Commit f1a0a376ca ("sched/core: Initialize the idle task with
preemption disabled") removed the init_idle() call from
idle_thread_get(). This was the sole call-path on hotplug that resets
the Shadow Call Stack (scs) Stack Pointer (sp).
Not resetting the scs-sp leads to scs overflow after enough hotplug
cycles. Therefore add an explicit scs_task_reset() to the hotplug code
to make sure the scs-sp does get reset on hotplug.
Fixes: f1a0a376ca ("sched/core: Initialize the idle task with preemption disabled")
Signed-off-by: Woody Lin <woodylin@google.com>
[peterz: Changelog]
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lore.kernel.org/r/20211012083521.973587-1-woodylin@google.com
Merge misc fixes from Andrew Morton:
"19 patches.
Subsystems affected by this patch series: mm (userfaultfd, migration,
memblock, mempolicy, slub, secretmem, and thp), ocfs2, binfmt, vfs,
and misc"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
mailmap: add Andrej Shadura
mm/thp: decrease nr_thps in file's mapping on THP split
mm/secretmem: fix NULL page->mapping dereference in page_is_secretmem()
vfs: check fd has read access in kernel_read_file_from_fd()
elfcore: correct reference to CONFIG_UML
mm, slub: fix incorrect memcg slab count for bulk free
mm, slub: fix potential use-after-free in slab_debugfs_fops
mm, slub: fix potential memoryleak in kmem_cache_open()
mm, slub: fix mismatch between reconstructed freelist depth and cnt
mm, slub: fix two bugs in slab_debug_trace_open()
mm/mempolicy: do not allow illegal MPOL_F_NUMA_BALANCING | MPOL_LOCAL in mbind()
memblock: check memory total_size
ocfs2: mount fails with buffer overflow in strlen
ocfs2: fix data corruption after conversion from inline format
mm/migrate: fix CPUHP state to update node demotion order
mm/migrate: add CPU hotplug to demotion #ifdef
mm/migrate: optimize hotplug-time demotion order updates
userfaultfd: fix a race between writeprotect and exit_mmap()
mm/userfaultfd: selftests: fix memory corruption with thp enabled
Marc Kleine-Budde says:
====================
pull-request: can 2021-10-19
this is a pull request of a single patch for net/master.
The patch is by me and fixes the error handling in case of a FC
timeout in the TX path of the ISOTOP CAN protocol.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
During the process of driver probing, the probe function should return < 0
for failure, otherwise, the kernel will treat value > 0 as success.
Signed-off-by: Zheyu Ma <zheyuma97@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
During the process of driver probing, the probe function should return < 0
for failure, otherwise, the kernel will treat value > 0 as success.
Signed-off-by: Zheyu Ma <zheyuma97@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The Winbond MMC driver fails to build on ARCH=m68k so prevent
that build config. Silences these build errors:
../drivers/mmc/host/wbsd.c: In function 'wbsd_request_end':
../drivers/mmc/host/wbsd.c:212:28: error: implicit declaration of function 'claim_dma_lock' [-Werror=implicit-function-declaration]
212 | dmaflags = claim_dma_lock();
../drivers/mmc/host/wbsd.c:215:17: error: implicit declaration of function 'release_dma_lock'; did you mean 'release_task'? [-Werror=implicit-function-declaration]
215 | release_dma_lock(dmaflags);
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Pierre Ossman <pierre@ossman.eu>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20211017175949.23838-1-rdunlap@infradead.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
To reset standard tuning circuit completely, after clear ESDHC_MIX_CTRL_EXE_TUNE,
also need to clear bit buffer_read_ready, this operation will finally clear the
USDHC IP internal logic flag execute_tuning_with_clr_buf, make sure the following
normal data transfer will not be impacted by standard tuning logic used before.
Find this issue when do quick SD card insert/remove stress test. During standard
tuning prodedure, if remove SD card, USDHC standard tuning logic can't clear the
internal flag execute_tuning_with_clr_buf. Next time when insert SD card, all
data related commands can't get any data related interrupts, include data transfer
complete interrupt, data timeout interrupt, data CRC interrupt, data end bit interrupt.
Always trigger software timeout issue. Even reset the USDHC through bits in register
SYS_CTRL (0x2C, bit28 reset tuning, bit26 reset data, bit 25 reset command, bit 24
reset all) can't recover this. From the user's point of view, USDHC stuck, SD can't
be recognized any more.
