Skip to content

cloud-kernel
cloud-kernel

openanolis / cloud-kernel
1 年多 前同步成功

通知 161
Star 36
Fork 7
cloud-kernel
cloud-kernel
切换分支/标签
  1. 06 3月, 2019 20 次提交
    • P
      x86/mm/mem_encrypt: Fix erroneous sizeof() · 7ff77864
      由 Peng Hao 提交于 
      [ Upstream commit bf7d28c53453ea904584960de55e33e03b9d93b1 ]

Using sizeof(pointer) for determining the size of a memset() only works
when the size of the pointer and the size of type to which it points are
the same. For pte_t this is only true for 64bit and 32bit-NONPAE. On 32bit
PAE systems this is wrong as the pointer size is 4 byte but the PTE entry
is 8 bytes. It's actually not a real world issue as this code depends on
64bit, but it's wrong nevertheless.

Use sizeof(*p) for correctness sake.

Fixes: aad98391 ("x86/mm/encrypt: Simplify sme_populate_pgd() and sme_populate_pgd_large()")
Signed-off-by: NPeng Hao <peng.hao2@zte.com.cn>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Cc: dave.hansen@linux.intel.com
Cc: peterz@infradead.org
Cc: luto@kernel.org
Link: https://lkml.kernel.org/r/1546065252-97996-1-git-send-email-peng.hao2@zte.com.cnSigned-off-by: NSasha Levin <sashal@kernel.org>
      7ff77864
    • S
      genirq: Make sure the initial affinity is not empty · 17fab891
      由 Srinivas Ramana 提交于 
      [ Upstream commit bddda606ec76550dd63592e32a6e87e7d32583f7 ]

If all CPUs in the irq_default_affinity mask are offline when an interrupt
is initialized then irq_setup_affinity() can set an empty affinity mask for
a newly allocated interrupt.

Fix this by falling back to cpu_online_mask in case the resulting affinity
mask is zero.
Signed-off-by: NSrinivas Ramana <sramana@codeaurora.org>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: linux-arm-msm@vger.kernel.org
Link: https://lkml.kernel.org/r/1545312957-8504-1-git-send-email-sramana@codeaurora.orgSigned-off-by: NSasha Levin <sashal@kernel.org>
      17fab891
    • A
      selftests: rtc: rtctest: add alarm test on minute boundary · 7746dd64
      由 Alexandre Belloni 提交于 
      [ Upstream commit 7b3027728f4d4f6763f4d7e771acfc9424cdd0e6 ]

Unfortunately, some RTC don't have a second resolution for alarm so also
test for alarm on a minute boundary.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: NShuah Khan <shuah@kernel.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      7746dd64
    • A
      selftests: rtc: rtctest: fix alarm tests · 2409a869
      由 Alexandre Belloni 提交于 
      [ Upstream commit fdac94489c4d247088b3885875b39b3e1eb621ef ]

Return values for select are not checked properly and timeouts may not be
detected.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: NShuah Khan <shuah@kernel.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      2409a869
    • D
      usb: gadget: Potential NULL dereference on allocation error · 4670e839
      由 Dan Carpenter 提交于 
      [ Upstream commit df28169e1538e4a8bcd8b779b043e5aa6524545c ]

The source_sink_alloc_func() function is supposed to return error
pointers on error.  The function is called from usb_get_function() which
doesn't check for NULL returns so it would result in an Oops.

Of course, in the current kernel, small allocations always succeed so
this doesn't affect runtime.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      4670e839
    • Z
      usb: dwc3: gadget: Fix the uninitialized link_state when udc starts · 08c937f9
      由 Zeng Tao 提交于 
      [ Upstream commit 88b1bb1f3b88e0bf20b05d543a53a5b99bd7ceb6 ]

Currently the link_state is uninitialized and the default value is 0(U0)
before the first time we start the udc, and after we start the udc then
 stop the udc, the link_state will be undefined.
We may have the following warnings if we start the udc again with
an undefined link_state:

