- 21 11月, 2019 3 次提交
-
-
由 Tomer Tayar 提交于
s/paerser/parser/ s/requeusted/requested/ s/an JOB/a JOB/ Signed-off-by: NTomer Tayar <ttayar@habana.ai> Reviewed-by: NOded Gabbay <oded.gabbay@gmail.com> Signed-off-by: NOded Gabbay <oded.gabbay@gmail.com>
-
由 YueHaibing 提交于
Fixes gcc '-Wunused-but-set-variable' warning: drivers/misc/habanalabs/device.c: In function hpriv_release: drivers/misc/habanalabs/device.c:45:17: warning: variable ctx set but not used [-Wunused-but-set-variable] It is never used since commit eb7caf84 ("habanalabs: maintain a list of file private data objects") Reported-by: NHulk Robot <hulkci@huawei.com> Signed-off-by: NYueHaibing <yuehaibing@huawei.com> Reviewed-by: NOded Gabbay <oded.gabbay@gmail.com> Signed-off-by: NOded Gabbay <oded.gabbay@gmail.com>
-
由 Oded Gabbay 提交于
In case the F/W fails to initialize the thermal sensors, print an appropriate error message to kernel log and fail the device initialization. Reviewed-by: NTomer Tayar <ttayar@habana.ai> Signed-off-by: NOded Gabbay <oded.gabbay@gmail.com>
-
- 20 11月, 2019 8 次提交
-
-
由 Krzysztof Kozlowski 提交于
Adjust indentation from spaces to tab (+optional two spaces) as in coding style with command like: $ sed -e 's/^ /\t/' -i */Kconfig Signed-off-by: NKrzysztof Kozlowski <krzk@kernel.org> Link: https://lore.kernel.org/r/20191120134056.14677-1-krzk@kernel.orgSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Krzysztof Kozlowski 提交于
Adjust indentation from spaces to tab (+optional two spaces) as in coding style with command like: $ sed -e 's/^ /\t/' -i */Kconfig Signed-off-by: NKrzysztof Kozlowski <krzk@kernel.org> Link: https://lore.kernel.org/r/20191120134247.16073-1-krzk@kernel.orgSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Alexander Shishkin 提交于
This adds support for the Trace Hub in Tiger Lake CPU. Signed-off-by: NAlexander Shishkin <alexander.shishkin@linux.intel.com> Reviewed-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20191120130806.44028-4-alexander.shishkin@linux.intel.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Alexander Shishkin 提交于
This adds support for the Trace Hub in Ice Lake CPU. Signed-off-by: NAlexander Shishkin <alexander.shishkin@linux.intel.com> Reviewed-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20191120130806.44028-3-alexander.shishkin@linux.intel.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Alexander Shishkin 提交于
Commit a753bfcf ("intel_th: Make the switch allocate its subdevices") factored out intel_th_subdevice_alloc() from intel_th_populate(), but got the error path wrong, resulting in two instances of a double put_device() on a freshly initialized, but not 'added' device. Fix this by only doing one put_device() in the error path. Signed-off-by: NAlexander Shishkin <alexander.shishkin@linux.intel.com> Fixes: a753bfcf ("intel_th: Make the switch allocate its subdevices") Reported-by: NWen Yang <wenyang@linux.alibaba.com> Reviewed-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: stable@vger.kernel.org # v4.14+ Link: https://lore.kernel.org/r/20191120130806.44028-2-alexander.shishkin@linux.intel.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Wei Yongjun 提交于
The driver allocates the spinlock but not initialize it. Use spin_lock_init() on it to initialize it correctly. This is detected by Coccinelle semantic patch. Signed-off-by: NWei Yongjun <weiyongjun1@huawei.com> Tested-by: NYabin Cui <yabinc@google.com> Signed-off-by: NMathieu Poirier <mathieu.poirier@linaro.org> Link: https://lore.kernel.org/r/20191118185207.30441-3-mathieu.poirier@linaro.orgSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Wei Yongjun 提交于
The driver allocates the spinlock but not initialize it. Use spin_lock_init() on it to initialize it correctly. This is detected by Coccinelle semantic patch. Signed-off-by: NWei Yongjun <weiyongjun1@huawei.com> Tested-by: NYabin Cui <yabinc@google.com> Signed-off-by: NMathieu Poirier <mathieu.poirier@linaro.org> Link: https://lore.kernel.org/r/20191118185207.30441-2-mathieu.poirier@linaro.orgSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Mika Westerberg 提交于
