Currently, utils.c is only useful if CONFIG_EROFS_FS_ZIP is on.
So let's rename it to zutil.c as well as avoid its inclusion if
CONFIG_EROFS_FS_ZIP is explicitly disabled.
Signed-off-by: Chunhai Guo <guochunhai@vivo.com>
Reviewed-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Link: https://lore.kernel.org/r/20240401135550.2550043-1-guochunhai@vivo.com
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Bug: 348591003
Change-Id: Iace477719fb803084955218a3e5ace74338f5ec8
(cherry picked from commit cacd5b04e24c74a813c694ec7b26a1a370b5d666)
[Chunhai: Resolved minor conflict in fs/erofs/Makefile and
fs/erofs/utils.c ]
Signed-off-by: Chunhai Guo <guochunhai@vivo.com>
Even with inplace decompression, sometimes very few temporary buffers
may be still needed for a single decompression shot (e.g. 16 pages for
64k sliding window or 4 pages for 16k sliding window). In low-memory
scenarios, it would be better to try to allocate with GFP_NOWAIT on
readahead first. That can help reduce the time spent on page allocation
under durative memory pressure.
Here are detailed performance numbers under multi-app launch benchmark
workload [1] on ARM64 Android devices (8-core CPU and 8GB of memory)
running a 5.15 LTS kernel with EROFS of 4k pclusters:
+----------------------------------------------+
| LZ4 | vanilla | patched | diff |
|----------------+---------+---------+---------|
| Average (ms) | 3364 | 2684 | -20.21% | [64k sliding window]
|----------------+---------+---------+---------|
| Average (ms) | 2079 | 1610 | -22.56% | [16k sliding window]
+----------------------------------------------+
The total size of system images for 4k pclusters is almost unchanged:
(64k sliding window) 9,117,044 KB
(16k sliding window) 9,113,096 KB
Therefore, in addition to switch the sliding window from 64k to 16k,
after applying this patch, it can eventually save 52.14% (3364 -> 1610)
on average with no memory reservation. That is particularly useful for
embedded devices with limited resources.
[1] https://lore.kernel.org/r/20240109074143.4138783-1-guochunhai@vivo.com
Suggested-by: Gao Xiang <xiang@kernel.org>
Signed-off-by: Chunhai Guo <guochunhai@vivo.com>
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Reviewed-by: Yue Hu <huyue2@coolpad.com>
Link: https://lore.kernel.org/r/20240126140142.201718-1-hsiangkao@linux.alibaba.com
Bug: 348591003
Change-Id: I84cf5b5dbccbd707b6f4339e525ca52efbcf167f
(cherry picked from commit d9281660ff3ffb4a05302b485cc59a87e709aefc)
[Chunhai: Resolved minor conflict in fs/erofs/zdata.c and remove
modification on fs/erofs/decompressor_deflate.c as it is not existed in
6.1 kernel ]
Signed-off-by: Chunhai Guo <guochunhai@vivo.com>
The padding VMA is never inserted into the VMA tree; therefore we
don't need to have the mmap lock in exclusive mode to modify it.
Test: v2/android-gki/ack_platform_integration_main_cf_arm64_boot_test
on kernel_virt_debug_aarch64
Bug: 346741763
Change-Id: I4ca3ed22dab45b6bb895cb41c5c6792344188b61
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
ufshcd_queuecommand() may be called two times in a row for a SCSI command
before it is completed. Hence make the following changes:
- In the functions that submit a command, do not check the old value of
lrbp->cmd nor clear lrbp->cmd in error paths.
- In ufshcd_release_scsi_cmd(), do not clear lrbp->cmd.
See also scsi_send_eh_cmnd().
This commit prevents that the following appears if a command times out:
WARNING: at drivers/ufs/core/ufshcd.c:2965 ufshcd_queuecommand+0x6f8/0x9a8
Call trace:
ufshcd_queuecommand+0x6f8/0x9a8
scsi_send_eh_cmnd+0x2c0/0x960
scsi_eh_test_devices+0x100/0x314
scsi_eh_ready_devs+0xd90/0x114c
scsi_error_handler+0x2b4/0xb70
kthread+0x16c/0x1e0
Fixes: 5a0b0cb9be ("[SCSI] ufs: Add support for sending NOP OUT UPIU")
Change-Id: Iebc077bc8ae8d8004519ebfc93a1c76a4807b6f2
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20230524203659.1394307-3-bvanassche@acm.org
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
(cherry picked from commit 549e91a9bb)
[Peter: Resolved minor conflict in drivers/ufs/core/ufshcd.c ]
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
When the lock is owned by readers and there is no RWSEM_FLAG_HANDOFF set,
we allow some specific new readers to acquire the lock immediately
in this hook, event if there are some writer tasks in the wait_list.
