| Age | Commit message (Collapse) | Author |
|---|
|
As reported by smatch:
drivers/media/pci/cx25821/cx25821-alsa.c:146 cx25821_alsa_dma_init() warn: should 'nr_pages << 12' be a 64 bit type?
the number of patches should be unsigned long.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
As reported by smatch:
drivers/media/v4l2-core/videobuf-dma-sg.c:245 videobuf_dma_init_kernel() warn: should 'nr_pages << 12' be a 64 bit type?
The printk should not be using %d for the number of pages.
After looking better, the real problem here is that the
number of pages should be long int.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
As warned by smatch:
drivers/media/rc/serial_ir.c:550 serial_ir_probe() warn: should '8 << ioshift' be a 64 bit type?
the "8" constant should be unsigned long.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
The usage of castings and float point when checking for
the setup based at the symbol_rate cause those warnings
with smatch:
drivers/media/dvb-frontends/tda10021.c:153 tda10021_set_symbolrate() warn: unsigned 'symbolrate' is never less than zero.
drivers/media/dvb-frontends/tda10021.c:155 tda10021_set_symbolrate() warn: unsigned 'symbolrate' is never less than zero.
drivers/media/dvb-frontends/tda10021.c:157 tda10021_set_symbolrate() warn: unsigned 'symbolrate' is never less than zero.
drivers/media/dvb-frontends/tda10021.c:159 tda10021_set_symbolrate() warn: unsigned 'symbolrate' is never less than zero.
While the code should work with gcc, as it will evaluate the
values into a constant before compiling, other compilers
could do otherwise. So, get rid of float pointing math on it,
avoiding the need of doing typecasts.
While here, cleanup some coding style issues at the related
code.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Sparse warns about this symbol:
drivers/media/platform/qcom/venus/hfi_venus.c:133:5: warning: symbol 'venus_fw_debug' was not declared. Should it be static?
Because hfi_venus.c doesn't include a header file with the
extern. So, move it to core.h, with is included by both
hfi_venus.c and dbgfs.c.
This way, if something changes with it, warnings or errors
will be produced.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
As reported by smatch:
drivers/media/radio/wl128x/fmdrv_common.c:736 fm_irq_handle_rdsdata_getcmd_resp() warn: potential spectre issue 'rds_fmt.data.groupdatabuff.buff>
Address it by using array_index_nospec().
While here, reorder the linux/ includes.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
While smatch reports an issue there:
drivers/media/pci/ttpci/av7110_v4l.c:163 ves1820_set_tv_freq() warn: unsigned 'freq' is never less than zero.
drivers/media/pci/ttpci/av7110_v4l.c:165 ves1820_set_tv_freq() warn: unsigned 'freq' is never less than zero.
The logic is actually fine. Yet, removing the typecast
shuts up smatch and makes the code more readable.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
As pointed by smatch:
drivers/media/tuners/qt1010.c:239 qt1010_init_meas1() error: uninitialized symbol 'val2'.
drivers/media/tuners/qt1010.c:273 qt1010_init_meas2() error: uninitialized symbol 'val'.
The logic is ok, but it is hard for static analyzers
to parse it, as it depends on a value read in the middle
of a loop.
Also, it takes a while for humans to verify.
Re-write the first function to use a more direct way.
At the second one, I opted to just initialize the read var,
in order to shut up the report.
While here, address a few coding style issues at the
function code.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
The original logic had an integer to unsigned integer
conversion, plus a float-point math. While gcc should be
able to do the match at compile time, other compilers might
not do the same. Also, those produce the following warnings
with static code analyzers:
drivers/media/dvb-frontends/tda10086.c:300 tda10086_set_symbol_rate() warn: unsigned 'symbol_rate' is never less than zero.
drivers/media/dvb-frontends/tda10086.c:303 tda10086_set_symbol_rate() warn: unsigned 'symbol_rate' is never less than zero.
drivers/media/dvb-frontends/tda10086.c:306 tda10086_set_symbol_rate() warn: unsigned 'symbol_rate' is never less than zero.
