IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
The input parameters may not be reliable, so check the vlan id before
using it, otherwise may set wrong vlan id into hardware.
Fixes: dc8131d846 ("net: hns3: Fix for packet loss due wrong filter config in VLAN tbls")
Signed-off-by: liaoguojia <liaoguojia@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The input parameters may not be reliable. Before using the
queue id, we should check this parameter. Otherwise, memory
overwriting may occur.
Fixes: d341001846 ("net: hns3: refactor the mailbox message between PF and VF")
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
vport_id include PF and VFs, vport_id = 0 means PF, other values mean VFs.
So the actual vf id is equal to vport_id minus 1.
Some VF print logs are actually vport, and logs of vf id actually use
vport id, so this patch fixes them.
Fixes: ac887be5b0 ("net: hns3: change print level of RAS error log from warning to error")
Fixes: adcf738b80 ("net: hns3: cleanup some print format warning")
Signed-off-by: Jiaran Zhang <zhangjiaran@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The vf id from ethtool is added 1 before configured to driver.
So it's necessary to minus 1 when printing it, in order to
keep consistent with user's configuration.
Fixes: dd74f815dd ("net: hns3: Add support for rule add/delete for flow director")
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When user change rss 'hfunc' without set rss 'hkey' by ethtool
-X command, the driver will ignore the 'hfunc' for the hkey is
NULL. It's unreasonable. So fix it.
Fixes: 46a3df9f97 ("net: hns3: Add HNS3 Acceleration Engine & Compatibility Layer Support")
Fixes: 374ad29176 ("net: hns3: Add RSS general configuration support for VF")
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Adding back the nonstandard ioctl commands caused -Wrestrict warnings
when building with 'make W=1':
drivers/staging/r8188eu/os_dep/ioctl_linux.c: In function 'rtw_mp_read_rf':
drivers/staging/r8188eu/os_dep/ioctl_linux.c:5515:27: error: 'sprintf' argument 3 overlaps destination object 'extra' [-Werror=restrict]
5515 | sprintf(extra, "%s %d", extra, strtou);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/staging/r8188eu/os_dep/ioctl_linux.c:5470:54: note: destination object referenced by 'restrict'-qualified argument 1 was declared here
5470 | struct iw_point *wrqu, char *extra)
| ~~~~~~^~~~~
Change these to the same construct used elsewhere in that driver,
with an offset to the string to make the warning go away.
The ioctl commands were previously removed, and it's unlikely that
anything is actually using them, so ideally I would prefer to have
them removed again.
The lack of range checking of the 'extra' output buffer is also
slightly worrying, but I did not check whether this could cause
harm.
Fixes: 2b42bd58b3 ("staging: r8188eu: introduce new os_dep dir for RTL8188eu driver")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20210920095525.1150678-1-arnd@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Without CONFIG_COMMON_CLK, this fails to link:
arm-linux-gnueabi-ld: drivers/ptp/ptp_ocp.o: in function `ptp_ocp_register_i2c':
ptp_ocp.c:(.text+0xcc0): undefined reference to `__clk_hw_register_fixed_rate'
arm-linux-gnueabi-ld: ptp_ocp.c:(.text+0xcf4): undefined reference to `devm_clk_hw_register_clkdev'
arm-linux-gnueabi-ld: drivers/ptp/ptp_ocp.o: in function `ptp_ocp_detach':
ptp_ocp.c:(.text+0x1c24): undefined reference to `clk_hw_unregister_fixed_rate'
Fixes: a7e1abad13 ("ptp: Add clock driver for the OpenCompute TimeCard.")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
0xac24 device ID is already defined and used via
BANDB_DEVICE_ID_USO9ML2_4. Remove the duplicate from the list.
Fixes: 27f1281d5f ("USB: serial: Extra device/vendor ID for mos7840 driver")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Cc: stable@vger.kernel.org
Signed-off-by: Johan Hovold <johan@kernel.org>
The smallest TX ring size we support must fit a TX SKB with MAX_SKB_FRAGS
+ 1. Because the first TX BD for a packet is always a long TX BD, we
need an extra TX BD to fit this packet. Define BNXT_MIN_TX_DESC_CNT with
this value to make this more clear. The current code uses a minimum
that is off by 1. Fix it using this constant.
The tx_wake_thresh to determine when to wake up the TX queue is half the
ring size but we must have at least BNXT_MIN_TX_DESC_CNT for the next
packet which may have maximum fragments. So the comparison of the
available TX BDs with tx_wake_thresh should be >= instead of > in the
current code. Otherwise, at the smallest ring size, we will never wake
up the TX queue and will cause TX timeout.
Fixes: c0c050c58d ("bnxt_en: New Broadcom ethernet driver.")
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadocm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The resilient nexthop group torture tests in fib_nexthop.sh exposed a
possible division by zero while replacing a resilient group [1]. The
division by zero occurs when the data path sees a resilient nexthop
group with zero buckets.
The tests replace a resilient nexthop group in a loop while traffic is
forwarded through it. The tests do not specify the number of buckets
while performing the replacement, resulting in the kernel allocating a
stub resilient table (i.e, 'struct nh_res_table') with zero buckets.
This table should never be visible to the data path, but the old nexthop
group (i.e., 'oldg') might still be used by the data path when the stub
table is assigned to it.
Fix this by only assigning the stub table to the old nexthop group after
making sure the group is no longer used by the data path.
Tested with fib_nexthops.sh:
Tests passed: 222
Tests failed: 0
[1]
divide error: 0000 [#1] PREEMPT SMP KASAN
CPU: 0 PID: 1850 Comm: ping Not tainted 5.14.0-custom-10271-ga86eb53057fe #1107
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-4.fc34 04/01/2014
RIP: 0010:nexthop_select_path+0x2d2/0x1a80
[...]