Fixes: d9370424c9 ("mmc: sdhci-esdhc-imx: reset tuning circuit when power on mmc card")
Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/1634263236-6111-1-git-send-email-haibo.chen@nxp.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
In randconfig builds, we sometimes come across this warning:
arm-linux-gnueabi-ld: XIP start address may cause MPU programming issues
While this is helpful for actual systems to figure out why it
fails, the warning does not provide any benefit for build testing,
so guard it in a check for CONFIG_COMPILE_TEST, which is usually
set on randconfig builds.
Fixes: 216218308c ("ARM: 8713/1: NOMMU: Support MPU in XIP configuration")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
When frame pointers are used instead of the ARM unwinder,
and the kernel is built using clang with an external assembler
and CONFIG_XIP_KERNEL, every file produces two warnings
like:
arm-linux-gnueabi-ld: warning: orphan section `.ARM.extab' from `net/mac802154/util.o' being placed in section `.ARM.extab'
arm-linux-gnueabi-ld: warning: orphan section `.ARM.exidx' from `net/mac802154/util.o' being placed in section `.ARM.exidx'
The same fix was already merged for the normal (non-XIP)
linker script, with a longer description.
Fixes: c39866f268 ("arm/build: Always handle .ARM.exidx and .ARM.extab sections")
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Both the decompressor code and the kasan logic try to override
the memcpy() and memmove() definitions, which leading to a clash
in a KASAN-enabled kernel with XZ decompression:
arch/arm/boot/compressed/decompress.c:50:9: error: 'memmove' macro redefined [-Werror,-Wmacro-redefined]
#define memmove memmove
^
arch/arm/include/asm/string.h:59:9: note: previous definition is here
#define memmove(dst, src, len) __memmove(dst, src, len)
^
arch/arm/boot/compressed/decompress.c:51:9: error: 'memcpy' macro redefined [-Werror,-Wmacro-redefined]
#define memcpy memcpy
^
arch/arm/include/asm/string.h:58:9: note: previous definition is here
#define memcpy(dst, src, len) __memcpy(dst, src, len)
^
Here we want the set of functions from the decompressor, so undefine
the other macros before the override.
Link: https://lore.kernel.org/linux-arm-kernel/CACRpkdZYJogU_SN3H9oeVq=zJkRgRT1gDz3xp59gdqWXxw-B=w@mail.gmail.com/
Link: https://lore.kernel.org/lkml/202105091112.F5rmd4By-lkp@intel.com/
Fixes: d6d51a96c7 ("ARM: 9014/2: Replace string mem* functions for KASan")
Fixes: a7f464f3db ("ARM: 7001/2: Wire up support for the XZ decompressor")
Reported-by: kernel test robot <lkp@intel.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
ARM: kasan: Fix __get_user_check failure with kasan
In macro __get_user_check defined in arch/arm/include/asm/uaccess.h,
error code is store in register int __e(r0). When kasan is
enabled, assigning value to kernel address might trigger kasan check,
which unexpectedly overwrites r0 and causes undefined behavior on arm
kasan images.
One example is failure in do_futex and results in process soft lockup.
Log:
watchdog: BUG: soft lockup - CPU#0 stuck for 62946ms! [rs:main
Q:Reg:1151]
...
(__asan_store4) from (futex_wait_setup+0xf8/0x2b4)
(futex_wait_setup) from (futex_wait+0x138/0x394)
(futex_wait) from (do_futex+0x164/0xe40)
(do_futex) from (sys_futex_time32+0x178/0x230)
(sys_futex_time32) from (ret_fast_syscall+0x0/0x50)
The soft lockup happens in function futex_wait_setup. The reason is
function get_futex_value_locked always return EINVAL, thus pc jump
back to retry label and causes looping.
This line in function get_futex_value_locked
ret = __get_user(*dest, from);
is expanded to
*dest = (typeof(*(p))) __r2; ,
in macro __get_user_check. Writing to pointer dest triggers kasan check
and overwrites the return value of __get_user_x function.