drivers/media/dvb-frontends/tda10086.c:309 tda10086_set_symbol_rate() warn: unsigned 'symbol_rate' is never less than zero.
drivers/media/dvb-frontends/tda10086.c:312 tda10086_set_symbol_rate() warn: unsigned 'symbol_rate' is never less than zero.
drivers/media/dvb-frontends/tda10086.c:315 tda10086_set_symbol_rate() warn: unsigned 'symbol_rate' is never less than zero.
drivers/media/dvb-frontends/tda10086.c:318 tda10086_set_symbol_rate() warn: unsigned 'symbol_rate' is never less than zero.
drivers/media/dvb-frontends/tda10086.c:321 tda10086_set_symbol_rate() warn: unsigned 'symbol_rate' is never less than zero.
drivers/media/dvb-frontends/tda10086.c:324 tda10086_set_symbol_rate() warn: unsigned 'symbol_rate' is never less than zero.
drivers/media/dvb-frontends/tda10086.c:327 tda10086_set_symbol_rate() warn: unsigned 'symbol_rate' is never less than zero.
Change the logic to declare the reference constant as unsigned
and to not use float point math.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Upon running Sparse, "warning: dubious: !x | !y" is brought to notice
for this file. Logical and bitwise OR are basically the same in this
context, so it doesn't cause a runtime bug. But let's change it to
logical OR to make it a little bit cleaner and silence the Sparse
warning.
Signed-off-by: Anant Thazhemadam <anant.thazhemadam@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Use semicolons and braces.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
When aspect_ratio_crop_init() fails, curr_stream needs
to be freed just like what we've done in the following
error paths. However, current code is returning directly
and ends up leaking memory.
Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
There are spelling mistakes in some dev_err messages. Fix these.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
`find_gmin_subdev()` that returns a pointer to `struct
gmin_subdev` can return NULL.
In `gmin_v2p8_ctrl()` there's a call to this function but the
possibility of a NULL was not checked before its being dereferenced,
i.e.:
/* Acquired here --------v */
struct gmin_subdev *gs = find_gmin_subdev(subdev);
/* v------Dereferenced here */
if (gs->v2p8_gpio >= 0) {
...
}
With this change we're null checking `find_gmin_subdev()` result
and we return an error if that's the case. We also WARN()
for the sake of debugging.
Signed-off-by: Cengiz Can <cengiz@kernel.wtf>
Reported-by: Coverity Static Analyzer CID 1465536
Suggested-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
There are a couple of debug messages using hardcoded function names
instead of the preferred __func__ magic constant.
Replace them:
WARNING: Prefer using '"%s...", __func__' to using 'misensor_rmw_reg', this function's name, in a string
215: FILE: ./media/atomisp/i2c/atomisp-mt9m114.c:215:
+ v4l2_err(client, "misensor_rmw_reg error exit, read failed\n");
WARNING: Prefer using '"%s...", __func__' to using 'misensor_rmw_reg', this function's name, in a string
236: FILE: ./media/atomisp/i2c/atomisp-mt9m114.c:236:
+ v4l2_err(client, "misensor_rmw_reg error exit, write failed\n");
Signed-off-by: Juan Antonio Aldea-Armenteros <juant.aldea@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Fix warnings and errors generated after running checkpatch.pl on pci/sh_css_firmware.c.
Signed-off-by: Rohit K Bharadwaj <bharadwaj.rohit8@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Replace deprecated MSI IRQ enabler and disabler
with pci_alloc_irq_vectors and pci_free_irq_vectors respectively.
And as a result handle the returned error as appropriate.
Compile tested.
[mchehab: solved a merge conflict that renamed dev->pdev]
Signed-off-by: Suraj Upadhyay <usuraj35@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
When CONFIG_RETPOLINE is disabled, Clang uses a jump table for the
switch statement in cmdline_find_option (jump tables are disabled when
CONFIG_RETPOLINE is enabled). This function is called very early in boot
from sme_enable() if CONFIG_AMD_MEM_ENCRYPT is enabled. At this time,
the kernel is still executing out of the identity mapping, but the jump
table will contain virtual addresses.
