELAY_BEFORE_CHK_RDY, },
{ PCI_DEVICE(0x1c58, 0x0023), /* WDC SN200 adapter */
.driver_data = NVME_QUIRK_DELAY_BEFORE_CHK_RDY, },
{ PCI_DEVICE(0x1c5f, 0x0540), /* Memblaze Pblaze4 adapter */
.driver_data = NVME_QUIRK_DELAY_BEFORE_CHK_RDY, },
{ PCI_DEVICE(0x144d, 0xa821), /* Samsung PM1725 */
.driver_data = NVME_QUIRK_DELAY_BEFORE_CHK_RDY, },
{ PCI_DEVICE(0x144d, 0xa822), /* Samsung PM1725a */
.driver_data = NVME_QUIRK_DELAY_BEFORE_CHK_RDY, },
{ PCI_DEVICE(0x1d1d, 0x1f1f), /* LighNVM qemu device */
.driver_data = NVME_QUIRK_LIGHTNVM, },
{ PCI_DEVICE(0x1d1d, 0x2807), /* CNEX WL */
.driver_data = NVME_QUIRK_LIGHTNVM, },
{ PCI_DEVICE(0x1d1d, 0x2601), /* CNEX Granby */
.driver_data = NVME_QUIRK_LIGHTNVM, },
{ PCI_DEVICE_CLASS(PCI_CLASS_STORAGE_EXPRESS, 0xffffff) },
{ PCI_DEVICE(PCI_VENDOR_ID_APPLE, 0x2001) },
{ PCI_DEVICE(PCI_VENDOR_ID_APPLE, 0x2003) },
{ 0, }
};
MODULE_DEVICE_TABLE(pci, nvme_id_table);
static struct pci_driver nvme_driver = {
.name = "nvme",
.id_table = nvme_id_table,
.probe = nvme_probe,
.remove = nvme_remove,
.shutdown = nvme_shutdown,
.driver = {
.pm = &nvme_dev_pm_ops,
},
.sriov_configure = nvme_pci_sriov_configure,
.err_handler = &nvme_err_handler,
};
static int __init nvme_init(void)
{
return pci_register_driver(&nvme_driver);
}
static void __exit nvme_exit(void)
{
pci_unregister_driver(&nvme_driver);
flush_workqueue(nvme_wq);
_nvme_check_size();
}
MODULE_AUTHOR("Matthew Wilcox <willy@linux.intel.com>");
MODULE_LICENSE("GPL");
MODULE_VERSION("1.0");
module_init(nvme_init);
module_exit(nvme_exit);