On zang's Dell XPS 13 9370 after Thunderbolt NVM firmware upgrade the Thunderbolt controller did not come back as expected. Only after the system was rebooted it became available again. It is not entirely clear what happened but I suspect the new NVM firmware image authentication failed for some reason. Regardless of this the router needs to be power cycled if NVM authentication fails in order to get it fully functional again. This modifies the driver to issue a power cycle in case the NVM authentication fails immediately when dma_port_flash_update_auth() returns. We also need to call tb_switch_set_uuid() earlier to be able to fetch possible NVM authentication failure when DMA port is added. Link: https://bugzilla.kernel.org/show_bug.cgi?id=205457Reported-by: Nzang <dump@tzib.net> Cc: stable <stable@vger.kernel.org> Signed-off-by: NMika Westerberg <mika.westerberg@linux.intel.com> Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
- 16 11月, 2019 1 次提交
-
-
由 Alexander Usyskin 提交于
Export more client attributes via sysfs that are usually obtained upon connection. In some cases, for example a monitoring application may wish to know the attributes without actually performing the connection. Added attributes: max number of connections, fixed address, max message length. Signed-off-by: NAlexander Usyskin <alexander.usyskin@intel.com> Signed-off-by: NTomas Winkler <tomas.winkler@intel.com> Link: https://lore.kernel.org/r/20191116142136.17535-1-tomas.winkler@intel.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
- 14 11月, 2019 10 次提交
-
-
由 Alexander Shishkin 提交于
Commit c7fd62bc ("stm class: Introduce framing protocol drivers") forgot to tear down the link between an stm device and its protocol driver when policy is removed. This leads to an invalid pointer reference if one tries to write to an stm device after the policy has been removed and the protocol driver module unloaded, leading to the below splat: > BUG: unable to handle page fault for address: ffffffffc0737068 > #PF: supervisor read access in kernel mode > #PF: error_code(0x0000) - not-present page > PGD 3d780f067 P4D 3d780f067 PUD 3d7811067 PMD 492781067 PTE 0 > Oops: 0000 [#1] SMP NOPTI > CPU: 1 PID: 26122 Comm: cat Not tainted 5.4.0-rc5+ #1 > RIP: 0010:stm_output_free+0x40/0xc0 [stm_core] > Call Trace: > stm_char_release+0x3e/0x70 [stm_core] > __fput+0xc6/0x260 > ____fput+0xe/0x10 > task_work_run+0x9d/0xc0 > exit_to_usermode_loop+0x103/0x110 > do_syscall_64+0x19d/0x1e0 > entry_SYSCALL_64_after_hwframe+0x44/0xa9 Fix this by tearing down the link from an stm device to its protocol driver when the policy involving that driver is removed. Signed-off-by: NAlexander Shishkin <alexander.shishkin@linux.intel.com> Fixes: c7fd62bc ("stm class: Introduce framing protocol drivers") Reported-by: NAmmy Yi <ammy.yi@intel.com> Tested-by: NAmmy Yi <ammy.yi@intel.com> CC: stable@vger.kernel.org # v4.20+ Link: https://lore.kernel.org/r/20191114064201.43089-2-alexander.shishkin@linux.intel.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Richard Gong 提交于
The older versions of remote system update (RSU) firmware don't support retry and notify features then the kernel module dies when it queries the RSU retry counter or performs notify operation. Update the Intel service layer and RSU drivers to be compatible with all versions of RSU firmware. Reported-by: NRadu Barcau <radu.bacrau@intel.com> Reported-by: Nkbuild test robot <lkp@intel.com> Signed-off-by: NRichard Gong <richard.gong@intel.com> Link: https://lore.kernel.org/r/1572884676-1385-1-git-send-email-richard.gong@linux.intel.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Rui Feng 提交于
Add support for new chip rts5261. In order to support rts5261, the definitions of some internal registers and workflow have to be modified and are different from its predecessors. So we need this patch to ensure RTS5261 can work. Signed-off-by: NRui Feng <rui_feng@realsil.com.cn> Link: https://lore.kernel.org/r/1571645105-5028-1-git-send-email-rui_feng@realsil.com.cnSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Angelo Dureghello 提交于