Fix this by disabling jump tables for cmdline.c when AMD_MEM_ENCRYPT is
enabled.
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200903023056.3914690-1-nivedita@alum.mit.edu
|
|
There are several bad whitespacing usage there. Remove them.
While here, place all Kconfig options for sensors at the
same place.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Removed braces in 'if else' condition as it only consists of
one line each and according to coding style rules , in this case
the braces are not required.
Signed-off-by: Ankit Baluni <b18007@students.iitmandi.ac.in>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Fix braces placement errors as reported by checkpatch.pl
Signed-off-by: Aditya Jain <aditya.jainadityajain.jain@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Removed braces for a 'if' condition as it contain only single line &
there is no need for braces for such case according to coding style
rules.
Signed-off-by: Ankit Baluni <b18007@students.iitmandi.ac.in>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
This fixes some coding style issues of multiline comments to
correct a few checkpatch.pl warnings.
Signed-off-by: Rahul Gottipati <rahul.blr97@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
This removes whitespaces at the beginning of a few lines to fix
some checkpatch.pl warnings.
Signed-off-by: Rahul Gottipati <rahul.blr97@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
clang static analysis flags this problem
stream.c:844:9: warning: Use of memory after
it is freed
kfree(bus->defer_msg.msg->buf);
^~~~~~~~~~~~~~~~~~~~~~~
This happens in an error handler cleaning up memory
allocated for elements in a list.
list_for_each_entry(m_rt, &stream->master_list, stream_node) {
bus = m_rt->bus;
kfree(bus->defer_msg.msg->buf);
kfree(bus->defer_msg.msg);
}
And is triggered when the call to sdw_bank_switch() fails.
There are a two problems.
First, when sdw_bank_switch() fails, though it frees memory it
does not clear bus's reference 'defer_msg.msg' to that memory.
The second problem is the freeing msg->buf. In some cases
msg will be NULL so this will dereference a null pointer.
Need to check before freeing.
Fixes: 99b8a5d608a6 ("soundwire: Add bank switch routine")
Signed-off-by: Tom Rix <trix@redhat.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200902202650.14189-1-trix@redhat.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
For disabled paths the 'interconnect_summary' in debugfs currently shows
the orginally requested bandwidths. This is confusing, since the bandwidth
requests aren't active. Instead show the bandwidths for disabled
paths/requests as zero.
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Reviewed-by: Evan Green <evgreen@chromium.org>
Link: https://lore.kernel.org/r/20200729104933.1.If8e80e4c0c7ddf99056f6e726e59505ed4e127f3@changeid
Signed-off-by: Georgi Djakov <georgi.djakov@linaro.org>
|
|
SYSFW ABI 3.0 has changed the rchan_oes_offset value for am654 to support
SR2.
Since the kernel now needs SYSFW API 3.0 to work because the merged irqchip
update, we need to also update the am654 rchan_oes_offset.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20200831091019.25273-1-peter.ujfalusi@ti.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
This reverts commit f2af201002a8bc22500c04cc474ea480bf361351.
It added a usage of cc-ldoption, but cc-ldoption was removed in commit
055efab3120b ("kbuild: drop support for cc-ldoption").
Reported-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
Thanks to NVIDIA for confirming this workaround, and clarifying which HW
is affected.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Tested-by: Alexander Kapshuk <alexander.kapshuk@gmail.com>
|
|
This was lost during the header conversion.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
|
|
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
|
|
Looks like when we converted everything over to Nvidia's class headers,
we mistakenly included the nvif/push507b.h instead of nvif/pushc37b.h,
which resulted in breaking CRC reporting for volta+:
nouveau 0000:1f:00.0: disp: chid 0 stat 10003361 reason 3
[RESERVED_METHOD] mthd 0d84 data 00000000 code 00000000
nouveau 0000:1f:00.0: disp: chid 0 stat 10003360 reason 3
[RESERVED_METHOD] mthd 0d80 data 00000000 code 00000000
nouveau 0000:1f:00.0: DRM: CRC notifier ctx for head 3 not finished
after 50ms
So, fix that.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Fixes: c4b27bc8682c ("drm/nouveau/kms/nv50-: convert core crc_set_src() to new push macros")
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
|
|
Commit 2e26ccb119bd ("drm/radeon: prefer lower reference dividers")
fixed screen flicker for HP Compaq nx9420 but breaks other laptops like
Asus X50SL.
