The list iterator 'pmem' will point to a bogus position containing
HEAD if the list is empty or no element is found. This case must
be checked before any use of the iterator, otherwise it will
lead to a invalid memory access.
To fix this bug, just gen_pool_free/set NULL/list_del() and return
when found, otherwise list_del HEAD and return;
The list iterator value 'buf' will *always* be set and non-NULL
by list_for_each_entry(), so it is incorrect to assume that the
iterator value will be NULL if the list is empty (in this case, the
check 'if (!buf) {' will always be false and never exit expectly).
To fix the bug, use a new variable 'iter' as the list iterator,
while use the original variable 'buf' as a dedicated pointer to
point to the found element.
The help message of 'get_abi.pl' is mistakenly saying it's
'abi_book.pl'. This commit fixes the wrong name in the help message.
Fixes: bbc249f2b859 ("scripts: add an script to parse the ABI files") Signed-off-by: SeongJae Park <sj@kernel.org> Link: https://lore.kernel.org/r/20220419121636.290407-1-sj@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
If the device is already in a runtime PM enabled state
pm_runtime_get_sync() will return 1, so a test for negative
value should be used to check for errors.
The list_for_each_entry_safe() macro saves the current item (n) and
the item after (n+1), so that n can be safely removed without
corrupting the list. However, when traversing the list and removing
items using gadget giveback, the DWC3 lock is briefly released,
allowing other routines to execute. There is a situation where, while
items are being removed from the cancelled_list using
dwc3_gadget_ep_cleanup_cancelled_requests(), the pullup disable
routine is running in parallel (due to UDC unbind). As the cleanup
routine removes n, and the pullup disable removes n+1, once the
cleanup retakes the DWC3 lock, it references a request who was already
removed/handled. With list debug enabled, this leads to a panic.
Ensure all instances of the macro are replaced where gadget giveback
is used.
Example call stack:
Thread#1:
__dwc3_gadget_ep_set_halt() - CLEAR HALT
-> dwc3_gadget_ep_cleanup_cancelled_requests()
->list_for_each_entry_safe()
->dwc3_gadget_giveback(n)
->dwc3_gadget_del_and_unmap_request()- n deleted[cancelled_list]
->spin_unlock
->Thread#2 executes
...
->dwc3_gadget_giveback(n+1)
->Already removed!
Thread#2:
dwc3_gadget_pullup()
->waiting for dwc3 spin_lock
...
->Thread#1 released lock
->dwc3_stop_active_transfers()
->dwc3_remove_requests()
->fetches n+1 item from cancelled_list (n removed by Thread#1)
->dwc3_gadget_giveback()
->dwc3_gadget_del_and_unmap_request()- n+1 deleted[cancelled_list]
->spin_unlock
The reg member of struct raspberrypi_pwm_prop is a little endian 32 bit
quantity. Explicitly convert the (native endian) value to little endian
on assignment as is already done in raspberrypi_pwm_set_property().
This fixes the following sparse warning:
drivers/pwm/pwm-raspberrypi-poe.c:69:24: warning: incorrect type in initializer (different base types)
drivers/pwm/pwm-raspberrypi-poe.c:69:24: expected restricted __le32 [usertype] reg
drivers/pwm/pwm-raspberrypi-poe.c:69:24: got unsigned int [usertype] reg
The hardware only supports periods <= 1.6 ms and if a bigger period is
requested it is clamped to 1.6 ms. In this case duty_cycle might be bigger
than 1.6 ms and then the duty cycle register is written with a value
bigger than LP3943_MAX_DUTY. So clamp duty_cycle accordingly.
If device_register() fails, device_unregister() should not be called
because it will free some resources that are not allocated.
put_device() should be used instead.
The function documentation of usb_set_configuration says that its
callers should hold the device lock. This lock is held for all
callsites except tweak_set_configuration_cmd. The code path can be
executed for example when attaching a remote USB device.
The solution is to surround the call by the device lock.
This bug was found using my experimental own-developed static analysis
tool, which reported the missing lock on v5.17.2. I manually verified
this bug report by doing code review as well. I runtime checked that
the required lock is not held. I compiled and runtime tested this on
x86_64 with a USB mouse. After applying this patch, my analyser no
longer reports this potential bug.
