From 32f57cb1b2c8d6f20aefec7052b1bfeb7e3b69d4 Mon Sep 17 00:00:00 2001 From: Mike Marciniszyn Date: Thu, 17 Feb 2022 08:59:32 -0500 Subject: [PATCH 1/5] IB/qib: Fix duplicate sysfs directory name The qib driver load has been failing with the following message: sysfs: cannot create duplicate filename '/devices/pci0000:80/0000:80:02.0/0000:81:00.0/infiniband/qib0/ports/1/linkcontrol' The patch below has two "linkcontrol" names causing the duplication. Fix by using the correct "diag_counters" name on the second instance. Fixes: 4a7aaf88c89f ("RDMA/qib: Use attributes for the port sysfs") Link: https://lore.kernel.org/r/1645106372-23004-1-git-send-email-mike.marciniszyn@cornelisnetworks.com Cc: Reviewed-by: Dennis Dalessandro Signed-off-by: Mike Marciniszyn Signed-off-by: Jason Gunthorpe --- drivers/infiniband/hw/qib/qib_sysfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/hw/qib/qib_sysfs.c b/drivers/infiniband/hw/qib/qib_sysfs.c index 0a3b28142c05..41c272980f91 100644 --- a/drivers/infiniband/hw/qib/qib_sysfs.c +++ b/drivers/infiniband/hw/qib/qib_sysfs.c @@ -541,7 +541,7 @@ static struct attribute *port_diagc_attributes[] = { }; static const struct attribute_group port_diagc_group = { - .name = "linkcontrol", + .name = "diag_counters", .attrs = port_diagc_attributes, }; From 8700af2cc18c919b2a83e74e0479038fd113c15d Mon Sep 17 00:00:00 2001 From: Md Haris Iqbal Date: Thu, 17 Feb 2022 04:09:28 +0100 Subject: [PATCH 2/5] RDMA/rtrs-clt: Fix possible double free in error case Callback function rtrs_clt_dev_release() for put_device() calls kfree(clt) to free memory. We shouldn't call kfree(clt) again, and we can't use the clt after kfree too. Replace device_register() with device_initialize() and device_add() so that dev_set_name can() be used appropriately. Move mutex_destroy() to the release function so it can be called in the alloc_clt err path. Fixes: eab098246625 ("RDMA/rtrs-clt: Refactor the failure cases in alloc_clt") Link: https://lore.kernel.org/r/20220217030929.323849-1-haris.iqbal@ionos.com Reported-by: Miaoqian Lin Signed-off-by: Md Haris Iqbal Reviewed-by: Jack Wang Signed-off-by: Jason Gunthorpe --- drivers/infiniband/ulp/rtrs/rtrs-clt.c | 37 ++++++++++++++------------ 1 file changed, 20 insertions(+), 17 deletions(-) diff --git a/drivers/infiniband/ulp/rtrs/rtrs-clt.c b/drivers/infiniband/ulp/rtrs/rtrs-clt.c index 7c3f98e57889..6164b07a1644 100644 --- a/drivers/infiniband/ulp/rtrs/rtrs-clt.c +++ b/drivers/infiniband/ulp/rtrs/rtrs-clt.c @@ -2682,6 +2682,8 @@ static void rtrs_clt_dev_release(struct device *dev) struct rtrs_clt_sess *clt = container_of(dev, struct rtrs_clt_sess, dev); + mutex_destroy(&clt->paths_ev_mutex); + mutex_destroy(&clt->paths_mutex); kfree(clt); } @@ -2711,6 +2713,8 @@ static struct rtrs_clt_sess *alloc_clt(const char *sessname, size_t paths_num, return ERR_PTR(-ENOMEM); } + clt->dev.class = rtrs_clt_dev_class; + clt->dev.release = rtrs_clt_dev_release; uuid_gen(&clt->paths_uuid); INIT_LIST_HEAD_RCU(&clt->paths_list); clt->paths_num = paths_num; @@ -2727,43 +2731,41 @@ static struct rtrs_clt_sess *alloc_clt(const char *sessname, size_t paths_num, init_waitqueue_head(&clt->permits_wait); mutex_init(&clt->paths_ev_mutex); mutex_init(&clt->paths_mutex); + device_initialize(&clt->dev); - clt->dev.class = rtrs_clt_dev_class; - clt->dev.release = rtrs_clt_dev_release; err = dev_set_name(&clt->dev, "%s", sessname); if (err) - goto err; + goto err_put; + /* * Suppress user space notification until * sysfs files are created */ dev_set_uevent_suppress(&clt->dev, true); - err = device_register(&clt->dev); - if (err) { - put_device(&clt->dev); - goto err; - } + err = device_add(&clt->dev); + if (err) + goto err_put; clt->kobj_paths = kobject_create_and_add("paths", &clt->dev.kobj); if (!clt->kobj_paths) { err = -ENOMEM; - goto err_dev; + goto err_del; } err = rtrs_clt_create_sysfs_root_files(clt); if (err) { kobject_del(clt->kobj_paths); kobject_put(clt->kobj_paths); - goto