Call Trace:
fib_select_multipath+0x79b/0x1530
fib_select_path+0x8fb/0x1c10
ip_route_output_key_hash_rcu+0x1198/0x2da0
ip_route_output_key_hash+0x190/0x340
ip_route_output_flow+0x21/0x120
raw_sendmsg+0x91d/0x2e10
inet_sendmsg+0x9e/0xe0
__sys_sendto+0x23d/0x360
__x64_sys_sendto+0xe1/0x1b0
do_syscall_64+0x35/0x80
entry_SYSCALL_64_after_hwframe+0x44/0xae
Cc: stable@vger.kernel.org
Fixes: 283a72a559 ("nexthop: Add implementation of resilient next-hop groups")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The process will cause napi.state to contain NAPI_STATE_SCHED and
not in the poll_list, which will cause napi_disable() to get stuck.
The prefix "NAPI_STATE_" is removed in the figure below, and
NAPI_STATE_HASHED is ignored in napi.state.
CPU0 | CPU1 | napi.state
===============================================================================
napi_disable() | | SCHED | NPSVC
napi_enable() | |
{ | |
smp_mb__before_atomic(); | |
clear_bit(SCHED, &n->state); | | NPSVC
| napi_schedule_prep() | SCHED | NPSVC
| napi_poll() |
| napi_complete_done() |
| { |
| if (n->state & (NPSVC | | (1)
| _BUSY_POLL))) |
| return false; |
| ................ |
| } | SCHED | NPSVC
| |
clear_bit(NPSVC, &n->state); | | SCHED
} | |
| |
napi_schedule_prep() | | SCHED | MISSED (2)
(1) Here return direct. Because of NAPI_STATE_NPSVC exists.
(2) NAPI_STATE_SCHED exists. So not add napi.poll_list to sd->poll_list
Since NAPI_STATE_SCHED already exists and napi is not in the
sd->poll_list queue, NAPI_STATE_SCHED cannot be cleared and will always
exist.
1. This will cause this queue to no longer receive packets.
2. If you encounter napi_disable under the protection of rtnl_lock, it
will cause the entire rtnl_lock to be locked, affecting the overall
system.
This patch uses cmpxchg to implement napi_enable(), which ensures that
there will be no race due to the separation of clear two bits.
Fixes: 2d8bff1269 ("netpoll: Close race condition between poll_one_napi and napi_disable")
Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Reviewed-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
FPGA Manager fixes for 5.15
Tom and Jiapeng's fixes address smatch warnings around missing return
values in error cases.
Russ' change addresses an issue where registers are being accessed too
early resulting in invalid data being read.
All patches have been reviewed on the mailing list, and have been in the
last few linux-next releases (as part of my fixes branch) without issues.
Signed-off-by: Moritz Fischer <mdf@kernel.org>
* tag 'fpga-fixes-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/mdf/linux-fpga:
fpga: dfl: Avoid reads to AFU CSRs during enumeration
fpga: machxo2-spi: Fix missing error code in machxo2_write_complete()
fpga: machxo2-spi: Return an error on failure
Oded writes:
This tag contains the following fixes for 5.15-rc3:
- Fix potential race when user waiting for interrupt ioctl
- Prevent possible kernel oops in staged CS ioctl
- Use direct MSI mechanism in Gaudi as a WA for a H/W issue
regarding FLR
- Don't support collective wait ioctl operation when it
is not supported. e.g. when the NIC ports are disabled
- Fix configuration of one of the security mechanism.
- Change error print to be rate-limited as it can be initiated
by the user and spam the kernel log
- Fix return value of CS ioctl when doing staged CS
- Fix CS ioctl code when user doesn't supply an offset for
the memory area that we use as fence.
- Spelling mistake fix
* tag 'misc-habanalabs-fixes-2021-09-19' of https://git.kernel.org/pub/scm/linux/kernel/git/ogabbay/linux:
habanalabs: expose a single cs seq in staged submissions
habanalabs: fix wait offset handling
habanalabs: rate limit multi CS completion errors
habanalabs/gaudi: fix LBW RR configuration
habanalabs: Fix spelling mistake "FEADBACK" -> "FEEDBACK"
habanalabs: fail collective wait when not supported
habanalabs/gaudi: use direct MSI in single mode
habanalabs: fix kernel OOPs related to staged cs
habanalabs: fix potential race in interrupt wait ioctl
Attempt to mount 9p file system as root gives the following kernel panic:
9pnet_virtio: no channels available for device root
Kernel panic - not syncing: VFS: Unable to mount root "root" (9p), err=-2
CPU: 2 PID: 1 Comm: swapper/0 Not tainted 5.15.0-rc1+ #127
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014
Call Trace:
dump_stack_lvl+0x45/0x59
panic+0x1e2/0x44b
? __warn_printk+0xf3/0xf3
? free_unref_page+0x2d4/0x4a0
? trace_hardirqs_on+0x32/0x120
? free_unref_page+0x2d4/0x4a0
mount_root+0x189/0x1e0
prepare_namespace+0x136/0x165
kernel_init_freeable+0x3b8/0x3cb
? rest_init+0x2e0/0x2e0
kernel_init+0x19/0x130
ret_from_fork+0x1f/0x30
Kernel Offset: disabled
---[ end Kernel panic - not syncing: VFS: Unable to mount root "root" (9p), err=-2 ]---
QEMU command line:
"qemu-system-x86_64 -append root=/dev/root rw rootfstype=9p rootflags=trans=virtio ..."
This error is because root_device_name is truncated in prepare_namespace() from
being "/dev/root" to be "root" prior to call to mount_nodev_root().
As a solution, don't treat errors in mount_nodev_root() as errors that
require panics and allow failback to the mount flow that existed before
patch citied in Fixes tag.