This commit fixes two issues with the muxed interrupt handler. First,
the OTG port has the "bvalid" interrupt enabled, not "linestate". Since
only the linestate interrupt was handled, and not the bvalid interrupt,
plugging in a cable to the OTG port caused an interrupt storm.
Second, the return values from the individual port IRQ handlers need to
be OR-ed together. Otherwise, the lack of an interrupt from the last
port would cause the handler to erroneously return IRQ_NONE.
Fixes: ed2b5a8e6b98 ("phy: phy-rockchip-inno-usb2: support muxed interrupts") Signed-off-by: Samuel Holland <samuel@sholland.org> Tested-by: Michael Riesch <michael.riesch@wolfvision.net> Link: https://lore.kernel.org/r/20220414032258.40984-2-samuel@sholland.org Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Resource table is used by Linux to get information published by
remote processor. It should be not be used for memory allocation, so
not create rproc mem entry.
Now fsl_lpuart driver use both of_alias_get_id() and ida_simple_get() in
.probe(), which has the potential bug. For example, when remove the
lpuart7 alias in dts, of_alias_get_id() will return error, then call
ida_simple_get() to allocate the id 0 for lpuart7, this may confilct
with the lpuart4 which has alias 0.
So remove the ida_simple_get() in .probe(), return an error directly
when calling of_alias_get_id() fails, which is consistent with other
uart drivers behavior.
Fixes: 3bc3206e1c0f ("serial: fsl_lpuart: Remove the alias node dependence") Signed-off-by: Sherry Sun <sherry.sun@nxp.com> Link: https://lore.kernel.org/r/20220321112211.8895-1-sherry.sun@nxp.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
platform_get_resource() may fail and return NULL, so we should
better check it's return value to avoid a NULL pointer dereference.
Fixes: 54da3e381c2b ("serial: 8250_aspeed_vuart: use UPF_IOREMAP to set up register mapping") Signed-off-by: Miaoqian Lin <linmq006@gmail.com> Link: https://lore.kernel.org/r/20220404143842.16960-1-linmq006@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
TTYs in ICANON mode have a special case that allows "pushing" a line
without a regular EOL character (like newline), by using EOF (the EOT
character - ASCII 0x4) as a pseudo-EOL. It is silently discarded, so
the reader of the PTS will receive the line *without* EOF or any other
terminating character.
This special case has an edge case: What happens if the readers buffer
is the same size as the line (without EOF)? Will they be able to tell
if the whole line is received, i.e. if the next read() will return more
of the same line or the next line?
There are two possibilities, that both have (dis)advantages:
1. The next read() returns 0. FreeBSD (13.0) and OSX (10.11) do this.
Advantage: The reader can interpret this as "the line is over".
Disadvantage: read() returning 0 means EOF, the reader could also
interpret it as "there's no more data" and stop reading or even
close the PT.
2. The next read() returns the next line, the EOF is silently discarded.
Solaris (or at least OpenIndiana 2021.10) does this, Linux has done
do this since commit 40d5e0905a03 ("n_tty: Fix EOF push handling");
this behavior was recently broken by commit 359303076163 ("tty:
n_tty: do not look ahead for EOL character past the end of the buffer").
Advantage: read() won't return 0 (EOF), reader less likely to be
confused (and things like `while(read(..)>0)` don't break)
Disadvantage: The reader can't really know if the read() continues
the last line (that filled the whole read buffer) or starts a
new line.
As both options are defensible (and are used by other Unix-likes), it's
best to stick to the "old" behavior since "n_tty: Fix EOF push handling"
of 2013, i.e. silently discard that EOF.
This patch - that I actually got from Linus for testing and only
modified slightly - restores that behavior by skipping an EOF
character if it's the next character after reading is done.
Based on a patch from Linus Torvalds.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=215611 Fixes: 359303076163 ("tty: n_tty: do not look ahead for EOL character past the end of the buffer") Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Jiri Slaby <jirislaby@kernel.org> Reviewed-and-tested-by: Daniel Gibson <daniel@gibson.sh> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Daniel Gibson <daniel@gibson.sh> Link: https://lore.kernel.org/r/20220329235810.452513-2-daniel@gibson.sh Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In goldfish_tty_probe(), the port initialized through tty_port_init()
should be destroyed in error paths.In goldfish_tty_remove(), qtty->port
also should be destroyed or else might leak resources.