add support for ds2430, 1 page, 256bit (32bytes) eeprom (family 0x14). Tests done: 32 bytes dump: x@y:~# hexdump -C -n 32 /sys/bus/w1/devices/14-00000158556e/eeprom 00000000 39 39 0a 00 00 36 0a ff ff ff ff ff ff ff ff ff 00000010 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff 00000020 34 bytes dump: 32 only displayed x@y:~# hexdump -C -n 34 /sys/bus/w1/devices/14-00000158556e/eeprom 00000000 39 39 0a 00 00 36 0a ff ff ff ff ff ff ff ff ff 00000010 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff 00000020 pattern write: x@y:~# echo 123456789 > /sys/bus/w1/devices/14-00000158556e/eeprom x@y:~# hexdump -C -n 54 /sys/bus/w1/devices/14-00000158556e/eeprom 00000000 31 32 33 34 35 36 37 38 39 0a ff ff ff ff ff ff 00000010 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff 00000020 specific address 1-byte write x@y:~# dd if=/dev/zero of=/sys/bus/w1/devices/14-00000158556e/eeprom \ count=1 bs=1 seek=4 1+0 records in 1+0 records out x@y:~# hexdump -C -n 54 /sys/bus/w1/devices/14-00000158556e/eeprom 00000000 31 32 33 34 00 36 37 38 39 0a ff ff ff ff ff ff 00000010 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff 00000020 writing binary block x@y:~# cat dump-128bytes.bin > /sys/bus/w1/devices/14-00000158556e/eeprom cat: write error: File too large x@y:~# cat dump-32bytes.bin > /sys/bus/w1/devices/14-00000158556e/eeprom x@y:~# hexdump -C -n 54 /sys/bus/w1/devices/14-00000158556e/eeprom 00000000 10 0b 5b ff ff ff ff ff ff ff ff ff ff ff ff ff 00000010 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff 40 00000020 Signed-off-by: NAngelo Dureghello <angelo.dureghello@timesys.com> Link: https://lore.kernel.org/r/20191019204015.61474-1-angelo.dureghello@timesys.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Ben Dooks (Codethink) 提交于
Include cs_internal.h (and pcmcia/cistpl.h as required by cs_internal.h) for the declearions of cb_alloc and cb_free to silence the following sparse warnings; drivers/pcmcia/cardbus.c:64:11: warning: symbol 'cb_alloc' was not declared. Should it be static? drivers/pcmcia/cardbus.c:103:6: warning: symbol 'cb_free' was not declared. Should it be static? Signed-off-by: NBen Dooks <ben.dooks@codethink.co.uk> Link: https://lore.kernel.org/r/20191017114059.10989-1-ben.dooks@codethink.co.ukSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Ben Dooks (Codethink) 提交于
Include <pcmcia/ds.h> for pcmcia_parse_tuple declaration to fix the following sparse warning: drivers/pcmcia/cistpl.c:1287:5: warning: symbol 'pcmcia_parse_tuple' was not declared. Should it be static? Signed-off-by: NBen Dooks <ben.dooks@codethink.co.uk> Link: https://lore.kernel.org/r/20191017114447.20455-1-ben.dooks@codethink.co.ukSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Alexandru Ardelean 提交于
This change also does a bit of a unification for the IRQ init code. But the actual problem is that UIO_IRQ_NONE == 0, so for the DT case where UIO_IRQ_NONE gets assigned to `uioinfo->irq`, a 2nd initialization will get triggered (for the IRQ) and this one will exit via `goto bad1`. As far as things seem to go, the only case where UIO_IRQ_NONE seems valid, is when using a device-tree. The driver has some legacy support for old platform_data structures. It looks like, for platform_data a non-existent IRQ is an invalid case (or was considered an invalid case). Which is why -ENXIO is treated only when a DT is used. Signed-off-by: NDragos Bogdan <dragos.bogdan@analog.com> Signed-off-by: NAlexandru Ardelean <alexandru.ardelean@analog.com> Acked-by: NDamian Hobson-Garcia <dhobsong@igel.co.jp> Link: https://lore.kernel.org/r/20191105073212.16719-1-alexandru.ardelean@analog.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Jann Horn 提交于