Fixes: f9259be6a9 ("init: allow mounting arbitrary non-blockdevice filesystems as root")
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
split_fs_names() currently takes comma separate list of filesystems
and converts it into individual filesystem strings. Pleaces these
strings in the input buffer passed by caller and returns number of
strings.
If caller manages to pass input string bigger than buffer, then we
can write beyond the buffer. Or if string just fits buffer, we will
still write beyond the buffer as we append a '\0' byte at the end.
Pass size of input buffer to split_fs_names() and put enough checks
in place so such buffer overrun possibilities do not occur.
This patch does few things.
- Add a parameter "size" to split_fs_names(). This specifies size
of input buffer.
- Use strlcpy() (instead of strcpy()) so that we can't go beyond
buffer size. If input string "names" is larger than passed in
buffer, input string will be truncated to fit in buffer.
- Stop appending extra '\0' character at the end and avoid one
possibility of going beyond the input buffer size.
- Do not use extra loop to count number of strings.
- Previously if one passed "rootfstype=foo,,bar", split_fs_names()
will return only 1 string "foo" (and "bar" will be truncated
due to extra ,). After this patch, now split_fs_names() will
return 3 strings ("foo", zero-sized-string, and "bar").
Callers of split_fs_names() have been modified to check for
zero sized string and skip to next one.
Reported-by: xu xin <xu.xin16@zte.com.cn>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Nathan Chancellor reports that the recent change to pci_iounmap in
commit 9caea00076 ("parisc: Declare pci_iounmap() parisc version only
when CONFIG_PCI enabled") causes build errors on arm64.
It took me about two hours to convince myself that I think I know what
the logic of that mess of #ifdef's in the <asm-generic/io.h> header file
really aim to do, and rewrite it to be easier to follow.
Famous last words.
Anyway, the code has now been lifted from that grotty header file into
lib/pci_iomap.c, and has fairly extensive comments about what the logic
is. It also avoids indirecting through another confusing (and badly
named) helper function that has other preprocessor config conditionals.
Let's see what odd architecture did something else strange in this area
to break things. But my arm64 cross build is clean.
Fixes: 9caea00076 ("parisc: Declare pci_iounmap() parisc version only when CONFIG_PCI enabled")
Reported-by: Nathan Chancellor <nathan@kernel.org>
Cc: Helge Deller <deller@gmx.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Guenter Roeck <linux@roeck-us.net>
Cc: Ulrich Teichert <krypton@ulrich-teichert.org>
Cc: James Bottomley <James.Bottomley@hansenpartnership.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull x86 fixes from Borislav Petkov:
- Prevent a infinite loop in the MCE recovery on return to user space,
which was caused by a second MCE queueing work for the same page and
thereby creating a circular work list.
- Make kern_addr_valid() handle existing PMD entries, which are marked
not present in the higher level page table, correctly instead of
blindly dereferencing them.
- Pass a valid address to sanitize_phys(). This was caused by the
mixture of inclusive and exclusive ranges. memtype_reserve() expect
'end' being exclusive, but sanitize_phys() wants it inclusive. This
worked so far, but with end being the end of the physical address
space the fail is exposed.
- Increase the maximum supported GPIO numbers for 64bit. Newer SoCs
exceed the previous maximum.
* tag 'x86_urgent_for_v5.15_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/mce: Avoid infinite loop for copy from user recovery
x86/mm: Fix kern_addr_valid() to cope with existing but not present entries
x86/platform: Increase maximum GPIO number for X86_64
x86/pat: Pass valid address to sanitize_phys()
Pull perf event fix from Thomas Gleixner:
"A single fix for the perf core where a value read with READ_ONCE() was
checked and then reread which makes all the checks invalid. Reuse the
already read value instead"
* tag 'perf-urgent-2021-09-19' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
events: Reuse value read using READ_ONCE instead of re-reading it
Pull locking fixes from Thomas Gleixner:
"A set of updates for the RT specific reader/writer locking base code:
- Make the fast path reader ordering guarantees correct.
- Code reshuffling to make the fix simpler"
[ This plays ugly games with atomic_add_return_release() because we
don't have a plain atomic_add_release(), and should really be cleaned
up, I think - Linus ]
* tag 'locking-urgent-2021-09-19' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
locking/rwbase: Take care of ordering guarantee for fastpath reader
locking/rwbase: Extract __rwbase_write_trylock()
locking/rwbase: Properly match set_and_save_state() to restore_state()
Pull powerpc fixes from Michael Ellerman:
- Fix crashes when scv (System Call Vectored) is used to make a syscall
when a transaction is active, on Power9 or later.
- Fix bad interactions between rfscv (Return-from scv) and Power9
fake-suspend mode.
- Fix crashes when handling machine checks in LPARs using the Hash MMU.
- Partly revert a recent change to our XICS interrupt controller code,
which broke the recently added Microwatt support.
Thanks to Cédric Le Goater, Eirik Fuller, Ganesh Goudar, Gustavo Romero,
Joel Stanley, Nicholas Piggin.
* tag 'powerpc-5.15-2' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
powerpc/xics: Set the IRQ chip data for the ICS native backend
powerpc/mce: Fix access error in mce handler
KVM: PPC: Book3S HV: Tolerate treclaim. in fake-suspend mode changing registers
powerpc/64s: system call rfscv workaround for TM bugs
selftests/powerpc: Add scv versions of the basic TM syscall tests
powerpc/64s: system call scv tabort fix for corrupt irq soft-mask state
Pull Kbuild fixes from Masahiro Yamada:
- Fix bugs in checkkconfigsymbols.py
- Fix missing sys import in gen_compile_commands.py
- Fix missing FORCE warning for ARCH=sh builds
- Fix -Wignored-optimization-argument warnings for Clang builds
- Turn -Wignored-optimization-argument into an error in order to stop
building instead of sprinkling warnings
* tag 'kbuild-fixes-v5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
kbuild: Add -Werror=ignored-optimization-argument to CLANG_FLAGS
x86/build: Do not add -falign flags unconditionally for clang
kbuild: Fix comment typo in scripts/Makefile.modpost
sh: Add missing FORCE prerequisites in Makefile
gen_compile_commands: fix missing 'sys' package
checkkconfigsymbols.py: Remove skipping of help lines in parse_kconfig_file
checkkconfigsymbols.py: Forbid passing 'HEAD' to --commit
Pull perf tools fixes from Arnaldo Carvalho de Melo:
- Fix ip display in 'perf script' when output type != attr->type.