# echo ARRAY_BOUNDS > /sys/kernel/debug/provoke-crash/DIRECT
[ 102.265827] ================================================================================
[ 102.278433] UBSAN: array-index-out-of-bounds in drivers/misc/lkdtm/bugs.c:342:16
[ 102.287207] index 8 is out of range for type 'char [8]'
[ 102.298722] ================================================================================
[ 102.313712] lkdtm: FAIL: survived array bounds overflow!
[ 102.318770] lkdtm: Unexpected! This kernel (5.16.0-rc1-s3k-dev-01884-g720dcf79314a ppc) was built with CONFIG_UBSAN_BOUNDS=y
It is not correct because when CONFIG_UBSAN_TRAP is not selected
you can't expect array bounds overflow to kill the thread.
Modify the logic so that when the kernel is built with
CONFIG_UBSAN_BOUNDS but without CONFIG_UBSAN_TRAP, you get a warning
about CONFIG_UBSAN_TRAP not been selected instead.
This also require a fix of pr_expected_config(), otherwise the
following error is encountered.
CC drivers/misc/lkdtm/bugs.o
drivers/misc/lkdtm/bugs.c: In function 'lkdtm_ARRAY_BOUNDS':
drivers/misc/lkdtm/bugs.c:351:2: error: 'else' without a previous 'if'
351 | else
| ^~~~
As the possible failure of the kmalloc(), the not_checked and checked
could be NULL pointer.
Therefore, it should be better to check it in order to avoid the
dereference of the NULL pointer.
Also, we need to kfree the 'not_checked' and 'checked' to avoid
the memory leak if fails.
And since it is just a test, it may directly return without error
number.
If the list does not exit early then data == NULL and 'module' does not
point to a valid list element.
Using 'module' in such a case is not valid and was therefore removed.
Fixes: 6dd67645f22c ("greybus: audio: Use single codec driver registration") Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by: Vaibhav Agarwal <vaibhav.sr@gmail.com> Reviewed-by: Mark Greer <mgreer@animalcreek.com> Signed-off-by: Jakob Koschel <jakobkoschel@gmail.com> Link: https://lore.kernel.org/r/20220321123626.3068639-1-jakobkoschel@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When the MIPS_ALCHEMY board selection is MIPS_XXS1500 instead of
MIPS_DB1XXX, the PCMCIA driver 'db1xxx_ss' has build errors due
to missing DB1XXX symbols. The PCMCIA driver should be restricted
to MIPS_DB1XXX instead of MIPS_ALCHEMY to fix this build error.
Commit d92c370a16cb ("block: really clone the block cgroup in
bio_clone_blkg_association") changed bio_clone_blkg_association() to
just clone bio->bi_blkg reference from source to destination bio. This
is however wrong if the source and destination bios are against
different block devices because struct blkcg_gq is different for each
bdev-blkcg pair. This will result in IOs being accounted (and throttled
as a result) multiple times against the same device (src bdev) while
throttling of the other device (dst bdev) is ignored. In case of BFQ the
inconsistency can even result in crashes in bfq_bic_update_cgroup().
Fix the problem by looking up correct blkcg_gq for the cloned bio.
Reported-by: Logan Gunthorpe <logang@deltatee.com> Reported-and-tested-by: Donald Buczek <buczek@molgen.mpg.de> Fixes: d92c370a16cb ("block: really clone the block cgroup in bio_clone_blkg_association") CC: stable@vger.kernel.org Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20220602081242.7731-1-jack@suse.cz Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The module param debug for n_gsm uses KERN_INFO level, but the hexdump
now uses KERN_DEBUG level. This started after commit 091cb0994edd
("lib/hexdump: make print_hex_dump_bytes() a nop on !DEBUG builds").
We now use dynamic_hex_dump() unless DEBUG is set.
This causes no packets to be seen with modprobe n_gsm debug=0x1f unlike
earlier. Let's fix this by adding gsm_hex_dump_bytes() that calls
print_hex_dump() with KERN_INFO to match what n_gsm is doing with the
other debug related output.