WARNING: CPU: 0 PID: 327 at drivers/usb/dwc3/gadget.c:294 dwc3_send_gadget_ep_cmd+0x304/0x308
dwc3 100e0000.hidwc3_0: wakeup failed --> -22
[...]
Call Trace:
[<c010f270>] (unwind_backtrace) from [<c010b3d8>] (show_stack+0x10/0x14)
[<c010b3d8>] (show_stack) from [<c034a4dc>] (dump_stack+0x84/0x98)
[<c034a4dc>] (dump_stack) from [<c0118000>] (__warn+0xe8/0x100)
[<c0118000>] (__warn) from [<c0118050>](warn_slowpath_fmt+0x38/0x48)
[<c0118050>] (warn_slowpath_fmt) from [<c0442ec0>](dwc3_send_gadget_ep_cmd+0x304/0x308)
[<c0442ec0>] (dwc3_send_gadget_ep_cmd) from [<c0445e68>](dwc3_ep0_start_trans+0x48/0xf4)
[<c0445e68>] (dwc3_ep0_start_trans) from [<c0446750>](dwc3_ep0_out_start+0x64/0x80)
[<c0446750>] (dwc3_ep0_out_start) from [<c04451c0>](__dwc3_gadget_start+0x1e0/0x278)
[<c04451c0>] (__dwc3_gadget_start) from [<c04452e0>](dwc3_gadget_start+0x88/0x10c)
[<c04452e0>] (dwc3_gadget_start) from [<c045ee54>](udc_bind_to_driver+0x88/0xbc)
[<c045ee54>] (udc_bind_to_driver) from [<c045f29c>](usb_gadget_probe_driver+0xf8/0x140)
[<c045f29c>] (usb_gadget_probe_driver) from [<bf005424>](gadget_dev_desc_UDC_store+0xac/0xc4 [libcomposite])
[<bf005424>] (gadget_dev_desc_UDC_store [libcomposite]) from[<c023d8e0>] (configfs_write_file+0xd4/0x160)
[<c023d8e0>] (configfs_write_file) from [<c01d51e8>] (__vfs_write+0x1c/0x114)
[<c01d51e8>] (__vfs_write) from [<c01d5ff4>] (vfs_write+0xa4/0x168)
[<c01d5ff4>] (vfs_write) from [<c01d6d40>] (SyS_write+0x3c/0x90)
[<c01d6d40>] (SyS_write) from [<c0107400>] (ret_fast_syscall+0x0/0x3c)
Signed-off-by: NZeng Tao <prime.zeng@hisilicon.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      08c937f9
    • B
      usb: dwc3: gadget: synchronize_irq dwc irq in suspend · 03a5d4d5
      由 Bo He 提交于 
      [ Upstream commit 01c10880d24291a96a4ab0da773e3c5ce4d12da8 ]

We see dwc3 endpoint stopped by unwanted irq during
suspend resume test, which is caused dwc3 ep can't be started
with error "No Resource".

Here, add synchronize_irq before suspend to sync the
pending IRQ handlers complete.
Signed-off-by: NBo He <bo.he@intel.com>
Signed-off-by: NYu Wang <yu.y.wang@intel.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      03a5d4d5
    • D
      thermal: int340x_thermal: Fix a NULL vs IS_ERR() check · f29024c0
      由 Dan Carpenter 提交于 
      [ Upstream commit 3fe931b31a4078395c1967f0495dcc9e5ec6b5e3 ]

The intel_soc_dts_iosf_init() function doesn't return NULL, it returns
error pointers.

Fixes: 4d0dd6c1 ("Thermal/int340x/processor_thermal: Enable auxiliary DTS for Braswell")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NZhang Rui <rui.zhang@intel.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      f29024c0
    • M
      clk: vc5: Abort clock configuration without upstream clock · fc1073df
      由 Marek Vasut 提交于 
      [ Upstream commit 2137a109a5e39c2bdccfffe65230ed3fadbaac0e ]

In case the upstream clock are not set, which can happen in case the
VC5 has no valid upstream clock, the $src variable is used uninited
by regmap_update_bits(). Check for this condition and return -EINVAL
in such case.

Note that in case the VC5 has no valid upstream clock, the VC5 can
not operate correctly. That is a hardware property of the VC5. The
internal oscilator present in some VC5 models is also considered
upstream clock.
Signed-off-by: NMarek Vasut <marek.vasut+renesas@gmail.com>
Cc: Alexey Firago <alexey_firago@mentor.com>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Stephen Boyd <sboyd@kernel.org>
Cc: linux-renesas-soc@vger.kernel.org
[sboyd@kernel.org: Added comment about probe preventing this from
happening in the first place]
Signed-off-by: NStephen Boyd <sboyd@kernel.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      fc1073df
    • L
      clk: sysfs: fix invalid JSON in clk_dump · 71943c38
      由 Lubomir Rintel 提交于 
      [ Upstream commit c6e909972ef87aa2a479269f46b84126f99ec6db ]

Add a missing comma so that the output is valid JSON format again.