- Ignore deprecation warning when using libbpf'sg btf__get_from_id(),
fixing the build with libbpf v0.6+.
- Make use of FD() robust in libperf, fixing a segfault with 'perf stat
--iostat list'.
- Initialize addr_location:srcline pointer to NULL when resolving
callchain addresses.
- Fix fused instruction logic for assembly functions in 'perf
annotate'.
* tag 'perf-tools-fixes-for-v5.15-2021-09-18' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux:
perf bpf: Ignore deprecation warning when using libbpf's btf__get_from_id()
libperf evsel: Make use of FD robust.
perf machine: Initialize srcline string member in add_location struct
perf script: Fix ip display when type != attr->type
perf annotate: Fix fused instr logic for assembly functions
The old dmascc driver depends on the legacy ISA_DMA_API, and blindly
just casts the kernel virtual address to 'int' for set_dma_addr().
That works only incidentally, and because the high bits of the address
will be ignored anyway. And on 64-bit architectures it causes warnings.
Admittedly, 64-bit architectures with ISA are basically dead - I think
the only example of this is alpha, and nobody would ever use the dmascc
driver there. But hey, the fix is easy enough, the end result is
cleaner, and it's yet another configuration that now builds without
warnings.
If somebody actually uses this driver on an alpha and this fixes it for
you, please email me. Because that is just incredibly bizarre.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
With the previous commit (9caea00076: "parisc: Declare pci_iounmap()
parisc version only when CONFIG_PCI enabled") we can now enable
GENERIC_PCI_IOMAP unconditionally on alpha, and if PCI is not enabled we
will just get the nice empty helper functions that allow mixed-bus
drivers to build.
Example driver: the old 3com/3c59x.c driver works with either the PCI or
the EISA version of the 3x59x card, but wouldn't build in an EISA-only
configuration because of missing pci_iomap() and pci_iounmap() dummy
wrappers.
Most of the other PCI infrastructure just becomes empty wrappers even
without GENERIC_PCI_IOMAP, and it's not obvious that the pci_iomap
functionality shouldn't do the same, but this works.
Cc: Ulrich Teichert <krypton@ulrich-teichert.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Makefile uses TEST_PROGS instead of TEST_GEN_PROGS to define
executables. TEST_PROGS is for shell scripts that need to be
installed and run by the common lib.mk framework. The common
framework doesn't touch TEST_PROGS when it does build and clean.
As a result "make kselftest-clean" and "make clean" fail to remove
executables. Run and install work because the common framework runs
and installs TEST_PROGS. Build works because the Makefile defines
"all" rule which is unnecessary if TEST_GEN_PROGS is used.
Use TEST_GEN_PROGS so the common framework can handle build/run/
install/clean properly.
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Any link state change that's done prior to net device registration
isn't reflected on the state, thus the operational state is left
obsolete, with 'UNKNOWN' status.
To resolve the issue, query link state from FW upon open operations
to ensure operational state is updated.
Fixes: c27a02cd94 ("mlx4_en: Add driver for Mellanox ConnectX 10GbE NIC")
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix the args to fprintf(). Splitting the message ends up passing
incorrect arg for "sigurg %d" and an extra arg overall. The test
result message ends up incorrect.
test_unix_oob.c: In function ‘main’:
test_unix_oob.c:274:43: warning: format ‘%d’ expects argument of type ‘int’, but argument 3 has type ‘char *’ [-Wformat=]
274 | fprintf(stderr, "Test 3 failed, sigurg %d len %d OOB %c ",
| ~^
| |
| int
| %s
275 | "atmark %d\n", signal_recvd, len, oob, atmark);
| ~~~~~~~~~~~~~
| |
| char *
test_unix_oob.c:274:19: warning: too many arguments for format [-Wformat-extra-args]
274 | fprintf(stderr, "Test 3 failed, sigurg %d len %d OOB %c ",
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Due to the inclusion of nvmem handling into the mac-address getter
function of_get_mac_address() by
commit d01f449c00 ("of_net: add NVMEM support to of_get_mac_address")
it is now possible to get a -EPROBE_DEFER return code. Which did cause
bgmac to assign a random ethernet address.
This exact issue happened on my Meraki MR32. The nvmem provider is
an EEPROM (at24c64) which gets instantiated once the module
driver is loaded... This happens once the filesystem becomes available.
With this patch, bgmac_probe() will propagate the -EPROBE_DEFER error.
Then the driver subsystem will reschedule the probe at a later time.
Cc: Petr Štetiar <ynezz@true.cz>
Cc: Michael Walle <michael@walle.cc>
Fixes: d01f449c00 ("of_net: add NVMEM support to of_get_mac_address")
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 86f8b1c01a ("net: dsa: Do not make user port errors fatal")
decided it was fine to ignore errors on certain ports that fail to
probe, and go on with the ports that do probe fine.
Commit fb6ec87f72 ("net: dsa: Fix type was not set for devlink port")
noticed that devlink_port_type_eth_set(dlp, dp->slave); does not get
called, and devlink notices after a timeout of 3600 seconds and prints a
WARN_ON. So it went ahead to unregister the devlink port. And because
there exists an UNUSED port flavour, we actually re-register the devlink
port as UNUSED.