Fixes: 091cb0994edd ("lib/hexdump: make print_hex_dump_bytes() a nop on !DEBUG builds") Cc: Stephen Boyd <swboyd@chromium.org> Signed-off-by: Tony Lindgren <tony@atomide.com> Link: https://lore.kernel.org/r/20220512131506.1216-1-tony@atomide.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Now io_acct_set is alloc and free in personality. Remove the codes that
free io_acct_set in md_free and md_stop.
Fixes: 0c031fd37f69 (md: Move alloc/free acct bioset in to personality) Signed-off-by: Xiao Ni <xni@redhat.com> Signed-off-by: Song Liu <song@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In normal stop process, it does like this:
do_md_stop
|
__md_stop (pers->free(); mddev->private=NULL)
|
md_free (free mddev)
__md_stop sets mddev->private to NULL after pers->free. The raid device
will be stopped and mddev memory is free. But in reshape, it doesn't
free the mddev and mddev will still be used in new raid.
In reshape, it first sets mddev->private to new_pers and then runs
old_pers->free(). Now raid0 sets mddev->private to NULL in raid0_free.
The new raid can't work anymore. It will panic when dereference
mddev->private because of NULL pointer dereference.
Removing the code that sets mddev->private to NULL in raid0 can fix
problem.
Fixes: 0c031fd37f69 (md: Move alloc/free acct bioset in to personality) Reported-by: Fine Fan <ffan@redhat.com> Signed-off-by: Xiao Ni <xni@redhat.com> Signed-off-by: Song Liu <song@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
log_read_rst() returns ENOMEM error when there is not enough memory.
In this case, if info is returned without initialization,
it attempts to kfree the uninitialized info->r_page pointer. This patch
moves the memset initialization code to before log_read_rst() is called.
Reported-by: Gerald Lee <sundaywind2004@gmail.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org> Signed-off-by: Konstantin Komarov <almaz.alexandrovich@paragon-software.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Make the two locations where exportfs helpers check permission to lookup
a given inode idmapped mount aware by switching it to the lookup_one()
helper. This is a bugfix for the open_by_handle_at() system call which
doesn't take idmapped mounts into account currently. It's not tied to a
specific commit so we'll just Cc stable.
In addition this is required to support idmapped base layers in overlay.
The overlay filesystem uses exportfs to encode and decode file handles
for its index=on mount option and when nfs_export=on.
Cc: <stable@vger.kernel.org> Cc: <linux-fsdevel@vger.kernel.org> Tested-by: Giuseppe Scrivano <gscrivan@redhat.com> Reviewed-by: Amir Goldstein <amir73il@gmail.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Christian Brauner (Microsoft) <brauner@kernel.org> Signed-off-by: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Similar to the addition of lookup_one() add a version of
lookup_one_unlocked() and lookup_one_positive_unlocked() that take
idmapped mounts into account. This is required to port overlay to
support idmapped base layers.
Cc: <linux-fsdevel@vger.kernel.org> Tested-by: Giuseppe Scrivano <gscrivan@redhat.com> Reviewed-by: Amir Goldstein <amir73il@gmail.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Christian Brauner (Microsoft) <brauner@kernel.org> Signed-off-by: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Make the test_dummy_encryption mount option require that the encrypt
feature flag be already enabled on the filesystem, rather than
automatically enabling it. Practically, this means that "-O encrypt"
will need to be included in MKFS_OPTIONS when running xfstests with the
test_dummy_encryption mount option. (ext4/053 also needs an update.)
Moreover, as long as the preconditions for test_dummy_encryption are
being tightened anyway, take the opportunity to start rejecting it when
!CONFIG_FS_ENCRYPTION rather than ignoring it.
The motivation for requiring the encrypt feature flag is that:
- Having the filesystem auto-enable feature flags is problematic, as it
bypasses the usual sanity checks. The specific issue which came up
recently is that in kernel versions where ext4 supports casefold but
not encrypt+casefold (v5.1 through v5.10), the kernel will happily add
the encrypt flag to a filesystem that has the casefold flag, making it
unmountable -- but only for subsequent mounts, not the initial one.