The assembly code of get_futex_value_locked in kernel/futex.c:
...
c01f6dc8: eb0b020e bl c04b7608 <__get_user_4>
// "x = (typeof(*(p))) __r2;" triggers kasan check and r0 is overwritten
c01f6dCc: e1a00007 mov r0, r7
c01f6dd0: e1a05002 mov r5, r2
c01f6dd4: eb04f1e6 bl c0333574 <__asan_store4>
c01f6dd8: e5875000 str r5, [r7]
// save ret value of __get_user(*dest, from), which is dest address now
c01f6ddc: e1a05000 mov r5, r0
...
// checking return value of __get_user failed
c01f6e00: e3550000 cmp r5, #0
...
c01f6e0c: 01a00005 moveq r0, r5
// assign return value to EINVAL
c01f6e10: 13e0000d mvnne r0, #13
Return value is the destination address of get_user thus certainly
non-zero, so get_futex_value_locked always return EINVAL.
Fix it by using a tmp vairable to store the error code before the
assignment. This fix has no effects to non-kasan images thanks to compiler
optimization. It only affects cases that overwrite r0 due to kasan check.
This should fix bug discussed in Link:
[1] https://lore.kernel.org/linux-arm-kernel/0ef7c2a5-5d8b-c5e0-63fa-31693fd4495c@gmail.com/
Fixes: 421015713b ("ARM: 9017/2: Enable KASan for ARM")
Signed-off-by: Lexi Shao <shaolexi@huawei.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Commit 344179fc7e ("ARM: 9106/1: traps: use get_kernel_nofault instead
of set_fs()") replaced an occurrence of __get_user() with
get_kernel_nofault(), but inverted the sense of the conditional in the
process, resulting in no values to be printed at all.
I.e., every exception stack now looks like this:
Exception stack(0xc18d1fb0 to 0xc18d1ff8)
1fa0: ???????? ???????? ???????? ????????
1fc0: ???????? ???????? ???????? ???????? ???????? ???????? ???????? ????????
1fe0: ???????? ???????? ???????? ???????? ???????? ????????
which is rather unhelpful.
Fixes: 344179fc7e ("ARM: 9106/1: traps: use get_kernel_nofault instead of set_fs()")
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
tglx notes:
This function [futex_detect_cmpxchg] is only needed when an
architecture has to runtime discover whether the CPU supports it or
not. ARM has unconditional support for this, so the obvious thing to
do is the below.
Fixes linkage failure from Clang randconfigs:
kernel/futex.o:(.text.fixup+0x5c): relocation truncated to fit: R_ARM_JUMP24 against `.init.text'
and boot failures for CONFIG_THUMB2_KERNEL.
Link: https://github.com/ClangBuiltLinux/linux/issues/325
Comments from Nick Desaulniers:
See-also: 03b8c7b623 ("futex: Allow architectures to skip
futex_atomic_cmpxchg_inatomic() test")
Reported-by: Arnd Bergmann <arnd@arndb.de>
Reported-by: Nathan Chancellor <nathan@kernel.org>
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Cc: stable@vger.kernel.org # v3.14+
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Currently, we check the wb_err too early for directories, before all of
the unsafe child requests have been waited on. In order to fix that we
need to check the mapping->wb_err later nearer to the end of ceph_fsync.
We also have an overly-complex method for tracking errors after
blocklisting. The errors recorded in cleanup_session_requests go to a
completely separate field in the inode, but we end up reporting them the
same way we would for any other error (in fsync).
There's no real benefit to tracking these errors in two different
places, since the only reporting mechanism for them is in fsync, and
we'd need to advance them both every time.
Given that, we can just remove i_meta_err, and convert the places that
used it to instead just use mapping->wb_err instead. That also fixes
the original problem by ensuring that we do a check_and_advance of the
wb_err at the end of the fsync op.
Cc: stable@vger.kernel.org
URL: https://tracker.ceph.com/issues/52864
Reported-by: Patrick Donnelly <pdonnell@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Currently when mounting, we may end up finding an existing superblock
that corresponds to a blocklisted MDS client. This means that the new
mount ends up being unusable.
If we've found an existing superblock with a client that is already
blocklisted, and the client is not configured to recover on its own,
fail the match. Ditto if the superblock has been forcibly unmounted.
While we're in here, also rename "other" to the more conventional "fsc".
Cc: stable@vger.kernel.org
URL: https://bugzilla.redhat.com/show_bug.cgi?id=1901499
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
When the a large chunk of data send and the receiver does not send a
Flow Control frame back in time, the sendmsg() does not return a error
code, but the number of bytes sent corresponding to the size of the
packet.