Turns out we also need to favor lower feedback dividers.
Users confirmed this change fixes the regression and doesn't regress the
original fix.
Fixes: 2e26ccb119bd ("drm/radeon: prefer lower reference dividers")
BugLink: https://bugs.launchpad.net/bugs/1791312
BugLink: https://bugs.launchpad.net/bugs/1861554
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
On my R9 390, the voltage was reported as a constant 1000 mV.
This was due to a bug in smu7_hwmgr.c, in the smu7_read_sensor()
function, where some magic constants were used in a condition,
to determine whether the voltage should be read from PLANE2_VID
or PLANE1_VID. The VDDC mask was incorrectly used, instead of
the VDDGFX mask.
This patch changes the code to use the correct defined constants
(and apply the correct bitshift), thus resulting in correct voltage reporting.
Signed-off-by: Sandeep Raghuraman <sandy.8925@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Starting in Linux 5.8, the graphics and memory clock frequency were not being
reported for CIK cards. This is a regression, since they were reported correctly
in Linux 5.7.
After investigation, I discovered that the smum_send_msg_to_smc() function,
attempts to call the corresponding get_argument() function of ci_smu_funcs.
However, the get_argument() function is not defined in ci_smu_funcs.
This patch fixes the bug by specifying the correct get_argument() function.
Fixes: a0ec225633d9f6 ("drm/amd/powerplay: unified interfaces for message issuing and response checking")
Signed-off-by: Sandeep Raghuraman <sandy.8925@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
|
|
Enable MP0 clock DPM for sienna_cichlid.
Signed-off-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Normally softwareshutdowntemp should be greater than Thotspotlimit.
However, on some VEGA10 ASIC, the softwareshutdowntemp is 91C while
Thotspotlimit is 105C. This seems not right and may trigger some
false alarms.
Signed-off-by: Evan Quan <evan.quan@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
|
|
v1:
the C type "unsigned long" size is 32bit on 32bit system,
it will cause code logic error, so replace it with "uint64_t".
v2:
remove duplicate cast operation.
Signed-off-by: Kevin <kevin1.wang@amd.com>
Suggest-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Jiansong Chen <Jiansong.Chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Semi-automatic removing of localization macros changed the line
from "prompt = _(prompt);" to "prompt = prompt;". Drop the
reduntand assignment.
Fixes: 694c49a7c01c ("kconfig: drop localization support")
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
|
|
This is a general cleanup of kbuild/makefiles.rst:
* Use "Chapter" for major heading references and use "section" for
the next-level heading references, for consistency.
* Section 3.8 was deleted long ago.
* Drop the ending ':' in section names in the contents list.
* Correct some section numbering references.
* Correct verb agreement typo.
* Fix run-on sentence punctuation.
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
|
|
A user reported:
'Use of uninitialized value $ENV{"LMC_KEEP"} in split at
./scripts/kconfig/streamline_config.pl line 596.'
so first check that $ENV{LMC_KEEP} is defined before trying
to use it.
Fixes: c027b02d89fd ("streamline_config.pl: add LMC_KEEP to preserve some kconfigs")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
|
|
Block layer usually doesn't support or allow zero-length bvec. Since
commit 1bdc76aea115 ("iov_iter: use bvec iterator to implement
iterate_bvec()"), iterate_bvec() switches to bvec iterator. However,
Al mentioned that 'Zero-length segments are not disallowed' in iov_iter.