Commit 08156ba430 ("net: dsa: Add devlink port regions support to
DSA") added devlink port regions, which are set up by the driver and not
by DSA.
When we trigger the devlink port deregistration and reregistration as
unused, devlink now prints another WARN_ON, from here:
devlink_port_unregister:
WARN_ON(!list_empty(&devlink_port->region_list));
So the port still has regions, which makes sense, because they were set
up by the driver, and the driver doesn't know we're unregistering the
devlink port.
Somebody needs to tear them down, and optionally (actually it would be
nice, to be consistent) set them up again for the new devlink port.
But DSA's layering stays in our way quite badly here.
The options I've considered are:
1. Introduce a function in devlink to just change a port's type and
flavour. No dice, devlink keeps a lot of state, it really wants the
port to not be registered when you set its parameters, so changing
anything can only be done by destroying what we currently have and
recreating it.
2. Make DSA cache the parameters passed to dsa_devlink_port_region_create,
and the region returned, keep those in a list, then when the devlink
port unregister needs to take place, the existing devlink regions are
destroyed by DSA, and we replay the creation of new regions using the
cached parameters. Problem: mv88e6xxx keeps the region pointers in
chip->ports[port].region, and these will remain stale after DSA frees
them. There are many things DSA can do, but updating mv88e6xxx's
private pointers is not one of them.
3. Just let the driver do it (i.e. introduce a very specific method
called ds->ops->port_reinit_as_unused, which unregisters its devlink
port devlink regions, then the old devlink port, then registers the
new one, then the devlink port regions for it). While it does work,
as opposed to the others, it's pretty horrible from an API
perspective and we can do better.
4. Introduce a new pair of methods, ->port_setup and ->port_teardown,
which in the case of mv88e6xxx must register and unregister the
devlink port regions. Call these 2 methods when the port must be
reinitialized as unused.
Naturally, I went for the 4th approach.
Fixes: 08156ba430 ("net: dsa: Add devlink port regions support to DSA")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The MODULE_DEVICE_TABLE already creates proper alias for platform
driver. Having another MODULE_ALIAS causes the alias to be duplicated.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Colin Foster says:
====================
ocelot phylink fixes
When the ocelot driver was migrated to phylink, e6e12df625 ("net:
mscc: ocelot: convert to phylink") there were two additional writes to
registers that became stale. One write was to DEV_CLOCK_CFG and one was
to ANA_PFC_PCF_CFG.
Both of these writes referenced the variable "speed" which originally
was set to OCELOT_SPEED_{10,100,1000,2500}. These macros expand to
values of 3, 2, 1, or 0, respectively. After the update, the variable
speed is set to SPEED_{10,100,1000,2500} which expand to 10, 100, 1000,
and 2500. So invalid values were getting written to the two registers,
which would lead to either a lack of functionality or undefined
funcationality.
Fixing these values was the intent of v1 of this patch set - submitted
as "[PATCH v1 net] net: ethernet: mscc: ocelot: bug fix when writing MAC
speed"
During that review it was determined that both writes were actually
unnecessary. DEV_CLOCK_CFG is a duplicate write, so can be removed
entirely. This was accidentally submitted as as a new, lone patch titled
"[PATCH v1 net] net: mscc: ocelot: remove buggy duplicate write to
DEV_CLOCK_CFG". This is part of what is considered v2 of this patch set.
Additionally, the write to ANA_PFC_PFC_CFG is also unnecessary. Priority
flow contol is disabled, so configuring it is useless and should be
removed. This was also submitted as a new, lone patch titled "[PATCH v1
net] net: mscc: ocelot: remove buggy and useless write to ANA_PFC_PFC_CFG".
This is the rest of what is considered v2 of this patch set.
v3
Identical to v2, but fixes the patch numbering to v3 and submitting the
two changes as a patch set.
v2
Note: I misunderstood and submitted two new "v1" patches instead of a
single "v2" patch set.
- Remove the buggy writes altogher
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
When updating ocelot to use phylink, a second write to DEV_CLOCK_CFG was
mistakenly left in. It used the variable "speed" which, previously, would
would have been assigned a value of OCELOT_SPEED_1000. In phylink the
variable is be SPEED_1000, which is invalid for the
DEV_CLOCK_LINK_SPEED macro. Removing it as unnecessary and buggy.
Fixes: e6e12df625 ("net: mscc: ocelot: convert to phylink")
Signed-off-by: Colin Foster <colin.foster@in-advantage.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A useless write to ANA_PFC_PFC_CFG was left in while refactoring ocelot to
phylink. Since priority flow control is disabled, writing the speed has no
effect.
Further, it was using ethtool.h SPEED_ instead of OCELOT_SPEED_ macros,
which are incorrectly offset for GENMASK.
Lastly, for priority flow control to properly function, some scenarios
would rely on the rate adaptation from the PCS while the MAC speed would
be fixed. So it isn't used, and even if it was, neither "speed" nor
"mac_speed" are necessarily the correct values to be used.
Fixes: e6e12df625 ("net: mscc: ocelot: convert to phylink")
Signed-off-by: Colin Foster <colin.foster@in-advantage.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
lock_sock_fast() and lock_sock_nested() contain lockdep annotations for the
sock::sk_lock.owned 'mutex'. sock::sk_lock.owned is not a regular mutex. It
is just lockdep wise equivalent. In fact it's an open coded trivial mutex
implementation with some interesting features.
sock::sk_lock.slock is a regular spinlock protecting the 'mutex'
representation sock::sk_lock.owned which is a plain boolean. If 'owned' is
true, then some other task holds the 'mutex', otherwise it is uncontended.