If a timeout occurs the isotp_tx_timer_handler() is fired, sets
sk->sk_err and calls the sk->sk_error_report() function. It was
wrongly expected that the error would be propagated to user space in
every case. For isotp_sendmsg() blocking on wait_event_interruptible()
this is not the case.
This patch fixes the problem by checking if sk->sk_err is set and
returning the error to user space.
Fixes: e057dd3fc2 ("can: add ISO 15765-2:2016 transport protocol")
Link: https://github.com/hartkopp/can-isotp/issues/42
Link: https://github.com/hartkopp/can-isotp/pull/43
Link: https://lore.kernel.org/all/20210507091839.1366379-1-mkl@pengutronix.de
Cc: stable@vger.kernel.org
Reported-by: Sottas Guillaume (LMB) <Guillaume.Sottas@liebherr.com>
Tested-by: Oliver Hartkopp <socketcan@hartkopp.net>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Starting with kernel 5.11 built with CONFIG_FORTIFY_SOURCE mouting an
ocfs2 filesystem with either o2cb or pcmk cluster stack fails with the
trace below. Problem seems to be that strings for cluster stack and
cluster name are not guaranteed to be null terminated in the disk
representation, while strlcpy assumes that the source string is always
null terminated. This causes a read outside of the source string
triggering the buffer overflow detection.
detected buffer overflow in strlen
------------[ cut here ]------------
kernel BUG at lib/string.c:1149!
invalid opcode: 0000 [#1] SMP PTI
CPU: 1 PID: 910 Comm: mount.ocfs2 Not tainted 5.14.0-1-amd64 #1
Debian 5.14.6-2
RIP: 0010:fortify_panic+0xf/0x11
...
Call Trace:
ocfs2_initialize_super.isra.0.cold+0xc/0x18 [ocfs2]
ocfs2_fill_super+0x359/0x19b0 [ocfs2]
mount_bdev+0x185/0x1b0
legacy_get_tree+0x27/0x40
vfs_get_tree+0x25/0xb0
path_mount+0x454/0xa20
__x64_sys_mount+0x103/0x140
do_syscall_64+0x3b/0xc0
entry_SYSCALL_64_after_hwframe+0x44/0xae
Link: https://lkml.kernel.org/r/20210929180654.32460-1-vvidic@valentin-vidic.from.hr
Signed-off-by: Valentin Vidic <vvidic@valentin-vidic.from.hr>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Gang He <ghe@suse.com>
Cc: Jun Piao <piaojun@huawei.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit 6dbf7bb555 ("fs: Don't invalidate page buffers in
block_write_full_page()") uncovered a latent bug in ocfs2 conversion
from inline inode format to a normal inode format.
The code in ocfs2_convert_inline_data_to_extents() attempts to zero out
the whole cluster allocated for file data by grabbing, zeroing, and
dirtying all pages covering this cluster. However these pages are
beyond i_size, thus writeback code generally ignores these dirty pages
and no blocks were ever actually zeroed on the disk.
This oversight was fixed by commit 693c241a5f ("ocfs2: No need to zero
pages past i_size.") for standard ocfs2 write path, inline conversion
path was apparently forgotten; the commit log also has a reasoning why
the zeroing actually is not needed.
After commit 6dbf7bb555, things became worse as writeback code stopped
invalidating buffers on pages beyond i_size and thus these pages end up
with clean PageDirty bit but with buffers attached to these pages being
still dirty. So when a file is converted from inline format, then
writeback triggers, and then the file is grown so that these pages
become valid, the invalid dirtiness state is preserved,
mark_buffer_dirty() does nothing on these pages (buffers are already
dirty) but page is never written back because it is clean. So data
written to these pages is lost once pages are reclaimed.
Simple reproducer for the problem is:
xfs_io -f -c "pwrite 0 2000" -c "pwrite 2000 2000" -c "fsync" \
-c "pwrite 4000 2000" ocfs2_file
After unmounting and mounting the fs again, you can observe that end of
'ocfs2_file' has lost its contents.
Fix the problem by not doing the pointless zeroing during conversion
from inline format similarly as in the standard write path.