The old loop wouldn't stop when reaching `start` if `start==NULL`, instead continuing backwards to index -1 and crashing. Luckily you need to be highly privileged to map things at NULL, so it's not a big problem. Fix it by adjusting the loop so that the loop variable is always in bounds. This patch is deliberately minimal to simplify backporting, but IMO this function could use a refactor. The jump labels in the second loop body are horrible (the error gotos should be jumping to free_range instead), and both loops would look nicer if they just iterated upwards through indices. And the up_read()+mmput() shouldn't be duplicated like that. Cc: stable@vger.kernel.org Fixes: 457b9a6f ("Staging: android: add binder driver") Signed-off-by: NJann Horn <jannh@google.com> Acked-by: NChristian Brauner <christian.brauner@ubuntu.com> Link: https://lore.kernel.org/r/20191018205631.248274-3-jannh@google.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Jann Horn 提交于
binder_alloc_mmap_handler() attempts to detect the use of ->mmap() on a binder_proc whose binder_alloc has already been initialized by checking whether alloc->buffer is non-zero. Before commit 88021166 ("binder: remove kernel vm_area for buffer space"), alloc->buffer was a kernel mapping address, which is always non-zero, but since that commit, it is a userspace mapping address. A sufficiently privileged user can map /dev/binder at NULL, tricking binder_alloc_mmap_handler() into assuming that the binder_proc has not been mapped yet. This leads to memory unsafety. Luckily, no context on Android has such privileges, and on a typical Linux desktop system, you need to be root to do that. Fix it by using the mapping size instead of the mapping address to distinguish the mapped case. A valid VMA can't have size zero. Fixes: 88021166 ("binder: remove kernel vm_area for buffer space") Cc: stable@vger.kernel.org Signed-off-by: NJann Horn <jannh@google.com> Acked-by: NChristian Brauner <christian.brauner@ubuntu.com> Link: https://lore.kernel.org/r/20191018205631.248274-2-jannh@google.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Jann Horn 提交于
binder_alloc_print_pages() iterates over alloc->pages[0..alloc->buffer_size-1] under alloc->mutex. binder_alloc_mmap_handler() writes alloc->pages and alloc->buffer_size without holding that lock, and even writes them before the last bailout point. Unfortunately we can't take the alloc->mutex in the ->mmap() handler because mmap_sem can be taken while alloc->mutex is held. So instead, we have to locklessly check whether the binder_alloc has been fully initialized with binder_alloc_get_vma(), like in binder_alloc_new_buf_locked(). Fixes: 8ef4665a ("android: binder: Add page usage in binder stats") Cc: stable@vger.kernel.org Signed-off-by: NJann Horn <jannh@google.com> Acked-by: NChristian Brauner <christian.brauner@ubuntu.com> Link: https://lore.kernel.org/r/20191018205631.248274-1-jannh@google.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
- 13 11月, 2019 6 次提交
-
-
由 Sudip Mukherjee 提交于
Modify parport daisy driver to use the new parallel port device model. Last attempt was '1aec4211 ("parport: daisy: use new parport device model")' which failed as daisy was also trying to load the low level driver and that resulted in a deadlock. Cc: Michal Kubecek <mkubecek@suse.cz> Signed-off-by: NSudip Mukherjee <sudipm.mukherjee@gmail.com> Tested-by: NSteven Rostedt (VMware) <rostedt@goodmis.org> Link: https://lore.kernel.org/r/20191016144540.18810-4-sudipm.mukherjee@gmail.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Sudip Mukherjee 提交于
Usually all the distro will load the parport low level driver as part of their initialization. But we can get into a situation where all the parallel port drivers are built as module and we unload all the modules at a later time. Then if we just do "modprobe parport" it will only load the parport module and will not load the low level driver which will actually register the ports. So, check the bus if there is any parport registered, if not, load the low level driver. We can get into the above situation with all distro but only Suse has setup the alias for "parport_lowlevel" and so it only works in Suse. Users of Debian based distro will need to load the lowlevel module manually. Signed-off-by: NSudip Mukherjee <sudipm.mukherjee@gmail.com> Link: https://lore.kernel.org/r/20191016144540.18810-3-sudipm.mukherjee@gmail.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Sudip Mukherjee 提交于