As this locking construct is obviously endangered by lock ordering issues as
any other locking primitive it got lockdep annotated via a dedicated
dependency map sock::sk_lock.dep_map which has to be updated at the lock
and unlock sites.
lock_sock_nested() is a straight forward 'mutex' lock operation:
might_sleep();
spin_lock_bh(sock::sk_lock.slock)
while (!try_lock(sock::sk_lock.owned)) {
spin_unlock_bh(sock::sk_lock.slock);
wait_for_release();
spin_lock_bh(sock::sk_lock.slock);
}
The lockdep annotation for sock::sk_lock.owned is for unknown reasons
_after_ the lock has been acquired, i.e. after the code block above and
after releasing sock::sk_lock.slock, but inside the bottom halves disabled
region:
spin_unlock(sock::sk_lock.slock);
mutex_acquire(&sk->sk_lock.dep_map, subclass, 0, _RET_IP_);
local_bh_enable();
The placement after the unlock is obvious because otherwise the
mutex_acquire() would nest into the spin lock held region.
But that's from the lockdep perspective still the wrong place:
1) The mutex_acquire() is issued _after_ the successful acquisition which
is pointless because in a dead lock scenario this point is never
reached which means that if the deadlock is the first instance of
exposing the wrong lock order lockdep does not have a chance to detect
it.
2) It only works because lockdep is rather lax on the context from which
the mutex_acquire() is issued. Acquiring a mutex inside a bottom halves
and therefore non-preemptible region is obviously invalid, except for a
trylock which is clearly not the case here.
This 'works' stops working on RT enabled kernels where the bottom halves
serialization is done via a local lock, which exposes this misplacement
because the 'mutex' and the local lock nest the wrong way around and
lockdep complains rightfully about a lock inversion.
The placement is wrong since the initial commit a5b5bb9a05 ("[PATCH]
lockdep: annotate sk_locks") which introduced this.
Fix it by moving the mutex_acquire() in front of the actual lock
acquisition, which is what the regular mutex_lock() operation does as well.
lock_sock_fast() is not that straight forward. It looks at the first glance
like a convoluted trylock operation:
spin_lock_bh(sock::sk_lock.slock)
if (!sock::sk_lock.owned)
return false;
while (!try_lock(sock::sk_lock.owned)) {
spin_unlock_bh(sock::sk_lock.slock);
wait_for_release();
spin_lock_bh(sock::sk_lock.slock);
}
spin_unlock(sock::sk_lock.slock);
mutex_acquire(&sk->sk_lock.dep_map, subclass, 0, _RET_IP_);
local_bh_enable();
return true;
But that's not the case: lock_sock_fast() is an interesting optimization
for short critical sections which can run with bottom halves disabled and
sock::sk_lock.slock held. This allows to shortcut the 'mutex' operation in
the non contended case by preventing other lockers to acquire
sock::sk_lock.owned because they are blocked on sock::sk_lock.slock, which
in turn avoids the overhead of doing the heavy processing in release_sock()
including waking up wait queue waiters.
In the contended case, i.e. when sock::sk_lock.owned == true the behavior
is the same as lock_sock_nested().
Semantically this shortcut means, that the task acquired the 'mutex' even
if it does not touch the sock::sk_lock.owned field in the non-contended
case. Not telling lockdep about this shortcut acquisition is hiding
potential lock ordering violations in the fast path.
As a consequence the same reasoning as for the above lock_sock_nested()
case vs. the placement of the lockdep annotation applies.
The current placement of the lockdep annotation was just copied from
the original lock_sock(), now renamed to lock_sock_nested(),
implementation.
Fix this by moving the mutex_acquire() in front of the actual lock
acquisition and adding the corresponding mutex_release() into
unlock_sock_fast(). Also document the fast path return case with a comment.
Reported-by: Sebastian Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: netdev@vger.kernel.org
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The file sja1105.txt was converted to nxp,sja1105.yaml.
Signed-off-by: Alejandro Concepcion-Rodriguez <asconcepcion@acoro.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>
When IGC=y and PTP_1588_CLOCK=m, the ptp_*() interface family is
not available to the igc driver. Make this driver depend on
PTP_1588_CLOCK_OPTIONAL so that it will build without errors.
Various igc commits have used ptp_*() functions without checking
that PTP_1588_CLOCK is enabled. Fix all of these here.
Fixes these build errors:
ld: drivers/net/ethernet/intel/igc/igc_main.o: in function `igc_msix_other':
igc_main.c:(.text+0x6494): undefined reference to `ptp_clock_event'
ld: igc_main.c:(.text+0x64ef): undefined reference to `ptp_clock_event'
ld: igc_main.c:(.text+0x6559): undefined reference to `ptp_clock_event'
ld: drivers/net/ethernet/intel/igc/igc_ethtool.o: in function `igc_ethtool_get_ts_info':
igc_ethtool.c:(.text+0xc7a): undefined reference to `ptp_clock_index'
ld: drivers/net/ethernet/intel/igc/igc_ptp.o: in function `igc_ptp_feature_enable_i225':
igc_ptp.c:(.text+0x330): undefined reference to `ptp_find_pin'
ld: igc_ptp.c:(.text+0x36f): undefined reference to `ptp_find_pin'
ld: drivers/net/ethernet/intel/igc/igc_ptp.o: in function `igc_ptp_init':
igc_ptp.c:(.text+0x11cd): undefined reference to `ptp_clock_register'
ld: drivers/net/ethernet/intel/igc/igc_ptp.o: in function `igc_ptp_stop':
igc_ptp.c:(.text+0x12dd): undefined reference to `ptp_clock_unregister'
ld: drivers/platform/x86/dell/dell-wmi-privacy.o: in function `dell_privacy_wmi_probe':
Fixes: 64433e5bf4 ("igc: Enable internal i225 PPS")
Fixes: 60dbede0c4 ("igc: Add support for ethtool GET_TS_INFO command")
Fixes: 87938851b6 ("igc: enable auxiliary PHC functions for the i225")
Fixes: 5f2958052c ("igc: Add basic skeleton for PTP")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Ederson de Souza <ederson.desouza@intel.com>
Cc: Tony Nguyen <anthony.l.nguyen@intel.com>
Cc: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Cc: intel-wired-lan@lists.osuosl.org
Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The only struct dim_sample member that does not get
initialized by dim_update_sample() is comp_ctr. (There
is special API to initialize comp_ctr:
dim_update_sample_with_comps(), and it is currently used
only for RDMA.) comp_ctr is used to compute curr_stats->cmps
and curr_stats->cpe_ratio (see dim_calc_stats()) which in
turn are consumed by the rdma_dim_*() API. Therefore,
functionally, the net_dim*() API consumers are not affected.