[akpm@linux-foundation.org: fix whitespace, per Joseph]
Link: https://lkml.kernel.org/r/20210930095405.21433-1-jack@suse.cz
Fixes: 6dbf7bb555 ("fs: Don't invalidate page buffers in block_write_full_page()")
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Tested-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Acked-by: Gang He <ghe@suse.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Jun Piao <piaojun@huawei.com>
Cc: "Markov, Andrey" <Markov.Andrey@Dell.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The node demotion order needs to be updated during CPU hotplug. Because
whether a NUMA node has CPU may influence the demotion order. The
update function should be called during CPU online/offline after the
node_states[N_CPU] has been updated. That is done in
CPUHP_AP_ONLINE_DYN during CPU online and in CPUHP_MM_VMSTAT_DEAD during
CPU offline. But in commit 884a6e5d1f ("mm/migrate: update node
demotion order on hotplug events"), the function to update node demotion
order is called in CPUHP_AP_ONLINE_DYN during CPU online/offline. This
doesn't satisfy the order requirement.
For example, there are 4 CPUs (P0, P1, P2, P3) in 2 sockets (P0, P1 in S0
and P2, P3 in S1), the demotion order is
- S0 -> NUMA_NO_NODE
- S1 -> NUMA_NO_NODE
After P2 and P3 is offlined, because S1 has no CPU now, the demotion
order should have been changed to
- S0 -> S1
- S1 -> NO_NODE
but it isn't changed, because the order updating callback for CPU
hotplug doesn't see the new nodemask. After that, if P1 is offlined,
the demotion order is changed to the expected order as above.
So in this patch, we added CPUHP_AP_MM_DEMOTION_ONLINE and
CPUHP_MM_DEMOTION_DEAD to be called after CPUHP_AP_ONLINE_DYN and
CPUHP_MM_VMSTAT_DEAD during CPU online and offline, and register the
update function on them.
Link: https://lkml.kernel.org/r/20210929060351.7293-1-ying.huang@intel.com
Fixes: 884a6e5d1f ("mm/migrate: update node demotion order on hotplug events")
Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Yang Shi <shy828301@gmail.com>
Cc: Zi Yan <ziy@nvidia.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Wei Xu <weixugc@google.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: David Rientjes <rientjes@google.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Keith Busch <kbusch@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "mm/migrate: 5.15 fixes for automatic demotion", v2.
This contains two fixes for the "automatic demotion" code which was
merged into 5.15:
* Fix memory hotplug performance regression by watching
suppressing any real action on irrelevant hotplug events.
* Ensure CPU hotplug handler is registered when memory hotplug
is disabled.
This patch (of 2):
== tl;dr ==
Automatic demotion opted for a simple, lazy approach to handling hotplug
events. This noticeably slows down memory hotplug[1]. Optimize away
updates to the demotion order when memory hotplug events should have no
effect.
This has no effect on CPU hotplug. There is no known problem on the CPU
side and any work there will be in a separate series.
== Background ==
Automatic demotion is a memory migration strategy to ensure that new
allocations have room in faster memory tiers on tiered memory systems.
The kernel maintains an array (node_demotion[]) to drive these
migrations.
The node_demotion[] path is calculated by starting at nodes with CPUs
and then "walking" to nodes with memory. Only hotplug events which
online or offline a node with memory (N_ONLINE) or CPUs (N_CPU) will
actually affect the migration order.
== Problem ==
However, the current code is lazy. It completely regenerates the
migration order on *any* CPU or memory hotplug event. The logic was
that these events are extremely rare and that the overhead from
indiscriminate order regeneration is minimal.
Part of the update logic involves a synchronize_rcu(), which is a pretty
big hammer. Its overhead was large enough to be detected by some 0day
tests that watch memory hotplug performance[1].
== Solution ==
Add a new helper (node_demotion_topo_changed()) which can differentiate
between superfluous and impactful hotplug events. Skip the expensive
update operation for superfluous events.
== Aside: Locking ==
It took me a few moments to declare the locking to be safe enough for
node_demotion_topo_changed() to work. It all hinges on the memory
hotplug lock:
During memory hotplug events, 'mem_hotplug_lock' is held for write.
This ensures that two memory hotplug events can not be called
simultaneously.