Fixes for_each_bvec() so that it can move on after seeing one zero
length bvec.
Fixes: 1bdc76aea115 ("iov_iter: use bvec iterator to implement iterate_bvec()")
Reported-by: syzbot <syzbot+61acc40a49a3e46e25ea@syzkaller.appspotmail.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Tested-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: <stable@vger.kernel.org>
Link: https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg2262077.html
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
disable_irq() might sleep, replace it with disable_irq_nosync(). For
synchronisation 'irq_poll_scheduled' is sufficient
Fixes: 320e77acb3 scsi: mpt3sas: Irq poll to avoid CPU hard lockups
Link: https://lore.kernel.org/r/20200901145026.12174-1-thenzl@redhat.com
Signed-off-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
disable_irq() might sleep. Replace it with disable_irq_nosync() which is
sufficient as irq_poll_scheduled protects against concurrently running
complete_cmd_fusion() from megasas_irqpoll() and megasas_isr_fusion().
Link: https://lore.kernel.org/r/20200827165332.8432-1-thenzl@redhat.com
Fixes: a6ffd5bf681 scsi: megaraid_sas: Call disable_irq from process IRQ poll
Signed-off-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
tpg->np_login_sem
The iSCSI target login thread might get stuck with the following stack:
cat /proc/`pidof iscsi_np`/stack
[<0>] down_interruptible+0x42/0x50
[<0>] iscsit_access_np+0xe3/0x167
[<0>] iscsi_target_locate_portal+0x695/0x8ac
[<0>] __iscsi_target_login_thread+0x855/0xb82
[<0>] iscsi_target_login_thread+0x2f/0x5a
[<0>] kthread+0xfa/0x130
[<0>] ret_from_fork+0x1f/0x30
This can be reproduced via the following steps:
1. Initiator A tries to log in to iqn1-tpg1 on port 3260. After finishing
PDU exchange in the login thread and before the negotiation is finished
the the network link goes down. At this point A has not finished login
and tpg->np_login_sem is held.
2. Initiator B tries to log in to iqn2-tpg1 on port 3260. After finishing
PDU exchange in the login thread the target expects to process remaining
login PDUs in workqueue context.
3. Initiator A' tries to log in to iqn1-tpg1 on port 3260 from a new
socket. A' will wait for tpg->np_login_sem with np->np_login_timer
loaded to wait for at most 15 seconds. The lock is held by A so A'
eventually times out.
4. Before A' got timeout initiator B gets negotiation failed and calls
iscsi_target_login_drop()->iscsi_target_login_sess_out(). The
np->np_login_timer is canceled and initiator A' will hang forever.
Because A' is now in the login thread, no new login requests can be
serviced.
Fix this by moving iscsi_stop_login_thread_timer() out of
iscsi_target_login_sess_out(). Also remove iscsi_np parameter from
iscsi_target_login_sess_out().
Link: https://lore.kernel.org/r/20200729130343.24976-1-houpu@bytedance.com
Cc: stable@vger.kernel.org
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Hou Pu <houpu@bytedance.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Fix the registers being written to as the values were being over written
when writing the same registers.
Fixes: caabee5b53f5 ("net: phy: dp83867: support Wake on LAN")
Signed-off-by: Dan Murphy <dmurphy@ti.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Fix an issue where the driver wrongly detected ipv6 neighbour updates
from the NFP as corrupt. Add a reserved field on the kernel side so
it is similar to the ipv4 version of the struct and has space for the
extra bytes from the card.
Fixes: 9ea9bfa12240 ("nfp: flower: support ipv6 tunnel keep-alive messages from fw")
Signed-off-by: Louis Peens <louis.peens@netronome.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
syzbot is reporting hung task at nbd_ioctl() [1], for there are two
problems regarding TIPC's connectionless socket's shutdown() operation.