Nevertheless, fix the computation of statistics based
on an uninitialized variable, even if the mentioned statistics
are not used at the moment.
Fixes: ae0e6a5d16 ("enetc: Add adaptive interrupt coalescing")
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
irq_set_affinity_hit() stores a reference to the cpumask_t
parameter in the irq descriptor, and that reference can be
accessed later from irq_affinity_hint_proc_show(). Since
the cpu_mask parameter passed to irq_set_affinity_hit() has
only temporary storage (it's on the stack memory), later
accesses to it are illegal. Thus reads from the corresponding
procfs affinity_hint file can result in paging request oops.
The issue is fixed by the get_cpu_mask() helper, which provides
a permanent storage for the cpumask_t parameter.
Fixes: d4fd0404c1 ("enetc: Introduce basic PF and VF ENETC ethernet drivers")
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We try to use build_skb() if we had sufficient tailroom. But we forget
to release the unused pages chained via private in big mode which will
leak pages. Fixing this by release the pages after building the skb in
big mode.
Cc: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Fixes: fb32856b16 ("virtio-net: page_to_skb() use build_skb when there's sufficient tailroom")
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When re-entering the main loop of xenvif_tx_check_gop() a 2nd time, the
special considerations for the head of the SKB no longer apply. Don't
mistakenly report ERROR to the frontend for the first entry in the list,
even if - from all I can tell - this shouldn't matter much as the overall
transmit will need to be considered failed anyway.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Paul Durrant <paul@xen.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean says:
====================
Make DSA switch drivers compatible with masters which unregister on shutdown
Changes in v2:
- fix build for b53_mmap
- use unregister_netdevice_many
It was reported by Lino here:
https://lore.kernel.org/netdev/20210909095324.12978-1-LinoSanfilippo@gmx.de/
that when the DSA master attempts to unregister its net_device on
shutdown, DSA should prevent that operation from succeeding because it
holds a reference to it. This hangs the shutdown process.
This issue was essentially introduced in commit 2f1e8ea726 ("net: dsa:
link interfaces with the DSA master to get rid of lockdep warnings").
The present series patches all DSA drivers to handle that case,
depending on whether those drivers were introduced before or after the
offending commit, a different Fixes: tag is specified for them.
The approach taken by this series solves the issue in essentially the
same way as Lino's patches, except for three key differences:
- this series takes a more minimal approach in what is done on shutdown,
we do not attempt a full tree teardown as that is not strictly
necessary. I might revisit this if there are compelling reasons to do
otherwise
- this series fixes the issues for all DSA drivers, not just KSZ9897
- this series works even if the ->remove driver method gets called for
the same device too, not just ->shutdown. This is really possible to
happen for SPI device drivers, and potentially possible for other bus
device drivers too.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit 2f1e8ea726 ("net: dsa: link interfaces with the DSA
master to get rid of lockdep warnings"), DSA gained a requirement which
it did not fulfill, which is to unlink itself from the DSA master at
shutdown time.
Since the Arrow SpeedChips XRS700x driver was introduced after the bad
commit, it has never worked with DSA masters which decide to unregister
their net_device on shutdown, effectively hanging the reboot process.
To fix that, we need to call dsa_switch_shutdown.
These devices can be connected by I2C or by MDIO, and if I search for
I2C or MDIO bus drivers that implement their ->shutdown by redirecting
it to ->remove I don't see any, however this does not mean it would not
be possible. To be compatible with that pattern, it is necessary to
implement an "if this then not that" scheme, to avoid ->remove and
->shutdown from being called both for the same struct device.
Fixes: ee00b24f32 ("net: dsa: add Arrow SpeedChips XRS700x driver")
Link: https://lore.kernel.org/netdev/20210909095324.12978-1-LinoSanfilippo@gmx.de/
Reported-by: Lino Sanfilippo <LinoSanfilippo@gmx.de>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: George McCollister <george.mccollister@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit 2f1e8ea726 ("net: dsa: link interfaces with the DSA
master to get rid of lockdep warnings"), DSA gained a requirement which
it did not fulfill, which is to unlink itself from the DSA master at
shutdown time.
Since the Microchip sub-driver for KSZ8863 was introduced after the bad
commit, it has never worked with DSA masters which decide to unregister
their net_device on shutdown, effectively hanging the reboot process.
To fix that, we need to call dsa_switch_shutdown.
Since this driver expects the MDIO bus to be backed by mdio_bitbang, I
don't think there is currently any MDIO bus driver which implements its
->shutdown by redirecting it to ->remove, but in any case, to be
compatible with that pattern, it is necessary to implement an "if this
then not that" scheme, to avoid ->remove and ->shutdown from being
called both for the same struct device.