This features can optimize the priority inversion problem caused by
rwsem and improve system responsiveness and
performance.
Bug: 348699619
Bug: 336506800
Change-Id: I7e8fded73579933d1f61faa9fb6e5f300ffd71bf
Signed-off-by: zhujingpeng <zhujingpeng@vivo.com>
[jstultz: rebased, resolved collision]
Signed-off-by: John Stultz <jstultz@google.com>
Add hooks to fuse queue request and request end so we can do boost
to those background tasks which block the UX related task.
Bug: 333220630
Change-Id: I9be59ed88675c5102c57ba9cbd26cf4df3d2fd7f
Signed-off-by: liliangliang <liliangliang@vivo.com>
(cherry picked from commit e520c2932df0d1bbf83ae45c82ac01fd41655d77)
HOSTCC is always wanted when building objtool. Setting CC to HOSTCC
happens after tools/scripts/Makefile.include is included, meaning
flags (like CFLAGS) are set assuming say CC is gcc, but then it can be
later set to HOSTCC which may be clang. tools/scripts/Makefile.include
is needed for host set up and common macros in objtool's
Makefile. Rather than override the CC variable to HOSTCC, just pass CC
as HOSTCC to the sub-makes of Makefile.build, the libsubcmd builds and
also to the linkage step.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20230126190606.40739-4-irogers@google.com
Signed-off-by: Josh Poimboeuf <jpoimboe@kernel.org>
Bug: 336872347
Bug: 335829879
Test: build x86_64 kernel with glibc 2.38
Change-Id: I1d672d0bb64f72d3fc571537de5f75d4068e79cc
(cherry picked from commit cd955bdd6a)
Signed-off-by: Yifan Hong <elsk@google.com>
Including from tools/lib can create inadvertent dependencies. Install
libsubcmd in the objtool build and then include the headers from
there.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20230126190606.40739-2-irogers@google.com
Signed-off-by: Josh Poimboeuf <jpoimboe@kernel.org>
Bug: 336872347
Bug: 335829879
Test: build x86_64 kernel with glibc 2.38
Change-Id: Id09c5b222519073214dbc01e151e59b18afb1ea8
(cherry picked from commit bdb8bf7d56)
Signed-off-by: Yifan Hong <elsk@google.com>
[ Upstream commit 22dd70eb2c3d754862964377a75abafd3167346b ]
Currently, we can read OOB data without MSG_OOB by using MSG_PEEK
when OOB data is sitting on the front row, which is apparently
wrong.
>>> from socket import *
>>> c1, c2 = socketpair(AF_UNIX, SOCK_STREAM)
>>> c1.send(b'a', MSG_OOB)
1
>>> c2.recv(1, MSG_PEEK | MSG_DONTWAIT)
b'a'
If manage_oob() is called when no data has been copied, we only
check if the socket enables SO_OOBINLINE or MSG_PEEK is not used.
Otherwise, the skb is returned as is.
However, here we should return NULL if MSG_PEEK is set and no data
has been copied.
Also, in such a case, we should not jump to the redo label because
we will be caught in the loop and hog the CPU until normal data
comes in.
Then, we need to handle skb == NULL case with the if-clause below
the manage_oob() block.
With this patch:
>>> from socket import *
>>> c1, c2 = socketpair(AF_UNIX, SOCK_STREAM)
>>> c1.send(b'a', MSG_OOB)
1
>>> c2.recv(1, MSG_PEEK | MSG_DONTWAIT)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
BlockingIOError: [Errno 11] Resource temporarily unavailable
Bug: 342490466
Fixes: 314001f0bf ("af_unix: Add OOB support")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20240410171016.7621-3-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
(cherry picked from commit 022d81a709)
Signed-off-by: Lee Jones <joneslee@google.com>
Change-Id: I4728977f8f908c19dfa5c861c7381a50499b7fe0
[ Upstream commit b46f4eaa4f0ec38909fb0072eea3aeddb32f954e ]
syzkaller started to report deadlock of unix_gc_lock after commit
4090fa373f0e ("af_unix: Replace garbage collection algorithm."), but
it just uncovers the bug that has been there since commit 314001f0bf
("af_unix: Add OOB support").