We do not need to maintain a list of ports when we are using the device-model. The base layer is going to maintain the list for us and we can get the list of ports just using bus_for_each_dev(). Signed-off-by: NSudip Mukherjee <sudipm.mukherjee@gmail.com> Link: https://lore.kernel.org/r/20191016144540.18810-2-sudipm.mukherjee@gmail.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Sudip Mukherjee 提交于
The daisy device name is hardcoded, define it in the header file and use it in the code. Signed-off-by: NSudip Mukherjee <sudipm.mukherjee@gmail.com> Link: https://lore.kernel.org/r/20191016144540.18810-1-sudipm.mukherjee@gmail.comSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Arnd Bergmann 提交于
The layout of struct timeval is different on sparc64 from anything else, and the patch I did long ago failed to take this into account. Change it now to handle sparc64 user space correctly again. Quite likely nobody cares about parallel ports on sparc64, but there is no reason not to fix it. Cc: stable@vger.kernel.org Fixes: 9a450484 ("lp: support 64-bit time_t user space") Signed-off-by: NArnd Bergmann <arnd@arndb.de> Link: https://lore.kernel.org/r/20191108203435.112759-7-arnd@arndb.deSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
由 Arnd Bergmann 提交于
Going through the uses of timeval in the user space API, I noticed two bugs in ppdev that were introduced in the y2038 conversion: * The range check was accidentally moved from ppsettime to ppgettime * On sparc64, the microseconds are in the other half of the 64-bit word. Fix both, and mark the fix for stable backports. Cc: stable@vger.kernel.org Fixes: 3b9ab374 ("ppdev: convert to y2038 safe") Signed-off-by: NArnd Bergmann <arnd@arndb.de> Link: https://lore.kernel.org/r/20191108203435.112759-8-arnd@arndb.deSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
-
- 09 11月, 2019 12 次提交
-
-
由 Pierre-Louis Bossart 提交于
The uniqueID is useful when there are two or more devices of the same type (identical manufacturer ID, part ID) on the same link. When there is a single device of a given type on a link, its uniqueID is irrelevant. It's not uncommon on actual platforms to see variations of the uniqueID, or differences between devID registers and ACPI _ADR fields. This patch suggests a filter on startup to identify 'single' devices and tag them accordingly. The uniqueID is then not used for the probe, and the device name omits the uniqueID as well. Signed-off-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20191022234808.17432-4-pierre-louis.bossart@linux.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>
-
由 Pierre-Louis Bossart 提交于
Simplify the loop with a helper. The only functionality change is that we continue the loop even with an ACPI error. Follow-up patches will build on this change. Signed-off-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20191022234808.17432-3-pierre-louis.bossart@linux.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>
-
由 Pierre-Louis Bossart 提交于
The previous formula is incorrect for PDI0/1, the mapping is not linear but has a discontinuity between PDI1 and PDI2. This change has no effect on PCM PDIs (same mapping). Signed-off-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20191022232948.17156-1-pierre-louis.bossart@linux.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>
-
由 Magnus Karlsson 提交于