CPU hotplug has a similar lock (cpuhp_state_mutex) which also provides
mutual exclusion between CPU hotplug events. In addition, the demotion
code acquire and hold the mem_hotplug_lock for read during its CPU
hotplug handlers. This provides mutual exclusion between the demotion
memory hotplug callbacks and the CPU hotplug callbacks.
This effectively allows treating the migration target generation code to
act as if it is single-threaded.
1. https://lore.kernel.org/all/20210905135932.GE15026@xsang-OptiPlex-9020/
Link: https://lkml.kernel.org/r/20210924161251.093CCD06@davehans-spike.ostc.intel.com
Link: https://lkml.kernel.org/r/20210924161253.D7673E31@davehans-spike.ostc.intel.com
Fixes: 884a6e5d1f ("mm/migrate: update node demotion order on hotplug events")
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reported-by: kernel test robot <oliver.sang@intel.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: "Huang, Ying" <ying.huang@intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Wei Xu <weixugc@google.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: David Rientjes <rientjes@google.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Yang Shi <yang.shi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In RHEL's gating selftests we've encountered memory corruption in the
uffd event test even with upstream kernel:
# ./userfaultfd anon 128 4
nr_pages: 32768, nr_pages_per_cpu: 32768
bounces: 3, mode: rnd racing read, userfaults: 6240 missing (6240) 14729 wp (14729)
bounces: 2, mode: racing read, userfaults: 1444 missing (1444) 28877 wp (28877)
bounces: 1, mode: rnd read, userfaults: 6055 missing (6055) 14699 wp (14699)
bounces: 0, mode: read, userfaults: 82 missing (82) 25196 wp (25196)
testing uffd-wp with pagemap (pgsize=4096): done
testing uffd-wp with pagemap (pgsize=2097152): done
testing events (fork, remap, remove): ERROR: nr 32427 memory corruption 0 1 (errno=0, line=963)
ERROR: faulting process failed (errno=0, line=1117)
It can be easily reproduced when global thp enabled, which is the
default for RHEL.
It's also known as a side effect of commit 0db282ba2c ("selftest: use
mmap instead of posix_memalign to allocate memory", 2021-07-23), which
is imho right itself on using mmap() to make sure the addresses will be
untagged even on arm.
The problem is, for each test we allocate buffers using two
allocate_area() calls. We assumed these two buffers won't affect each
other, however they could, because mmap() could have found that the two
buffers are near each other and having the same VMA flags, so they got
merged into one VMA.
It won't be a big problem if thp is not enabled, but when thp is
agressively enabled it means when initializing the src buffer it could
accidentally setup part of the dest buffer too when there's a shared THP
that overlaps the two regions. Then some of the dest buffer won't be
able to be trapped by userfaultfd missing mode, then it'll cause memory
corruption as described.
To fix it, do release_pages() after initializing the src buffer.
Since the previous two release_pages() calls are after
uffd_test_ctx_clear() which will unmap all the buffers anyway (which is
stronger than release pages; as unmap() also tear town pgtables), drop
them as they shouldn't really be anything useful.
We can mark the Fixes tag upon 0db282ba2c as it's reported to only
happen there, however the real "Fixes" IMHO should be 8ba6e86408, as
before that commit we'll always do explicit release_pages() before
registration of uffd, and 8ba6e86408 changed that logic by adding
extra unmap/map and we didn't release the pages at the right place.
Meanwhile I don't have a solid glue anyway on whether posix_memalign()
could always avoid triggering this bug, hence it's safer to attach this
fix to commit 8ba6e86408.
Link: https://lkml.kernel.org/r/20210923232512.210092-1-peterx@redhat.com
Fixes: 8ba6e86408 ("userfaultfd/selftests: reinitialize test context in each test")
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1994931
Signed-off-by: Peter Xu <peterx@redhat.com>
Reported-by: Li Wang <liwan@redhat.com>
Tested-by: Li Wang <liwang@redhat.com>
Reviewed-by: Axel Rasmussen <axelrasmussen@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Nadav Amit <nadav.amit@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When a Jieli Technology USB Webcam is connected, the video part works
well, but the mic sound is speeded up. On dmesg there are messages
about different rates from the runtime rates, warnings about volume
resolution and lastly, the log is filled, every 5 seconds, with
retire_capture_urb error messages.