Fixes: 9fba738a ("clk: add duty cycle support")
Signed-off-by: NLubomir Rintel <lkundrak@v3.sk>
Signed-off-by: NStephen Boyd <sboyd@kernel.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      71943c38
    • D
      clk: tegra: dfll: Fix a potential Oop in remove() · acc934f5
      由 Dan Carpenter 提交于 
      [ Upstream commit d39eca547f3ec67140a5d765a426eb157b978a59 ]

If tegra_dfll_unregister() fails then "soc" is an error pointer.  We
should just return instead of dereferencing it.

Fixes: 1752c9ee ("clk: tegra: dfll: Fix drvdata overwriting issue")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NStephen Boyd <sboyd@kernel.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      acc934f5
    • Y
      ASoC: Variable "val" in function rt274_i2c_probe() could be uninitialized · 651023ed
      由 Yizhuo 提交于 
      [ Upstream commit 8c3590de0a378c2449fc1aec127cc693632458e4 ]

Inside function rt274_i2c_probe(), if regmap_read() function
returns -EINVAL, then local variable "val" leaves uninitialized
but used in if statement. This is potentially unsafe.
Signed-off-by: NYizhuo <yzhai003@ucr.edu>
Signed-off-by: NMark Brown <broonie@kernel.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      651023ed
    • D
      ALSA: compress: prevent potential divide by zero bugs · e7b2f9f2
      由 Dan Carpenter 提交于 
      [ Upstream commit 678e2b44c8e3fec3afc7202f1996a4500a50be93 ]

The problem is seen in the q6asm_dai_compr_set_params() function:

	ret = q6asm_map_memory_regions(dir, prtd->audio_client, prtd->phys,
				       (prtd->pcm_size / prtd->periods),
                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
				       prtd->periods);

In this code prtd->pcm_size is the buffer_size and prtd->periods comes
from params->buffer.fragments.  If we allow the number of fragments to
be zero then it results in a divide by zero bug.  One possible fix would
be to use prtd->pcm_count directly instead of using the division to
re-calculate it.  But I decided that it doesn't really make sense to
allow zero fragments.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NMark Brown <broonie@kernel.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      e7b2f9f2
    • R
      ASoC: Intel: Haswell/Broadwell: fix setting for .dynamic field · a4964959
      由 Rander Wang 提交于 
      [ Upstream commit 906a9abc5de73c383af518f5a806f4be2993a0c7 ]

For some reason this field was set to zero when all other drivers use
.dynamic = 1 for front-ends. This change was tested on Dell XPS13 and
has no impact with the existing legacy driver. The SOF driver also works
with this change which enables it to override the fixed topology.
Signed-off-by: NRander Wang <rander.wang@linux.intel.com>
Acked-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NMark Brown <broonie@kernel.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      a4964959
    • K
      drm/msm: Unblock writer if reader closes file · 5a700533
      由 Kristian H. Kristensen 提交于 
      [ Upstream commit 99c66bc051e7407fe0bf0607b142ec0be1a1d1dd ]

Prevents deadlock when fifo is full and reader closes file.
Signed-off-by: NKristian H. Kristensen <hoegsberg@chromium.org>
Signed-off-by: NRob Clark <robdclark@gmail.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>
      5a700533
    • J
      scsi: libsas: Fix rphy phy_identifier for PHYs with end devices attached · 0f978ec3
      由 John Garry 提交于 
      commit ffeafdd2bf0b280d67ec1a47ea6287910d271f3f upstream.

The sysfs phy_identifier attribute for a sas_end_device comes from the rphy
phy_identifier value.

Currently this is not being set for rphys with an end device attached, so
we see incorrect symlinks from systemd disk/by-path:

root@localhost:~# ls -l /dev/disk/by-path/
total 0
lrwxrwxrwx 1 root root  9 Feb 13 12:26 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy0-lun-0 -> ../../sdb
lrwxrwxrwx 1 root root 10 Feb 13 12:26 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy0-lun-0-part1 -> ../../sdb1
lrwxrwxrwx 1 root root 10 Feb 13 12:26 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy0-lun-0-part2 -> ../../sdb2
lrwxrwxrwx 1 root root 10 Feb 13 12:26 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy0-lun-0-part3 -> ../../sdc3

Indeed, each sas_end_device phy_identifier value is 0:

root@localhost:/# more sys/class/sas_device/end_device-0\:0\:2/phy_identifier
0
root@localhost:/# more sys/class/sas_device/end_device-0\:0\:10/phy_identifier
0