Fixes: 60a3647600 ("net: dsa: microchip: Add Microchip KSZ8863 SMI based driver support")
Link: https://lore.kernel.org/netdev/20210909095324.12978-1-LinoSanfilippo@gmx.de/
Reported-by: Lino Sanfilippo <LinoSanfilippo@gmx.de>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit 2f1e8ea726 ("net: dsa: link interfaces with the DSA
master to get rid of lockdep warnings"), DSA gained a requirement which
it did not fulfill, which is to unlink itself from the DSA master at
shutdown time.
Since the hellcreek driver was introduced after the bad commit, it has
never worked with DSA masters which decide to unregister their
net_device on shutdown, effectively hanging the reboot process.
Hellcreek is a platform device driver, so we probably cannot have the
oddities of ->shutdown and ->remove getting both called for the exact
same struct device. But to be in line with the pattern from the other
device drivers which are on slow buses, implement the same "if this then
not that" pattern of either running the ->shutdown or the ->remove hook.
The driver's current ->remove implementation makes that very easy
because it already zeroes out its device_drvdata on ->remove.
Fixes: e4b27ebc78 ("net: dsa: Add DSA driver for Hirschmann Hellcreek switches")
Link: https://lore.kernel.org/netdev/20210909095324.12978-1-LinoSanfilippo@gmx.de/
Reported-by: Lino Sanfilippo <LinoSanfilippo@gmx.de>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Kurt Kanzenbach <kurt@linutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Lino reports that on his system with bcmgenet as DSA master and KSZ9897
as a switch, rebooting or shutting down never works properly.
What does the bcmgenet driver have special to trigger this, that other
DSA masters do not? It has an implementation of ->shutdown which simply
calls its ->remove implementation. Otherwise said, it unregisters its
network interface on shutdown.
This message can be seen in a loop, and it hangs the reboot process there:
unregister_netdevice: waiting for eth0 to become free. Usage count = 3
So why 3?
A usage count of 1 is normal for a registered network interface, and any
virtual interface which links itself as an upper of that will increment
it via dev_hold. In the case of DSA, this is the call path:
dsa_slave_create
-> netdev_upper_dev_link
-> __netdev_upper_dev_link
-> __netdev_adjacent_dev_insert
-> dev_hold
So a DSA switch with 3 interfaces will result in a usage count elevated
by two, and netdev_wait_allrefs will wait until they have gone away.
Other stacked interfaces, like VLAN, watch NETDEV_UNREGISTER events and
delete themselves, but DSA cannot just vanish and go poof, at most it
can unbind itself from the switch devices, but that must happen strictly
earlier compared to when the DSA master unregisters its net_device, so
reacting on the NETDEV_UNREGISTER event is way too late.
It seems that it is a pretty established pattern to have a driver's
->shutdown hook redirect to its ->remove hook, so the same code is
executed regardless of whether the driver is unbound from the device, or
the system is just shutting down. As Florian puts it, it is quite a big
hammer for bcmgenet to unregister its net_device during shutdown, but
having a common code path with the driver unbind helps ensure it is well
tested.
So DSA, for better or for worse, has to live with that and engage in an
arms race of implementing the ->shutdown hook too, from all individual
drivers, and do something sane when paired with masters that unregister
their net_device there. The only sane thing to do, of course, is to
unlink from the master.
However, complications arise really quickly.
The pattern of redirecting ->shutdown to ->remove is not unique to
bcmgenet or even to net_device drivers. In fact, SPI controllers do it
too (see dspi_shutdown -> dspi_remove), and presumably, I2C controllers
and MDIO controllers do it too (this is something I have not researched
too deeply, but even if this is not the case today, it is certainly
plausible to happen in the future, and must be taken into consideration).
Since DSA switches might be SPI devices, I2C devices, MDIO devices, the
insane implication is that for the exact same DSA switch device, we
might have both ->shutdown and ->remove getting called.
So we need to do something with that insane environment. The pattern
I've come up with is "if this, then not that", so if either ->shutdown
or ->remove gets called, we set the device's drvdata to NULL, and in the
other hook, we check whether the drvdata is NULL and just do nothing.
This is probably not necessary for platform devices, just for devices on
buses, but I would really insist for consistency among drivers, because
when code is copy-pasted, it is not always copy-pasted from the best
sources.
So depending on whether the DSA switch's ->remove or ->shutdown will get
called first, we cannot really guarantee even for the same driver if
rebooting will result in the same code path on all platforms. But
nonetheless, we need to do something minimally reasonable on ->shutdown
too to fix the bug. Of course, the ->remove will do more (a full
teardown of the tree, with all data structures freed, and this is why
the bug was not caught for so long). The new ->shutdown method is kept
separate from dsa_unregister_switch not because we couldn't have
unregistered the switch, but simply in the interest of doing something
quick and to the point.
The big question is: does the DSA switch's ->shutdown get called earlier
than the DSA master's ->shutdown? If not, there is still a risk that we
might still trigger the WARN_ON in unregister_netdevice that says we are
attempting to unregister a net_device which has uppers. That's no good.
Although the reference to the master net_device won't physically go away
even if DSA's ->shutdown comes afterwards, remember we have a dev_hold
on it.
The answer to that question lies in this comment above device_link_add:
* A side effect of the link creation is re-ordering of dpm_list and the
* devices_kset list by moving the consumer device and all devices depending
* on it to the ends of these lists (that does not happen to devices that have
* not been registered when this function is called).
so the fact that DSA uses device_link_add towards its master is not
exactly for nothing. device_shutdown() walks devices_kset from the back,
so this is our guarantee that DSA's shutdown happens before the master's
shutdown.
Fixes: 2f1e8ea726 ("net: dsa: link interfaces with the DSA master to get rid of lockdep warnings")
Link: https://lore.kernel.org/netdev/20210909095324.12978-1-LinoSanfilippo@gmx.de/
Reported-by: Lino Sanfilippo <LinoSanfilippo@gmx.de>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