The mic works only when ep packet size is set to wMaxPacketSize (normal
sound and no more retire_capture_urb error messages). Skipping reading
sample rate, fixes the messages about different rates and forcing a volume
resolution, fixes warnings about volume range. I have arbitrarily choosed
the value (16): I read in a comment that there should be no more than 255
levels, so 4096 (max volume) / 16 = 0-255.
Signed-off-by: Marco Giunta <giun7a@gmail.com>
Link: https://lore.kernel.org/r/20211018162552.12082-1-giun7a@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Implement the ->restore() PM operation and set the link to off, which will
force a full reset and restore. This ensures that Host Performance Booster
is reset after suspend-to-disk.
The Host Performance Booster feature caches logical-to-physical mapping
information in the host memory. After suspend-to-disk, such information is
not valid, so a full reset and restore is needed.
A full reset and restore is done if the SPM level is 5 or 6, but not for
other SPM levels, so this change fixes those cases.
A full reset and restore also restores base address registers, so that code
is removed.
Link: https://lore.kernel.org/r/20211018151004.284200-2-adrian.hunter@intel.com
Reviewed-by: Avri Altman <avri.altman@wdc.com>
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The sgl is freed in the target stack in target_release_cmd_kref() before
calling qlt_free_cmd() but there is an unmap of sgl in qlt_free_cmd() that
causes a panic if sgl is not yet DMA unmapped:
NIP dma_direct_unmap_sg+0xdc/0x180
LR dma_direct_unmap_sg+0xc8/0x180
Call Trace:
ql_dbg_prefix+0x68/0xc0 [qla2xxx] (unreliable)
dma_unmap_sg_attrs+0x54/0xf0
qlt_unmap_sg.part.19+0x54/0x1c0 [qla2xxx]
qlt_free_cmd+0x124/0x1d0 [qla2xxx]
tcm_qla2xxx_release_cmd+0x4c/0xa0 [tcm_qla2xxx]
target_put_sess_cmd+0x198/0x370 [target_core_mod]
transport_generic_free_cmd+0x6c/0x1b0 [target_core_mod]
tcm_qla2xxx_complete_free+0x6c/0x90 [tcm_qla2xxx]
The sgl may be left unmapped in error cases of response sending. For
instance, qlt_rdy_to_xfer() maps sgl and exits when session is being
deleted keeping the sgl mapped.
This patch removes use-after-free of the sgl and ensures that the sgl is
unmapped for any command that was not sent to firmware.
Link: https://lore.kernel.org/r/20211018122650.11846-1-d.bogdanov@yadro.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Dmitry Bogdanov <d.bogdanov@yadro.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Commit 8c0eb596ba ("[SCSI] qla2xxx: Fix a memory leak in an error path of
qla2x00_process_els()"), intended to change:
bsg_job->request->msgcode == FC_BSG_HST_ELS_NOLOGIN
to:
bsg_job->request->msgcode != FC_BSG_RPT_ELS
but changed it to:
bsg_job->request->msgcode == FC_BSG_RPT_ELS
instead.
Change the == to a != to avoid leaking the fcport structure or freeing
unallocated memory.
Link: https://lore.kernel.org/r/20211012191834.90306-2-jgu@purestorage.com
Fixes: 8c0eb596ba ("[SCSI] qla2xxx: Fix a memory leak in an error path of qla2x00_process_els()")
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Joy Gu <jgu@purestorage.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
While writing an email explaining the "bit = 0" logic for a discussion on
making ftrace_test_recursion_trylock() disable preemption, I discovered a
path that makes the "not do the logic if bit is zero" unsafe.
The recursion logic is done in hot paths like the function tracer. Thus,
any code executed causes noticeable overhead. Thus, tricks are done to try
to limit the amount of code executed. This included the recursion testing
logic.
Having recursion testing is important, as there are many paths that can
end up in an infinite recursion cycle when tracing every function in the
kernel. Thus protection is needed to prevent that from happening.
Because it is OK to recurse due to different running context levels (e.g.
an interrupt preempts a trace, and then a trace occurs in the interrupt
handler), a set of bits are used to know which context one is in (normal,
softirq, irq and NMI). If a recursion occurs in the same level, it is
prevented*.
Then there are infrastructure levels of recursion as well. When more than
one callback is attached to the same function to trace, it calls a loop
function to iterate over all the callbacks. Both the callbacks and the
loop function have recursion protection. The callbacks use the
"ftrace_test_recursion_trylock()" which has a "function" set of context
bits to test, and the loop function calls the internal
trace_test_and_set_recursion() directly, with an "internal" set of bits.