This patch fixes the discovery code to set the phy_identifier.  With this,
we now get proper symlinks:

root@localhost:~# ls -l /dev/disk/by-path/
total 0
lrwxrwxrwx 1 root root  9 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy10-lun-0 -> ../../sdg
lrwxrwxrwx 1 root root  9 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy11-lun-0 -> ../../sdh
lrwxrwxrwx 1 root root  9 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy2-lun-0 -> ../../sda
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy2-lun-0-part1 -> ../../sda1
lrwxrwxrwx 1 root root  9 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy3-lun-0 -> ../../sdb
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy3-lun-0-part1 -> ../../sdb1
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy3-lun-0-part2 -> ../../sdb2
lrwxrwxrwx 1 root root  9 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy4-lun-0 -> ../../sdc
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy4-lun-0-part1 -> ../../sdc1
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy4-lun-0-part2 -> ../../sdc2
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy4-lun-0-part3 -> ../../sdc3
lrwxrwxrwx 1 root root  9 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy5-lun-0 -> ../../sdd
lrwxrwxrwx 1 root root  9 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy7-lun-0 -> ../../sde
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy7-lun-0-part1 -> ../../sde1
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy7-lun-0-part2 -> ../../sde2
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy7-lun-0-part3 -> ../../sde3
lrwxrwxrwx 1 root root  9 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy8-lun-0 -> ../../sdf
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy8-lun-0-part1 -> ../../sdf1
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy8-lun-0-part2 -> ../../sdf2
lrwxrwxrwx 1 root root 10 Feb 13 11:53 platform-HISI0162:01-sas-exp0x500e004aaaaaaa1f-phy8-lun-0-part3 -> ../../sdf3

Fixes: 2908d778 ("[SCSI] aic94xx: new driver")
Reported-by: Ndann frazier <dann.frazier@canonical.com>
Signed-off-by: NJohn Garry <john.garry@huawei.com>
Reviewed-by: NJason Yan <yanaijie@huawei.com>
Tested-by: Ndann frazier <dann.frazier@canonical.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      0f978ec3
    • T
      mac80211: Change default tx_sk_pacing_shift to 7 · a7c6cf3b
      由 Toke Høiland-Jørgensen 提交于 
      commit 5c14a4d05f68415af9e41a4e667d1748d41d1baf upstream.

When we did the original tests for the optimal value of sk_pacing_shift, we
came up with 6 ms of buffering as the default. Sadly, 6 is not a power of
two, so when picking the shift value I erred on the size of less buffering
and picked 4 ms instead of 8. This was probably wrong; those 2 ms of extra
buffering makes a larger difference than I thought.

So, change the default pacing shift to 7, which corresponds to 8 ms of
buffering. The point of diminishing returns really kicks in after 8 ms, and
so having this as a default should cut down on the need for extensive
per-device testing and overrides needed in the drivers.

Cc: stable@vger.kernel.org
Signed-off-by: NToke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      a7c6cf3b
    • L
      genirq/matrix: Improve target CPU selection for managed interrupts. · 765c30b3
      由 Long Li 提交于 
      [ Upstream commit e8da8794a7fd9eef1ec9a07f0d4897c68581c72b ]

On large systems with multiple devices of the same class (e.g. NVMe disks,
using managed interrupts), the kernel can affinitize these interrupts to a
small subset of CPUs instead of spreading them out evenly.

irq_matrix_alloc_managed() tries to select the CPU in the supplied cpumask
of possible target CPUs which has the lowest number of interrupt vectors
allocated.

This is done by searching the CPU with the highest number of available
vectors. While this is correct for non-managed CPUs it can select the wrong
CPU for managed interrupts. Under certain constellations this results in
affinitizing the managed interrupts of several devices to a single CPU in
a set.

The book keeping of available vectors works the following way:

 1) Non-managed interrupts:

    available is decremented when the interrupt is actually requested by
    the device driver and a vector is assigned. It's incremented when the
    interrupt and the vector are freed.

 2) Managed interrupts:

    Managed interrupts guarantee vector reservation when the MSI/MSI-X
    functionality of a device is enabled, which is achieved by reserving
    vectors in the bitmaps of the possible target CPUs. This reservation
    decrements the available count on each possible target CPU.

    When the interrupt is requested by the device driver then a vector is
    allocated from the reserved region. The operation is reversed when the
    interrupt is freed by the device driver. Neither of these operations
    affect the available count.

    The reservation persist up to the point where the MSI/MSI-X
    functionality is disabled and only this operation increments the
    available count again.

For non-managed interrupts the available count is the correct selection
criterion because the guaranteed reservations need to be taken into
account. Using the allocated counter could lead to a failing allocation in
the following situation (total vector space of 10 assumed):

		 CPU0	CPU1
 available:	    2	   0
 allocated:	    5	   3   <--- CPU1 is selected, but available space = 0
 managed reserved:  3	   7

 while available yields the correct result.