The need_wakeup flag for Tx might not be set for AF_XDP sockets that are only used to send packets. This happens if there is at least one outstanding packet that has not been completed by the hardware and we get that corresponding completion (which will not generate an interrupt since interrupts are disabled in the napi poll loop) between the time we stopped processing the Tx completions and interrupts are enabled again. In this case, the need_wakeup flag will have been cleared at the end of the Tx completion processing as we believe we will get an interrupt from the outstanding completion at a later point in time. But if this completion interrupt occurs before interrupts are enable, we lose it and should at that point really have set the need_wakeup flag since there are no more outstanding completions that can generate an interrupt to continue the processing. When this happens, user space will see a Tx queue need_wakeup of 0 and skip issuing a syscall, which means will never get into the Tx processing again and we have a deadlock. This patch introduces a quick fix for this issue by just setting the need_wakeup flag for Tx to 1 all the time. I am working on a proper fix for this that will toggle the flag appropriately, but it is more challenging than I anticipated and I am afraid that this patch will not be completed before the merge window closes, therefore this easier fix for now. This fix has a negative performance impact in the range of 0% to 4%. Towards the higher end of the scale if you have driver and application on the same core and issue a lot of packets, and towards no negative impact if you use two cores, lower transmission speeds and/or a workload that also receives packets. Signed-off-by: NMagnus Karlsson <magnus.karlsson@intel.com> Tested-by: NAndrew Bowers <andrewx.bowers@intel.com> Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
-
由 Magnus Karlsson 提交于
The need_wakeup flag for Tx might not be set for AF_XDP sockets that are only used to send packets. This happens if there is at least one outstanding packet that has not been completed by the hardware and we get that corresponding completion (which will not generate an interrupt since interrupts are disabled in the napi poll loop) between the time we stopped processing the Tx completions and interrupts are enabled again. In this case, the need_wakeup flag will have been cleared at the end of the Tx completion processing as we believe we will get an interrupt from the outstanding completion at a later point in time. But if this completion interrupt occurs before interrupts are enable, we lose it and should at that point really have set the need_wakeup flag since there are no more outstanding completions that can generate an interrupt to continue the processing. When this happens, user space will see a Tx queue need_wakeup of 0 and skip issuing a syscall, which means will never get into the Tx processing again and we have a deadlock. This patch introduces a quick fix for this issue by just setting the need_wakeup flag for Tx to 1 all the time. I am working on a proper fix for this that will toggle the flag appropriately, but it is more challenging than I anticipated and I am afraid that this patch will not be completed before the merge window closes, therefore this easier fix for now. This fix has a negative performance impact in the range of 0% to 4%. Towards the higher end of the scale if you have driver and application on the same core and issue a lot of packets, and towards no negative impact if you use two cores, lower transmission speeds and/or a workload that also receives packets. Signed-off-by: NMagnus Karlsson <magnus.karlsson@intel.com> Tested-by: NAndrew Bowers <andrewx.bowers@intel.com> Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
-
由 Jacob Keller 提交于
When implementing launch time support in the igb and igc drivers, the skb->tstamp value is assumed to be a s64, but it's declared as a ktime_t value. Although ktime_t is typedef'd to s64 it wasn't always, and the kernel provides accessors for ktime_t values. Use the ktime_to_timespec64 and ktime_set accessors instead of directly assuming that the variable is always an s64. This improves portability if the code is ever moved to another kernel version, or if the definition of ktime_t ever changes again in the future. Signed-off-by: NJacob Keller <jacob.e.keller@intel.com> Acked-by: NVinicius Costa Gomes <vinicius.gomes@intel.com> Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
-
由 Arkadiusz Kubalewski 提交于