The repro basically does the following.
from socket import *
from array import array
c1, c2 = socketpair(AF_UNIX, SOCK_STREAM)
c1.sendmsg([b'a'], [(SOL_SOCKET, SCM_RIGHTS, array("i", [c2.fileno()]))], MSG_OOB)
c2.recv(1) # blocked as no normal data in recv queue
c2.close() # done async and unblock recv()
c1.close() # done async and trigger GC
A socket sends its file descriptor to itself as OOB data and tries to
receive normal data, but finally recv() fails due to async close().
The problem here is wrong handling of OOB skb in manage_oob(). When
recvmsg() is called without MSG_OOB, manage_oob() is called to check
if the peeked skb is OOB skb. In such a case, manage_oob() pops it
out of the receive queue but does not clear unix_sock(sk)->oob_skb.
This is wrong in terms of uAPI.
Let's say we send "hello" with MSG_OOB, and "world" without MSG_OOB.
The 'o' is handled as OOB data. When recv() is called twice without
MSG_OOB, the OOB data should be lost.
>>> from socket import *
>>> c1, c2 = socketpair(AF_UNIX, SOCK_STREAM, 0)
>>> c1.send(b'hello', MSG_OOB) # 'o' is OOB data
5
>>> c1.send(b'world')
5
>>> c2.recv(5) # OOB data is not received
b'hell'
>>> c2.recv(5) # OOB date is skipped
b'world'
>>> c2.recv(5, MSG_OOB) # This should return an error
b'o'
In the same situation, TCP actually returns -EINVAL for the last
recv().
Also, if we do not clear unix_sk(sk)->oob_skb, unix_poll() always set
EPOLLPRI even though the data has passed through by previous recv().
To avoid these issues, we must clear unix_sk(sk)->oob_skb when dequeuing
it from recv queue.
The reason why the old GC did not trigger the deadlock is because the
old GC relied on the receive queue to detect the loop.
When it is triggered, the socket with OOB data is marked as GC candidate
because file refcount == inflight count (1). However, after traversing
all inflight sockets, the socket still has a positive inflight count (1),
thus the socket is excluded from candidates. Then, the old GC lose the
chance to garbage-collect the socket.
With the old GC, the repro continues to create true garbage that will
never be freed nor detected by kmemleak as it's linked to the global
inflight list. That's why we couldn't even notice the issue.
Bug: 342490466
Fixes: 314001f0bf ("af_unix: Add OOB support")
Reported-by: syzbot+7f7f201cc2668a8fd169@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=7f7f201cc2668a8fd169
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20240405221057.2406-1-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
(cherry picked from commit 601a89ea24)
Signed-off-by: Lee Jones <joneslee@google.com>
Change-Id: Ib4a11eed6b5710d9934d4f31cd29dfd4c7b3658f
The correct fragment is the one in build/kernel,
enabled by --page_size or kernel_build.page_size.
This fragment:
- Does not correctly enable incremental FS
- Does not correctly clear LOCALVERSION that
has 4k for the 4k build.
Bug: 347036722
Bug: 340631213
Bug: 338659380
Change-Id: I31cb004ca639d8ec3dd6201112391c7214971eba
Signed-off-by: Yifan Hong <elsk@google.com>
Recent changes to gki_defconfig enable CONFIG_MEMORY which results
in CONFIG_BRCMSTB_DPFE and CONFIG_BRCMSTB_MEMC being enabled as
builtin modules. Disable the two drivers as the Broadcom STB SoCs
expect them to be configured as loadable modules.