For managed interrupts the available count is not the appropriate
selection criterion because as explained above the available count is not
affected by the actual vector allocation.

The following example illustrates that. Total vector space of 10
assumed. The starting point is:

		 CPU0	CPU1
 available:	    5	   4
 allocated:	    2	   3
 managed reserved:  3	   3

 Allocating vectors for three non-managed interrupts will result in
 affinitizing the first two to CPU0 and the third one to CPU1 because the
 available count is adjusted with each allocation:

		  CPU0	CPU1
 available:	     5	   4	<- Select CPU0 for 1st allocation
 --> allocated:	     3	   3

 available:	     4	   4	<- Select CPU0 for 2nd allocation
 --> allocated:	     4	   3

 available:	     3	   4	<- Select CPU1 for 3rd allocation
 --> allocated:	     4	   4

 But the allocation of three managed interrupts starting from the same
 point will affinitize all of them to CPU0 because the available count is
 not affected by the allocation (see above). So the end result is:

		  CPU0	CPU1
 available:	     5	   4
 allocated:	     5	   3

Introduce a "managed_allocated" field in struct cpumap to track the vector
allocation for managed interrupts separately. Use this information to
select the target CPU when a vector is allocated for a managed interrupt,
which results in more evenly distributed vector assignments. The above
example results in the following allocations:

		 CPU0	CPU1
 managed_allocated: 0	   0	<- Select CPU0 for 1st allocation
 --> allocated:	    3	   3

 managed_allocated: 1	   0	<- Select CPU1 for 2nd allocation
 --> allocated:	    3	   4

 managed_allocated: 1	   1	<- Select CPU0 for 3rd allocation
 --> allocated:	    4	   4

The allocation of non-managed interrupts is not affected by this change and
is still evaluating the available count.

The overall distribution of interrupt vectors for both types of interrupts
might still not be perfectly even depending on the number of non-managed
and managed interrupts in a system, but due to the reservation guarantee
for managed interrupts this cannot be avoided.

Expose the new field in debugfs as well.

[ tglx: Clarified the background of the problem in the changelog and
  	described it independent of NVME ]
Signed-off-by: NLong Li <longli@microsoft.com>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Michael Kelley <mikelley@microsoft.com>
Link: https://lkml.kernel.org/r/20181106040000.27316-1-longli@linuxonhyperv.comSigned-off-by: NSasha Levin <sashal@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      765c30b3
    • D
      irq/matrix: Spread managed interrupts on allocation · 8cae7757
      由 Dou Liyang 提交于 
      [ Upstream commit 76f99ae5b54d48430d1f0c5512a84da0ff9761e0 ]

Linux spreads out the non managed interrupt across the possible target CPUs
to avoid vector space exhaustion.

Managed interrupts are treated differently, as for them the vectors are
reserved (with guarantee) when the interrupt descriptors are initialized.

When the interrupt is requested a real vector is assigned. The assignment
logic uses the first CPU in the affinity mask for assignment. If the
interrupt has more than one CPU in the affinity mask, which happens when a
multi queue device has less queues than CPUs, then doing the same search as
for non managed interrupts makes sense as it puts the interrupt on the
least interrupt plagued CPU. For single CPU affine vectors that's obviously
a NOOP.

Restructre the matrix allocation code so it does the 'best CPU' search, add
the sanity check for an empty affinity mask and adapt the call site in the
x86 vector management code.

[ tglx: Added the empty mask check to the core and improved change log ]
Signed-off-by: NDou Liyang <douly.fnst@cn.fujitsu.com>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: hpa@zytor.com
Link: https://lkml.kernel.org/r/20180908175838.14450-2-dou_liyang@163.comSigned-off-by: NSasha Levin <sashal@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      8cae7757
    • D
      irq/matrix: Split out the CPU selection code into a helper · 2948b887
      由 Dou Liyang 提交于 
      [ Upstream commit 8ffe4e61c06a48324cfd97f1199bb9838acce2f2 ]

Linux finds the CPU which has the lowest vector allocation count to spread
out the non managed interrupts across the possible target CPUs, but does
not do so for managed interrupts.

Split out the CPU selection code into a helper function for reuse. No
functional change.
Signed-off-by: NDou Liyang <douly.fnst@cn.fujitsu.com>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: hpa@zytor.com
Link: https://lkml.kernel.org/r/20180908175838.14450-1-dou_liyang@163.comSigned-off-by: NSasha Levin <sashal@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      2948b887
  3. 27 2月, 2019 20 次提交