This confused the casefold support detection in xfstests, causing
generic/556 to fail rather than be skipped.
- The xfstests-bld test runners (kvm-xfstests et al.) already use the
required mkfs flag, so they will not be affected by this change. Only
users of test_dummy_encryption alone will be affected. But, this
option has always been for testing only, so it should be fine to
require that the few users of this option update their test scripts.
- f2fs already requires it (for its equivalent feature flag).
Signed-off-by: Eric Biggers <ebiggers@google.com> Reviewed-by: Gabriel Krisman Bertazi <krisman@collabora.com> Link: https://lore.kernel.org/r/20220519204437.61645-1-ebiggers@kernel.org Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove unsupported forcing of `cpu_has_fpu' to 1, which makes the `nofpu'
kernel parameter non-functional, and also causes a link error:
ld: arch/mips/kernel/traps.o: in function `trap_init':
./arch/mips/include/asm/msa.h:(.init.text+0x348): undefined reference to `handle_fpe'
ld: ./arch/mips/include/asm/msa.h:(.init.text+0x354): undefined reference to `handle_fpe'
ld: ./arch/mips/include/asm/msa.h:(.init.text+0x360): undefined reference to `handle_fpe'
where the CONFIG_MIPS_FP_SUPPORT configuration option has been disabled.
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk> Reported-by: Stephen Zhang <starzhangzsd@gmail.com> Fixes: 7505576d1c1a ("MIPS: add support for SGI Octane (IP30)") Cc: stable@vger.kernel.org # v5.5+ Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove unsupported forcing of `cpu_has_fpu' to 1, which makes the `nofpu'
kernel parameter non-functional, and also causes a link error:
ld: arch/mips/kernel/traps.o: in function `trap_init':
./arch/mips/include/asm/msa.h:(.init.text+0x348): undefined reference to `handle_fpe'
ld: ./arch/mips/include/asm/msa.h:(.init.text+0x354): undefined reference to `handle_fpe'
ld: ./arch/mips/include/asm/msa.h:(.init.text+0x360): undefined reference to `handle_fpe'
where the CONFIG_MIPS_FP_SUPPORT configuration option has been disabled.
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk> Reported-by: Stephen Zhang <starzhangzsd@gmail.com> Fixes: 0ebb2f4159af ("MIPS: IP27: Update/restructure CPU overrides") Cc: stable@vger.kernel.org # v4.2+ Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Current rxe_requester() doesn't generate a completion when processing an
unsupported/invalid opcode. If rxe driver doesn't support a new opcode
(e.g. RDMA Atomic Write) and RDMA library supports it, an application
using the new opcode can reproduce this issue. Fix the issue by calling
"goto err;".
The bt number of cqc_timer of HIP09 increases compared with that of HIP08.
Therefore, cqc_timer_bt_num and num_cqc_timer do not match. As a result,
the driver may fail to allocate cqc_timer. So the driver needs to uniquely
uses cqc_timer_bt_num to represent the bt number of cqc_timer.
Fixes: 0e40dc2f70cd ("RDMA/hns: Add timer allocation support for hip08") Link: https://lore.kernel.org/r/20220429093545.58070-1-liangwenpeng@huawei.com Signed-off-by: Yixing Liu <liuyixing1@huawei.com> Signed-off-by: Wenpeng Liang <liangwenpeng@huawei.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
list_is_head() was added recently[1], and didn't have a KUnit test. The
implementation is trivial, so it's not a particularly exciting test, but
it'd be nice to get back to full coverage of the list functions.
For the hybrid system, the "slots" event changes to "cpu_core/slots/", need
extend API arch_evsel__must_be_in_group() to support hybrid systems.
In the origin code, for hybrid system event "cpu_core/slots/", the output
of the API arch_evsel__must_be_in_group() is "false" (in fact,it should be
"true"). Currently only one API evsel__remove_from_group() calls it. In
evsel__remove_from_group(), it adds the second condition to check, so the
output of evsel__remove_from_group() still is correct. That's the reason
why there isn't an instant error. I'd like to fix the issue found in API
arch_evsel__must_be_in_group() in case someone else using the function in
the other place.