This patch contains fix for a problem with command: 'ethtool -m <dev>' which breaks functionality of: 'ethtool <dev>' when called on X722 NIC Disallowed update of link phy_types on X722 NIC Currently correct value cannot be obtained from FW Previously wrong value returned by FW was used and was a root cause for incorrect output of 'ethtool <dev>' command Signed-off-by: NArkadiusz Kubalewski <arkadiusz.kubalewski@intel.com> Tested-by: NAndrew Bowers <andrewx.bowers@intel.com> Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
-
由 Nicholas Nunley 提交于
Since commit 92418fb1 ("i40e/i40evf: Use usec value instead of reg value for ITR defines") the driver tracks the interrupt throttling intervals in single usec units, although the actual ITRN registers are programmed in 2 usec units. Most register programming flows in the driver correctly handle the conversion, although it is currently not applied when the registers are initialized to their default values. Most of the time this doesn't present a problem since the default values are usually immediately overwritten through the standard adaptive throttling mechanism, or updated manually by the user, but if adaptive throttling is disabled and the interval values are left alone then the incorrect value will persist. Since the intended default interval of 50 usecs (vs. 100 usecs as programmed) performs better for most traffic workloads, this can lead to performance regressions. This patch adds the correct conversion when writing the initial values to the ITRN registers. Signed-off-by: NNicholas Nunley <nicholas.d.nunley@intel.com> Tested-by: NAndrew Bowers <andrewx.bowers@intel.com> Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
-
由 Colin Ian King 提交于
Currently the for-loop counter i is a u8 however it is being checked against a maximum value hw->num_tx_sched_layers which is a u16. Hence there is a potential wrap-around of counter i back to zero if hw->num_tx_sched_layers is greater than 255. Fix this by making i a u16. Addresses-Coverity: ("Infinite loop") Fixes: b36c598c ("ice: Updates to Tx scheduler code") Signed-off-by: NColin Ian King <colin.king@canonical.com> Tested-by: NAndrew Bowers <andrewx.bowers@intel.com> Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
-
由 Manish Chopra 提交于
While rebooting the system with SR-IOV vfs enabled leads to below crash due to recurrence of __qede_remove() on the VF devices (first from .shutdown() flow of the VF itself and another from PF's .shutdown() flow executing pci_disable_sriov()) This patch adds a safeguard in __qede_remove() flow to fix this, so that driver doesn't attempt to remove "already removed" devices. [ 194.360134] BUG: unable to handle kernel NULL pointer dereference at 00000000000008dc [ 194.360227] IP: [<ffffffffc03553c4>] __qede_remove+0x24/0x130 [qede] [ 194.360304] PGD 0 [ 194.360325] Oops: 0000 [#1] SMP [ 194.360360] Modules linked in: tcp_lp fuse tun bridge stp llc devlink bonding ip_set nfnetlink ib_isert iscsi_target_mod ib_srpt target_core_mod ib_srp scsi_transport_srp scsi_tgt ib_ipoib ib_umad rpcrdma sunrpc rdma_ucm ib_uverbs ib_iser rdma_cm iw_cm ib_cm libiscsi scsi_transport_iscsi dell_smbios iTCO_wdt iTCO_vendor_support dell_wmi_descriptor dcdbas vfat fat pcc_cpufreq skx_edac intel_powerclamp coretemp intel_rapl iosf_mbi kvm_intel kvm irqbypass crc32_pclmul ghash_clmulni_intel aesni_intel lrw gf128mul glue_helper ablk_helper cryptd qedr ib_core pcspkr ses enclosure joydev ipmi_ssif sg i2c_i801 lpc_ich mei_me mei wmi ipmi_si ipmi_devintf ipmi_msghandler tpm_crb acpi_pad acpi_power_meter xfs libcrc32c sd_mod crc_t10dif crct10dif_generic crct10dif_pclmul crct10dif_common crc32c_intel mgag200 [ 194.361044] qede i2c_algo_bit drm_kms_helper qed syscopyarea sysfillrect nvme sysimgblt fb_sys_fops ttm nvme_core mpt3sas crc8 ptp drm pps_core ahci raid_class scsi_transport_sas libahci libata drm_panel_orientation_quirks nfit libnvdimm dm_mirror dm_region_hash dm_log dm_mod [last unloaded: ip_tables] [ 194.361297] CPU: 51 PID: 7996 Comm: reboot Kdump: loaded Not tainted 3.10.0-1062.el7.x86_64 #1 [ 194.361359] Hardware name: Dell Inc. PowerEdge MX840c/0740HW, BIOS 2.4.6 10/15/2019 [ 194.361412] task: ffff9cea9b360000 ti: ffff9ceabebdc000 task.ti: ffff9ceabebdc000 [ 194.361463] RIP: 0010:[<ffffffffc03553c4>] [<ffffffffc03553c4>] __qede_remove+0x24/0x130 [qede] [ 194.361534] RSP: 0018:ffff9ceabebdfac0 EFLAGS: 00010282 [ 194.361570] RAX: 0000000000000000 RBX: ffff9cd013846098 RCX: 0000000000000000 [ 194.361621] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffff9cd013846098 [ 194.361668] RBP: ffff9ceabebdfae8 R08: 0000000000000000 R09: 0000000000000000 [ 194.361715] R10: 00000000bfe14201 R11: ffff9ceabfe141e0 R12: 0000000000000000 [ 194.361762] R13: ffff9cd013846098 R14: 0000000000000000 R15: ffff9ceab5e48000 [ 194.361810] FS: 00007f799c02d880(0000) GS:ffff9ceacb0c0000(0000) knlGS:0000000000000000 [ 194.361865] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 194.361903] CR2: 00000000000008dc CR3: 0000001bdac76000 CR4: 00000000007607e0 [ 194.361953] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 194.362002] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 [ 194.362051] PKRU: 55555554 [ 194.362073] Call Trace: [ 194.362109] [<ffffffffc0355500>] qede_remove+0x10/0x20 [qede] [ 194.362180] [<ffffffffb97d0f3e>] pci_device_remove+0x3e/0xc0 [ 194.362240] [<ffffffffb98b3c52>] __device_release_driver+0x82/0xf0 [ 194.362285] [<ffffffffb98b3ce3>] device_release_driver+0x23/0x30 [ 194.362343] [<ffffffffb97c86d4>] pci_stop_bus_device+0x84/0xa0 [ 194.362388] [<ffffffffb97c87e2>] pci_stop_and_remove_bus_device+0x12/0x20 [ 194.362450] [<ffffffffb97f153f>] pci_iov_remove_virtfn+0xaf/0x160 [ 194.362496] [<ffffffffb97f1aec>] sriov_disable+0x3c/0xf0 [ 194.362534] [<ffffffffb97f1bc3>] pci_disable_sriov+0x23/0x30 [ 194.362599] [<ffffffffc02f83c3>] qed_sriov_disable+0x5e3/0x650 [qed] [ 194.362658] [<ffffffffb9622df6>] ? kfree+0x106/0x140 [ 194.362709] [<ffffffffc02cc0c0>] ? qed_free_stream_mem+0x70/0x90 [qed] [ 194.362754] [<ffffffffb9622df6>] ? kfree+0x106/0x140 [ 194.362803] [<ffffffffc02cd659>] qed_slowpath_stop+0x1a9/0x1d0 [qed] [ 194.362854] [<ffffffffc035544e>] __qede_remove+0xae/0x130 [qede] [ 194.362904] [<ffffffffc03554e0>] qede_shutdown+0x10/0x20 [qede] [ 194.362956] [<ffffffffb97cf90a>] pci_device_shutdown+0x3a/0x60 [ 194.363010] [<ffffffffb98b180b>] device_shutdown+0xfb/0x1f0 [ 194.363066] [<ffffffffb94b66c6>] kernel_restart_prepare+0x36/0x40 [ 194.363107] [<ffffffffb94b66e2>] kernel_restart+0x12/0x60 [ 194.363146] [<ffffffffb94b6959>] SYSC_reboot+0x229/0x260 [ 194.363196] [<ffffffffb95f200d>] ? handle_mm_fault+0x39d/0x9b0 [ 194.363253] [<ffffffffb942b621>] ? __switch_to+0x151/0x580 [ 194.363304] [<ffffffffb9b7ec28>] ? __schedule+0x448/0x9c0 [ 194.363343] [<ffffffffb94b69fe>] SyS_reboot+0xe/0x10 [ 194.363387] [<ffffffffb9b8bede>] system_call_fastpath+0x25/0x2a [ 194.363430] Code: f9 e9 37 ff ff ff 90 0f 1f 44 00 00 55 48 89 e5 41 57 41 56 41 55 4c 8d af 98 00 00 00 41 54 4c 89 ef 41 89 f4 53 e8 4c e4 55 f9 <80> b8 dc 08 00 00 01 48 89 c3 4c 8d b8 c0 08 00 00 4c 8b b0 c0 [ 194.363712] RIP [<ffffffffc03553c4>] __qede_remove+0x24/0x130 [qede] [ 194.363764] RSP <ffff9ceabebdfac0> [ 194.363791] CR2: 00000000000008dc Signed-off-by: NManish Chopra <manishc@marvell.com> Signed-off-by: NAriel Elior <aelior@marvell.com> Signed-off-by: NSudarsana Kalluru <skalluru@marvell.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Alexander Sverdlin 提交于
Octeon's input ring-buffer entry has 14 bits-wide size field, so to account for second possible VLAN header max_mtu must be further reduced. Fixes: 109cc165 ("ethernet/cavium: use core min/max MTU checking") Signed-off-by: NAlexander Sverdlin <alexander.sverdlin@nokia.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Uwe Kleine-König 提交于
The owner member of struct pwm_ops must be set to THIS_MODULE to increase the reference count of the module such that the module cannot be removed while its code is in use. Fixes: daa5abc4 ("pwm: Add support for Broadcom iProc PWM controller") Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de> Reviewed-by: NFlorian Fainelli <f.fainelli@gmail.com> Signed-off-by: NThierry Reding <thierry.reding@gmail.com>
-