----------
#include <fcntl.h>
#include <sys/socket.h>
#include <sys/ioctl.h>
#include <linux/nbd.h>
#include <unistd.h>
int main(int argc, char *argv[])
{
const int fd = open("/dev/nbd0", 3);
alarm(5);
ioctl(fd, NBD_SET_SOCK, socket(PF_TIPC, SOCK_DGRAM, 0));
ioctl(fd, NBD_DO_IT, 0); /* To be interrupted by SIGALRM. */
return 0;
}
----------
One problem is that wait_for_completion() from flush_workqueue() from
nbd_start_device_ioctl() from nbd_ioctl() cannot be completed when
nbd_start_device_ioctl() received a signal at wait_event_interruptible(),
for tipc_shutdown() from kernel_sock_shutdown(SHUT_RDWR) from
nbd_mark_nsock_dead() from sock_shutdown() from nbd_start_device_ioctl()
is failing to wake up a WQ thread sleeping at wait_woken() from
tipc_wait_for_rcvmsg() from sock_recvmsg() from sock_xmit() from
nbd_read_stat() from recv_work() scheduled by nbd_start_device() from
nbd_start_device_ioctl(). Fix this problem by always invoking
sk->sk_state_change() (like inet_shutdown() does) when tipc_shutdown() is
called.
The other problem is that tipc_wait_for_rcvmsg() cannot return when
tipc_shutdown() is called, for tipc_shutdown() sets sk->sk_shutdown to
SEND_SHUTDOWN (despite "how" is SHUT_RDWR) while tipc_wait_for_rcvmsg()
needs sk->sk_shutdown set to RCV_SHUTDOWN or SHUTDOWN_MASK. Fix this
problem by setting sk->sk_shutdown to SHUTDOWN_MASK (like inet_shutdown()
does) when the socket is connectionless.
[1] https://syzkaller.appspot.com/bug?id=3fe51d307c1f0a845485cf1798aa059d12bf18b2
Reported-by: syzbot <syzbot+e36f41d207137b5d12f7@syzkaller.appspotmail.com>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Cited commit added the possible value of '2', but it cannot be set. Fix
it by adjusting the maximum value to '2'. This is consistent with the
corresponding IPv4 sysctl.
Before:
# sysctl -w net.ipv6.fib_multipath_hash_policy=2
sysctl: setting key "net.ipv6.fib_multipath_hash_policy": Invalid argument
net.ipv6.fib_multipath_hash_policy = 2
# sysctl net.ipv6.fib_multipath_hash_policy
net.ipv6.fib_multipath_hash_policy = 0
After:
# sysctl -w net.ipv6.fib_multipath_hash_policy=2
net.ipv6.fib_multipath_hash_policy = 2
# sysctl net.ipv6.fib_multipath_hash_policy
net.ipv6.fib_multipath_hash_policy = 2
Fixes: d8f74f0975d8 ("ipv6: Support multipath hashing on inner IP pkts")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Stephen Suryaputra <ssuryaextr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Change the default value of hard_header_len in hdlc.c from 16 to 0.
Currently there are 6 HDLC protocol drivers, among them:
hdlc_raw_eth, hdlc_cisco, hdlc_ppp, hdlc_x25 set hard_header_len when
attaching the protocol, overriding the default. So this patch does not
affect them.
hdlc_raw and hdlc_fr don't set hard_header_len when attaching the
protocol. So this patch will change the hard_header_len of the HDLC
device for them from 16 to 0.
This is the correct change because both hdlc_raw and hdlc_fr don't have
header_ops, and the code in net/packet/af_packet.c expects the value of
hard_header_len to be consistent with header_ops.
In net/packet/af_packet.c, in the packet_snd function,
for AF_PACKET/DGRAM sockets it would reserve a headroom of
hard_header_len and call dev_hard_header to fill in that headroom,
and for AF_PACKET/RAW sockets, it does not reserve the headroom and
does not call dev_hard_header, but checks if the user has provided a
header of length hard_header_len (in function dev_validate_header).
Cc: Krzysztof Halasa <khc@pm.waw.pl>
Cc: Martin Schiller <ms@dev.tdt.de>
Signed-off-by: Xie He <xie.he.0141@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