Commit 54de76c01239 ("kselftest/cgroup: fix test_stress.sh to use OUTPUT
dir") changes the test_core command path from . to $OUTPUT. However,
variable OUTPUT may not be defined if the command is run interactively.
Fix that by using ${OUTPUT:-.} to cover both cases.
Currently the (possibly compound) pages used for receive buffers are
freed using __free_pages(). But according to this comment above the
definition of that function, that's wrong:
If you want to use the page's reference count to decide
when to free the allocation, you should allocate a compound
page, and use put_page() instead of __free_pages().
Convert the call to __free_pages() in ipa_endpoint_replenish_one()
to use put_page() instead.
Fixes: 6a606b90153b8 ("net: ipa: allocate transaction in replenish loop") Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently the (possibly compound) page used for receive buffers are
freed using __free_pages(). But according to this comment above the
definition of that function, that's wrong:
If you want to use the page's reference count to decide when
to free the allocation, you should allocate a compound page,
and use put_page() instead of __free_pages().
Convert the call to __free_pages() in ipa_endpoint_trans_release()
to use put_page() instead.
It is possibe that probe failure issue happens when the device
and its child_device's probe happens at the same time.
In coresight_make_links, has_conns_grp is true for parent, but
has_conns_grp is false for child device as has_conns_grp is set
to true in coresight_create_conns_sysfs_group. The probe of parent
device will fail at this condition. Add has_conns_grp check for
child device before make the links and make the process from
device_register to connection_create be atomic to avoid this
probe failure issue.
Cc: stable@vger.kernel.org Suggested-by: Suzuki K Poulose <suzuki.poulose@arm.com> Suggested-by: Mike Leach <mike.leach@linaro.org> Signed-off-by: Mao Jinlong <quic_jinlmao@quicinc.com> Link: https://lore.kernel.org/r/20220309142206.15632-1-quic_jinlmao@quicinc.com
[ Added Cc stable ] Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
iolatency needs to track the number of inflight IOs per cgroup. As this
tracking can be expensive, it is disabled when no cgroup has iolatency
configured for the device. To ensure that the inflight counters stay
balanced, iolatency_set_limit() freezes the request_queue while manipulating
the enabled counter, which ensures that no IO is in flight and thus all
counters are zero.
Unfortunately, iolatency_set_limit() isn't the only place where the enabled
counter is manipulated. iolatency_pd_offline() can also dec the counter and
trigger disabling. As this disabling happens without freezing the q, this
can easily happen while some IOs are in flight and thus leak the counts.
This can be easily demonstrated by turning on iolatency on an one empty
cgroup while IOs are in flight in other cgroups and then removing the
cgroup. Note that iolatency shouldn't have been enabled elsewhere in the
system to ensure that removing the cgroup disables iolatency for the whole
device.
The following keeps flipping on and off iolatency on sda:
If a cgroup with stuck inflight ends up getting throttled, the throttled IOs
will never get issued as there's no completion event to wake it up leading
to an indefinite hang.
This patch fixes the bug by unifying enable handling into a work item which
is automatically kicked off from iolatency_set_min_lat_nsec() which is
called from both iolatency_set_limit() and iolatency_pd_offline() paths.
Punting to a work item is necessary as iolatency_pd_offline() is called
under spinlocks while freezing a request_queue requires a sleepable context.
This also simplifies the code reducing LOC sans the comments and avoids the
unnecessary freezes which were happening whenever a cgroup's latency target
is newly set or cleared.
Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
if (cfile) {
cifsFileInfo_put(cfile); // sends SMB2_CLOSE to the server
cfile = NULL;
This is triggered by smb2_query_path_info operation that happens during
revalidate_dentry. In smb2_query_path_info, get_readable_path is called to
load the cfile, increasing the reference counter. If in the meantime, this
reference becomes the very last, this call to cifsFileInfo_put(cfile) will
trigger a SMB2_CLOSE request sent to the server just before sending this compound
request – and so then the compound request fails either with EBADF/EIO depending
on the timing at the server, because the handle is already closed.