err_dev; + goto err_del; } dev_set_uevent_suppress(&clt->dev, false); kobject_uevent(&clt->dev.kobj, KOBJ_ADD); return clt; -err_dev: - device_unregister(&clt->dev); -err: +err_del: + device_del(&clt->dev); +err_put: free_percpu(clt->pcpu_path); - kfree(clt); + put_device(&clt->dev); return ERR_PTR(err); } @@ -2771,9 +2773,10 @@ static void free_clt(struct rtrs_clt_sess *clt) { free_permits(clt); free_percpu(clt->pcpu_path); - mutex_destroy(&clt->paths_ev_mutex); - mutex_destroy(&clt->paths_mutex); - /* release callback will free clt in last put */ + + /* + * release callback will free clt and destroy mutexes in last put + */ device_unregister(&clt->dev); } From c46fa8911b17e3f808679061a8af8bee219f4602 Mon Sep 17 00:00:00 2001 From: Md Haris Iqbal Date: Thu, 17 Feb 2022 04:09:29 +0100 Subject: [PATCH 3/5] RDMA/rtrs-clt: Move free_permit from free_clt to rtrs_clt_close Error path of rtrs_clt_open() calls free_clt(), where free_permit is called. This is wrong since error path of rtrs_clt_open() does not need to call free_permit(). Also, moving free_permits() call to rtrs_clt_close(), makes it more aligned with the call to alloc_permit() in rtrs_clt_open(). Fixes: 6a98d71daea1 ("RDMA/rtrs: client: main functionality") Link: https://lore.kernel.org/r/20220217030929.323849-2-haris.iqbal@ionos.com Signed-off-by: Md Haris Iqbal Reviewed-by: Jack Wang Signed-off-by: Jason Gunthorpe --- drivers/infiniband/ulp/rtrs/rtrs-clt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/ulp/rtrs/rtrs-clt.c b/drivers/infiniband/ulp/rtrs/rtrs-clt.c index 6164b07a1644..759b85f03331 100644 --- a/drivers/infiniband/ulp/rtrs/rtrs-clt.c +++ b/drivers/infiniband/ulp/rtrs/rtrs-clt.c @@ -2771,7 +2771,6 @@ err_put: static void free_clt(struct rtrs_clt_sess *clt) { - free_permits(clt); free_percpu(clt->pcpu_path); /* @@ -2893,6 +2892,7 @@ void rtrs_clt_close(struct rtrs_clt_sess *clt) rtrs_clt_destroy_path_files(clt_path, NULL); kobject_put(&clt_path->kobj); } + free_permits(clt); free_clt(clt); } EXPORT_SYMBOL(rtrs_clt_close); From 081bdc9fe05bb23248f5effb6f811da3da4b8252 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Tue, 15 Feb 2022 13:05:11 -0800 Subject: [PATCH 4/5] RDMA/ib_srp: Fix a deadlock Remove the flush_workqueue(system_long_wq) call since flushing system_long_wq is deadlock-prone and since that call is redundant with a preceding cancel_work_sync() Link: https://lore.kernel.org/r/20220215210511.28303-3-bvanassche@acm.org Fixes: ef6c49d87c34 ("IB/srp: Eliminate state SRP_TARGET_DEAD") Reported-by: syzbot+831661966588c802aae9@syzkaller.appspotmail.com Signed-off-by: Bart Van Assche Reviewed-by: Leon Romanovsky Signed-off-by: Jason Gunthorpe --- drivers/infiniband/ulp/srp/ib_srp.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c index e174e853f8a4..285b766e4e70 100644 --- a/drivers/infiniband/ulp/srp/ib_srp.c +++ b/drivers/infiniband/ulp/srp/ib_srp.c @@ -4047,9 +4047,11 @@ static void srp_remove_one(struct ib_device *device, void *client_data) spin_unlock(&host->target_lock); /* - * Wait for tl_err and target port removal tasks. + * srp_queue_remove_work() queues a call to + * srp_remove_target(). The latter function cancels + * target->tl_err_work so waiting for the remove works to + * finish is sufficient. */ - flush_workqueue(system_long_wq); flush_workqueue(srp_remove_wq); kfree(host); From 22e9f71072fa605cbf033158db58e0790101928d Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Wed, 23 Feb 2022 11:23:57 -0400 Subject: [PATCH 5/5] RDMA/cma: Do not change route.addr.src_addr outside state checks If the state is not idle then resolve_prepare_src() should immediately fail and no change to global state should happen. However, it unconditionally overwrites the src_addr trying to build a temporary any address. For instance if the state is already RDMA_CM_LISTEN then this will corrupt the src_addr and would cause the test in cma_cancel_operation(): if (cma_any_addr(cma_src_addr(id_priv)) && !id_priv->cma_dev) Which