Bug: 346510475
Fixes: 974a6f430e ("ANDROID: gki_defconfig: Enable Tegra SoCs")
Change-Id: I80bf0ce419992d947468b7054327ba80f611d316
Signed-off-by: Danesh Petigara <danesh.petigara@broadcom.com>
Signed-off-by: Pierre Couillaud <pierre@broadcom.com>
Depending on the platform binary being executed, the linker
(interpreter) requested can be one of:
1) /system/bin/bootstrap/linker64
2) /system/bin/linker64
3) /apex/com.android.runtime/bin/linker64
Relax the check to the basename (linker64), instead of the path.
Bug: 330767927
Bug: 335584973
Change-Id: I4a1f95b7cecd126f85ad8cefd9ff10d272947f9e
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
Usually to modify vm_flags we need to take exclusive mmap_lock but here
only have the lock in read mode, to avoid all DONTNEED/DONTNEED_LOCKED
calls needing the write lock.
A race to the flags update can only happen with another MADV_DONTNEED on
the same process and same range (VMA).
In practice, this specific scenario is not possible because the action
that could cause it is usually performed at most once per VMA and only by the dynamic linker.
Forego protection for this case, to avoid penalties in the common cases.
Bug: 344634072
Change-Id: I54ac1f204e0445291f3df3872fbaa16b37722812
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
This introduces a new environment variable, KCPPFLAGS_COMPAT.
One use-case is to ensure -ffile-prefix-map is passed to the arm32
compiler to normalise compilation directory and make the ELF build ID
reproducible.
Bug: 345452375
Change-Id: I6ae1df58172f4dadeac1dbbee2e3241b704a9256
Signed-off-by: Giuliano Procida <gprocida@google.com>
Enable configs required for Broadcom Set Top Box SoCs.
Also explicitly disable the Broadcom Set Top Box SoC
drivers that we wish to build as loadable modules.
Bug: 343006993
Change-Id: Ia3e542d0eeb519539aba2990fb4eb4eb2970a867
Signed-off-by: Danesh Petigara <danesh.petigara@broadcom.com>
Requesting a retune before switching to the RPMB partition has been
observed to cause CRC errors on the RPMB reads (-EILSEQ).
Since RPMB reads can not be retried, the clients would be directly
affected by the errors.
This commit disables the retune request prior to switching to the RPMB
partition: mmc_retune_pause() no longer triggers a retune before the
pause period begins.
This was verified with the sdhci-of-arasan driver (ZynqMP) configured
for HS200 using two separate eMMC cards (DG4064 and 064GB2). In both
cases, the error was easy to reproduce triggering every few tenths of
reads.
With this commit, systems that were utilizing OP-TEE to access RPMB
variables will experience an enhanced performance. Specifically, when
OP-TEE is configured to employ RPMB as a secure storage solution, it not
only writes the data but also the secure filesystem within the
partition. As a result, retrieving any variable involves multiple RPMB
reads, typically around five.
For context, on ZynqMP, each retune request consumed approximately
8ms. Consequently, reading any RPMB variable used to take at the very
minimum 40ms.
After droping the need to retune before switching to the RPMB partition,
this is no longer the case.
Change-Id: I2320a38436b05b435ac528c6656b621a1e2eaeb7
Signed-off-by: Jorge Ramirez-Ortiz <jorge@foundries.io>
Acked-by: Avri Altman <avri.altman@wdc.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20240103112911.2954632-1-jorge@foundries.io
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
(cherry picked from commit 67380251e8bbd3302c64fea07f95c31971b91c22)
Signed-off-by: Danesh Petigara <danesh.petigara@broadcom.com>
The driver used to truncate several 64-bit registers such as PMCEID[n]
registers used to describe whether architectural and microarchitectural
events in range 0x4000-0x401f exist. Due to discarding the bits, the
driver made the events invisible, even if they existed.
Moreover, PMCCFILTR and PMCR registers have additional bits in the upper
32 bits. This patch makes them available although they aren't currently
used. Finally, functions handling PMXEVCNTR and PMXEVTYPER registers are
removed as they not being used at all.
Fixes: df29ddf4f0 ("arm64: perf: Abstract system register accesses away")
Reported-by: Carl Worth <carl@os.amperecomputing.com>
Change-Id: Ic2ec717caa2fdda2456c560a82490e374565eb2d
Signed-off-by: Ilkka Koskinen <ilkka@os.amperecomputing.com>
Acked-by: Will Deacon <will@kernel.org>
Closes: https://lore.kernel.org/..