If an architecture does not implement all the features supported by ftrace
then the callbacks are never called directly, and the loop function is
called instead, which will implement the features of ftrace.
Since both the loop function and the callbacks do recursion protection, it
was seemed unnecessary to do it in both locations. Thus, a trick was made
to have the internal set of recursion bits at a more significant bit
location than the function bits. Then, if any of the higher bits were set,
the logic of the function bits could be skipped, as any new recursion
would first have to go through the loop function.
This is true for architectures that do not support all the ftrace
features, because all functions being traced must first go through the
loop function before going to the callbacks. But this is not true for
architectures that support all the ftrace features. That's because the
loop function could be called due to two callbacks attached to the same
function, but then a recursion function inside the callback could be
called that does not share any other callback, and it will be called
directly.
i.e.
traced_function_1: [ more than one callback tracing it ]
call loop_func
loop_func:
trace_recursion set internal bit
call callback
callback:
trace_recursion [ skipped because internal bit is set, return 0 ]
call traced_function_2
traced_function_2: [ only traced by above callback ]
call callback
callback:
trace_recursion [ skipped because internal bit is set, return 0 ]
call traced_function_2
[ wash, rinse, repeat, BOOM! out of shampoo! ]
Thus, the "bit == 0 skip" trick is not safe, unless the loop function is
call for all functions.
Since we want to encourage architectures to implement all ftrace features,
having them slow down due to this extra logic may encourage the
maintainers to update to the latest ftrace features. And because this
logic is only safe for them, remove it completely.
[*] There is on layer of recursion that is allowed, and that is to allow
for the transition between interrupt context (normal -> softirq ->
irq -> NMI), because a trace may occur before the context update is
visible to the trace recursion logic.
Link: https://lore.kernel.org/all/609b565a-ed6e-a1da-f025-166691b5d994@linux.alibaba.com/
Link: https://lkml.kernel.org/r/20211018154412.09fcad3c@gandalf.local.home
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Petr Mladek <pmladek@suse.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "James E.J. Bottomley" <James.Bottomley@hansenpartnership.com>
Cc: Helge Deller <deller@gmx.de>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Albert Ou <aou@eecs.berkeley.edu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Borislav Petkov <bp@alien8.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Jiri Kosina <jikos@kernel.org>
Cc: Miroslav Benes <mbenes@suse.cz>
Cc: Joe Lawrence <joe.lawrence@redhat.com>
Cc: Colin Ian King <colin.king@canonical.com>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Jisheng Zhang <jszhang@kernel.org>
Cc: =?utf-8?b?546L6LSH?= <yun.wang@linux.alibaba.com>
Cc: Guo Ren <guoren@kernel.org>
Cc: stable@vger.kernel.org
Fixes: edc15cafcb ("tracing: Avoid unnecessary multiple recursion checks")
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
A new warning in clang points out two places in this driver where
boolean expressions are being used with a bitwise OR instead of a
logical one:
drivers/net/ethernet/netronome/nfp/nfp_asm.c:199:20: error: use of bitwise '|' with boolean operands [-Werror,-Wbitwise-instead-of-logical]
reg->src_lmextn = swreg_lmextn(lreg) | swreg_lmextn(rreg);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
||
drivers/net/ethernet/netronome/nfp/nfp_asm.c:199:20: note: cast one or both operands to int to silence this warning
drivers/net/ethernet/netronome/nfp/nfp_asm.c:280:20: error: use of bitwise '|' with boolean operands [-Werror,-Wbitwise-instead-of-logical]
reg->src_lmextn = swreg_lmextn(lreg) | swreg_lmextn(rreg);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
||
drivers/net/ethernet/netronome/nfp/nfp_asm.c:280:20: note: cast one or both operands to int to silence this warning
2 errors generated.
The motivation for the warning is that logical operations short circuit
while bitwise operations do not. In this case, it does not seem like
short circuiting is harmful so implement the suggested fix of changing
to a logical operation to fix the warning.
Link: https://github.com/ClangBuiltLinux/linux/issues/1479
Reported-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Link: https://lore.kernel.org/r/20211018193101.2340261-1-nathan@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