would manifest as this trace from syzkaller: BUG: KASAN: use-after-free in __list_add_valid+0x93/0xa0 lib/list_debug.c:26 Read of size 8 at addr ffff8881546491e0 by task syz-executor.1/32204 CPU: 1 PID: 32204 Comm: syz-executor.1 Not tainted 5.12.0-rc8-syzkaller #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 Call Trace: __dump_stack lib/dump_stack.c:79 [inline] dump_stack+0x141/0x1d7 lib/dump_stack.c:120 print_address_description.constprop.0.cold+0x5b/0x2f8 mm/kasan/report.c:232 __kasan_report mm/kasan/report.c:399 [inline] kasan_report.cold+0x7c/0xd8 mm/kasan/report.c:416 __list_add_valid+0x93/0xa0 lib/list_debug.c:26 __list_add include/linux/list.h:67 [inline] list_add_tail include/linux/list.h:100 [inline] cma_listen_on_all drivers/infiniband/core/cma.c:2557 [inline] rdma_listen+0x787/0xe00 drivers/infiniband/core/cma.c:3751 ucma_listen+0x16a/0x210 drivers/infiniband/core/ucma.c:1102 ucma_write+0x259/0x350 drivers/infiniband/core/ucma.c:1732 vfs_write+0x28e/0xa30 fs/read_write.c:603 ksys_write+0x1ee/0x250 fs/read_write.c:658 do_syscall_64+0x2d/0x70 arch/x86/entry/common.c:46 entry_SYSCALL_64_after_hwframe+0x44/0xae This is indicating that an rdma_id_private was destroyed without doing cma_cancel_listens(). Instead of trying to re-use the src_addr memory to indirectly create an any address derived from the dst build one explicitly on the stack and bind to that as any other normal flow would do. rdma_bind_addr() will copy it over the src_addr once it knows the state is valid. This is similar to commit bc0bdc5afaa7 ("RDMA/cma: Do not change route.addr.src_addr.ss_family") Link: https://lore.kernel.org/r/0-v2-e975c8fd9ef2+11e-syz_cma_srcaddr_jgg@nvidia.com Cc: stable@vger.kernel.org Fixes: 732d41c545bb ("RDMA/cma: Make the locking for automatic state transition more clear") Reported-by: syzbot+c94a3675a626f6333d74@syzkaller.appspotmail.com Reviewed-by: Leon Romanovsky Signed-off-by: Jason Gunthorpe --- drivers/infiniband/core/cma.c | 38 +++++++++++++++++++++-------------- 1 file changed, 23 insertions(+), 15 deletions(-) diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c index c447526288f4..50c53409ceb6 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -3370,22 +3370,30 @@ err: static int cma_bind_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, const struct sockaddr *dst_addr) { - if (!src_addr || !src_addr->sa_family) { - src_addr = (struct sockaddr *) &id->route.addr.src_addr; - src_addr->sa_family = dst_addr->sa_family; - if (IS_ENABLED(CONFIG_IPV6) && - dst_addr->sa_family == AF_INET6) { - struct sockaddr_in6 *src_addr6 = (struct sockaddr_in6 *) src_addr; - struct sockaddr_in6 *dst_addr6 = (struct sockaddr_in6 *) dst_addr; - src_addr6->sin6_scope_id = dst_addr6->sin6_scope_id; - if (ipv6_addr_type(&dst_addr6->sin6_addr) & IPV6_ADDR_LINKLOCAL) - id->route.addr.dev_addr.bound_dev_if = dst_addr6->sin6_scope_id; - } else if (dst_addr->sa_family == AF_IB) { - ((struct sockaddr_ib *) src_addr)->sib_pkey = - ((struct sockaddr_ib *) dst_addr)->sib_pkey; - } + struct sockaddr_storage zero_sock = {}; + + if (src_addr && src_addr->sa_family) + return rdma_bind_addr(id, src_addr); + + /* + * When the src_addr is not specified, automatically supply an any addr + */ + zero_sock.ss_family = dst_addr->sa_family; + if (IS_ENABLED(CONFIG_IPV6) && dst_addr->sa_family == AF_INET6) { + struct sockaddr_in6 *src_addr6 = + (struct sockaddr_in6 *)&zero_sock; + struct sockaddr_in6 *dst_addr6 = + (struct sockaddr_in6 *)dst_addr; + + src_addr6->sin6_scope_id = dst_addr6->sin6_scope_id; + if (ipv6_addr_type(&dst_addr6->sin6_addr) & IPV6_ADDR_LINKLOCAL) + id->route.addr.dev_addr.bound_dev_if = + dst_addr6->sin6_scope_id; + } else if (dst_addr->sa_family == AF_IB) { + ((struct sockaddr_ib *)&zero_sock)->sib_pkey = + ((struct sockaddr_ib *)dst_addr)->sib_pkey; } - return rdma_bind_addr(id, src_addr); + return rdma_bind_addr(id, (struct sockaddr *)&zero_sock); } /*