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Link: https://lore.kernel.org/r/20231102183012.1251410-1-ilkka@os.amperecomputing.com
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
(cherry picked from commit 403edfa436286b21f5ffe6856ae5b36396e8966c)
Signed-off-by: Danesh Petigara <danesh.petigara@broadcom.com>
A few PHY drivers are currently attempting to not suspend the PHY when
Wake-on-LAN is enabled, however that code is not currently executing at
all due to an early check in phy_suspend().
This prevents PHY drivers from making an appropriate decisions and put
the hardware into a low power state if desired.
In order to allow the PHY drivers to opt into getting their ->suspend
routine to be called, add a PHY_ALWAYS_CALL_SUSPEND bit which can be
set. A boolean that tracks whether the PHY or the attached MAC has
Wake-on-LAN enabled is also provided for convenience.
If phydev::wol_enabled then the PHY shall not prevent its own
Wake-on-LAN detection logic from working and shall not prevent the
Ethernet MAC from receiving packets for matching.
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Change-Id: I9077fe16d2515b60a2cb58753e6914be95a6923a
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit a7e3448086)
[danesh: Moved phydev->wol_enabled to avoid KMI-break]
Signed-off-by: Danesh Petigara <danesh.petigara@broadcom.com>
When just including <asm/arm_pmuv3.h>:
arch/arm/include/asm/arm_pmuv3.h:110:13: error: ‘write_pmevtypern’ defined but not used [-Werror=unused-function]
110 | static void write_pmevtypern(int n, unsigned long val)
| ^~~~~~~~~~~~~~~~
arch/arm/include/asm/arm_pmuv3.h:103:13: error: ‘write_pmevcntrn’ defined but not used [-Werror=unused-function]
103 | static void write_pmevcntrn(int n, unsigned long val)
| ^~~~~~~~~~~~~~~
arch/arm/include/asm/arm_pmuv3.h:95:22: error: ‘read_pmevcntrn’ defined but not used [-Werror=unused-function]
95 | static unsigned long read_pmevcntrn(int n)
| ^~~~~~~~~~~~~~
Fix this by adding the missing "inline" keywords to the three accessor
functions that lack them.
Fixes: 009d6dc87a ("ARM: perf: Allow the use of the PMUv3 driver on 32bit ARM")
Change-Id: I0877fa9d337a59857c8480a219515e2c641cfc55
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/3a7d9bc7470aa2d85696ee9765c74f8c03fb5454.1683561482.git.geert+renesas@glider.be
Signed-off-by: Will Deacon <will@kernel.org>
(cherry picked from commit 68e3f61eb9)
Signed-off-by: Danesh Petigara <danesh.petigara@broadcom.com>
When just including <asm/arm_pmuv3.h>:
arch/arm64/include/asm/arm_pmuv3.h:31:13: error: ‘write_pmevtypern’ defined but not used [-Werror=unused-function]
31 | static void write_pmevtypern(int n, unsigned long val)
| ^~~~~~~~~~~~~~~~
arch/arm64/include/asm/arm_pmuv3.h:24:13: error: ‘write_pmevcntrn’ defined but not used [-Werror=unused-function]
24 | static void write_pmevcntrn(int n, unsigned long val)
| ^~~~~~~~~~~~~~~
arch/arm64/include/asm/arm_pmuv3.h:16:22: error: ‘read_pmevcntrn’ defined but not used [-Werror=unused-function]
16 | static unsigned long read_pmevcntrn(int n)
| ^~~~~~~~~~~~~~
Fix this by adding the missing "inline" keywords to the three accessor
functions that lack them.
Fixes: df29ddf4f0 ("arm64: perf: Abstract system register accesses away")
Change-Id: I36e1360fc3b187b5e68619386ae9c142a9dda3c6
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Marc Zyngier <maz@kernel.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Link: https://lore.kernel.org/r/d53a19043c0c3bd25f6c203e73a2fb08a9661824.1683561482.git.geert+renesas@glider.be
Signed-off-by: Will Deacon <will@kernel.org>
(cherry picked from commit 3bc879e355)
Signed-off-by: Danesh Petigara <danesh.petigara@broadcom.com>
Reiji reports that the arm64 implementation of arch_perf_update_userpage()
is now ignored and replaced by the dummy stub in core code.