In the first scenario, the race seems to be happening between smb2_query_path_info
triggered by the rename operation, and between “cleanup” of asynchronous writes – while
fsync(fd) likely waits for the asynchronous writes to complete, releasing the writeback
structures can happen after the close(fd) call. So the EBADF/EIO errors will pop up if
the timing is such that:
1) There are still outstanding references after close(fd) in the writeback structures
2) smb2_query_path_info successfully fetches the cfile, increasing the refcounter by 1
3) All writeback structures release the same cfile, reducing refcounter to 1
4) smb2_compound_op is called with that cfile
In the second scenario, the race seems to be similar – here open triggers the
smb2_query_path_info operation, and if all other threads in the meantime decrease the
refcounter to 1 similarly to the first scenario, again SMB2_CLOSE will be sent to the
server just before issuing the compound request. This case is harder to reproduce.
See https://bugzilla.samba.org/show_bug.cgi?id=15051
Cc: stable@vger.kernel.org Fixes: 8de9e86c67ba ("cifs: create a helper to find a writeable handle by path name") Signed-off-by: Ondrej Hubsch <ohubsch@purestorage.com> Reviewed-by: Ronnie Sahlberg <lsahlber@redhat.com> Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The cephfs kernel client started to show the message:
ceph: mds0 session blocklisted
when mounting a filesystem. This is due to the fact that the session
messages are being incorrectly decoded: the skip needs to take into
account the 'len'.
While there, fixed some whitespaces too.
Cc: stable@vger.kernel.org Fixes: e1c9788cb397 ("ceph: don't rely on error_string to validate blocklisted session.") Signed-off-by: Luís Henriques <lhenriques@suse.de> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since commit 766c6b63aa04 ("spi: fix client driver breakages when using
GPIO descriptors"), the panel has been blank due to an inverted CS GPIO.
In order to correct this, drop the spi-cs-high from the panel SPI device.
Commit 4782c0a5dd88 ("clk: tegra: Don't deassert reset on enabling
clocks") removed deassertion of reset lines when enabling peripheral
clocks. This breaks the initialization of the DFLL driver which relied
on this behaviour.
Fix this problem by adding explicit deassert/assert requests to the
driver. Tested on Google Pixel C.
Commit 4782c0a5dd88 ("clk: tegra: Don't deassert reset on enabling
clocks") removed deassertion of reset lines when enabling peripheral
clocks. This breaks the initialization of the DFLL driver which relied
on this behaviour.
In order to be able to fix this, add the corresponding reset to the DT.
Tested on Google Pixel C.
The list iterator value 'crtc' will *always* be set and non-NULL by
list_for_each_entry(), so it is incorrect to assume that the iterator
value will be NULL if the list is empty or no element is found.
To fix the bug, return 'crtc' when found, otherwise return NULL.
The list iterator value 'encoder' will *always* be set and non-NULL
by list_for_each_entry(), so it is incorrect to assume that the
iterator value will be NULL if the list is empty or no element
is found.
To fix the bug, use a new variable 'iter' as the list iterator,
while use the original variable 'encoder' as a dedicated pointer
to point to the found element.
When x_char is to be sent, the TX path overwrites whatever is in the
circular buffer at offset 0 with x_char and sends it using
pch_uart_hal_write(). I don't understand how this was supposed to work
if xmit->buf[0] already contained some character. It must have been
lost.
Remove this whole pop_tx_x() concept and do the work directly in the
callers. (Without printing anything using dev_dbg().)
The journal no-space deadlock was reported time to time. Such deadlock
can happen in the following situation.
When all journal buckets are fully filled by active jset with heavy
write I/O load, the cache set registration (after a reboot) will load
all active jsets and inserting them into the btree again (which is
called journal replay). If a journaled bkey is inserted into a btree
node and results btree node split, new journal request might be
triggered. For example, the btree grows one more level after the node
split, then the root node record in cache device super block will be
upgrade by bch_journal_meta() from bch_btree_set_root(). But there is no
space in journal buckets, the journal replay has to wait for new journal
bucket to be reclaimed after at least one journal bucket replayed. This
is one example that how the journal no-space deadlock happens.
The solution to avoid the deadlock is to reserve 1 journal bucket in
run time, and only permit the reserved journal bucket to be used during
cache set registration procedure for things like journal replay. Then
the journal space will never be fully filled, there is no chance for
journal no-space deadlock to happen anymore.