This seems to happen since the PMUv3 driver was moved to driver/perf.
As it turns out, dropping the __weak attribute from the *prototype*
of the function solves the problem. You're right, this doesn't seem
to make much sense. And yet... It appears that both symbols get
flagged as weak, and that the first one to appear in the link order
wins:
$ nm drivers/perf/arm_pmuv3.o|grep arch_perf_update_userpage
0000000000001db0 W arch_perf_update_userpage
Dropping the attribute from the prototype restores the expected
behaviour, and arm64 is able to enjoy arch_perf_update_userpage()
again.
Fixes: 7755cec63a ("arm64: perf: Move PMUv3 driver to drivers/perf")
Fixes: f1ec3a517b ("kernel/events: Add a missing prototype for arch_perf_update_userpage()")
Reported-by: Reiji Watanabe <reijiw@google.com>
Change-Id: I497041999776e66ac77c3fccdd09ce47aae544e9
Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Reiji Watanabe <reijiw@google.com>
Link: https://lkml.kernel.org/r/20230616114831.3186980-1-maz@kernel.org
(cherry picked from commit b50f26a448)
Signed-off-by: Danesh Petigara <danesh.petigara@broadcom.com>
Having the ARM PMUv3 driver sitting in arch/arm64/kernel is getting
in the way of being able to use perf on ARMv8 cores running a 32bit
kernel, such as 32bit KVM guests.
This patch moves it into drivers/perf/arm_pmuv3.c, with an include
file in include/linux/perf/arm_pmuv3.h. The only thing left in
arch/arm64 is some mundane perf stuff.
Change-Id: I1928ceb684a107ef991401175458f2ea6c711826
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Zaid Al-Bassam <zalbassam@google.com>
Tested-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20230317195027.3746949-2-zalbassam@google.com
Signed-off-by: Will Deacon <will@kernel.org>
(cherry picked from commit 7755cec63a)
Signed-off-by: Danesh Petigara <danesh.petigara@broadcom.com>
The logic in blk_mq_complete_need_ipi() assumes SMP systems where all
CPUs have equal compute capacities and only LLC cache can make
a different on perceived performance. But this assumption falls apart on
HMP systems where LLC is shared, but the CPUs have different capacities.
Staying local then can have a big performance impact if the IO request
was done from a CPU with higher capacity but the interrupt is serviced
on a lower capacity CPU.
Use the new cpus_equal_capacity() function to check if we need to send
an IPI.
Without the patch I see the BLOCK softirq always running on little cores
(where the hardirq is serviced). With it I can see it running on all
cores.
This was noticed after the topology change [1] where now on a big.LITTLE
we truly get that the LLC is shared between all cores where as in the
past it was being misrepresented for historical reasons. The logic
exposed a missing dependency on capacities for such systems where there
can be a big performance difference between the CPUs.
This of course introduced a noticeable change in behavior depending on
how the topology is presented. Leading to regressions in some workloads
as the performance of the BLOCK softirq on littles can be noticeably
worse on some platforms.
Worth noting that we could have checked for capacities being greater
than or equal instead for equality. This will lead to favouring higher
performance always. But opted for equality instead to match the
performance of the requester without making an assumption that can lead
to power trade-offs which these systems tend to be sensitive about. If
the requester would like to run faster, it's better to rely on the
scheduler to give the IO requester via some facility to run on a faster
core; and then if the interrupt triggered on a CPU with different
capacity we'll make sure to match the performance the requester is
supposed to run at.
[1] https://lpc.events/event/16/contributions/1342/attachments/962/1883/LPC-2022-Android-MC-Phantom-Domains.pdf
Bug: 341551538
Signed-off-by: Qais Yousef <qyousef@layalina.io>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20240223155749.2958009-3-qyousef@layalina.io
Signed-off-by: Jens Axboe <axboe@kernel.dk>
(cherry picked from commit af550e4c968294398fc76b075f12d51c76caf753)
Signed-off-by: Qais Yousef <qyousef@google.com>
(cherry picked from https://android-review.googlesource.com/q/commit:d01dc1c7d020f49f31194254a89694bd23ad3dd5)
Merged-In: Iefed6a19b9d25102642e264118431f5e12e23fea
Change-Id: Iefed6a19b9d25102642e264118431f5e12e23fea