This patch adds a new member "bool do_reserve" in struct journal, it is
inititalized to 0 (false) when struct journal is allocated, and set to
1 (true) by bch_journal_space_reserve() when all initialization done in
run_cache_set(). In the run time when journal_reclaim() tries to
allocate a new journal bucket, free_journal_buckets() is called to check
whether there are enough free journal buckets to use. If there is only
1 free journal bucket and journal->do_reserve is 1 (true), the last
bucket is reserved and free_journal_buckets() will return 0 to indicate
no free journal bucket. Then journal_reclaim() will give up, and try
next time to see whetheer there is free journal bucket to allocate. By
this method, there is always 1 jouranl bucket reserved in run time.
During the cache set registration, journal->do_reserve is 0 (false), so
the reserved journal bucket can be used to avoid the no-space deadlock.
After making bch_sectors_dirty_init() being multithreaded, the existing
incremental dirty sector counting in bch_root_node_dirty_init() doesn't
release btree occupation after iterating 500000 (INIT_KEYS_EACH_TIME)
bkeys. Because a read lock is added on btree root node to prevent the
btree to be split during the dirty sectors counting, other I/O requester
has no chance to gain the write lock even restart bcache_btree().
That is to say, the incremental dirty sectors counting is incompatible
to the multhreaded bch_sectors_dirty_init(). We have to choose one and
drop another one.
In my testing, with 512 bytes random writes, I generate 1.2T dirty data
and a btree with 400K nodes. With single thread and incremental dirty
sectors counting, it takes 30+ minites to register the backing device.
And with multithreaded dirty sectors counting, the backing device
registration can be accomplished within 2 minutes.
The 30+ minutes V.S. 2- minutes difference makes me decide to keep
multithreaded bch_sectors_dirty_init() and drop the incremental dirty
sectors counting. This is what this patch does.
But INIT_KEYS_EACH_TIME is kept, in sectors_dirty_init_fn() the CPU
will be released by cond_resched() after every INIT_KEYS_EACH_TIME keys
iterated. This is to avoid the watchdog reports a bogus soft lockup
warning.
Fixes: b144e45fc576 ("bcache: make bch_sectors_dirty_init() to be multithreaded") Signed-off-by: Coly Li <colyli@suse.de> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20220524102336.10684-4-colyli@suse.de Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit b144e45fc576 ("bcache: make bch_sectors_dirty_init() to be
multithreaded") makes bch_sectors_dirty_init() to be much faster
when counting dirty sectors by iterating all dirty keys in the btree.
But it isn't in ideal shape yet, still can be improved.
This patch does the following changes to improve current parallel dirty
keys iteration on the btree,
- Add read lock to root node when multiple threads iterating the btree,
to prevent the root node gets split by I/Os from other registered
bcache devices.
- Remove local variable "char name[32]" and generate kernel thread name
string directly when calling kthread_run().
- Allocate "struct bch_dirty_init_state state" directly on stack and
avoid the unnecessary dynamic memory allocation for it.
- Decrease BCH_DIRTY_INIT_THRD_MAX from 64 to 12 which is enough indeed.
- Increase &state->started to count created kernel thread after it
succeeds to create.
- When wait for all dirty key counting threads to finish, use
wait_event() to replace wait_event_interruptible().
With the above changes, the code is more clear, and some potential error
conditions are avoided.
Fixes: b144e45fc576 ("bcache: make bch_sectors_dirty_init() to be multithreaded") Signed-off-by: Coly Li <colyli@suse.de> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20220524102336.10684-3-colyli@suse.de Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 8e7102273f59 ("bcache: make bch_btree_check() to be
multithreaded") makes bch_btree_check() to be much faster when checking
all btree nodes during cache device registration. But it isn't in ideal
shap yet, still can be improved.
This patch does the following thing to improve current parallel btree
nodes check by multiple threads in bch_btree_check(),
- Add read lock to root node while checking all the btree nodes with
multiple threads. Although currently it is not mandatory but it is
good to have a read lock in code logic.
- Remove local variable 'char name[32]', and generate kernel thread name
string directly when calling kthread_run().
- Allocate local variable "struct btree_check_state check_state" on the
stack and avoid unnecessary dynamic memory allocation for it.
- Reduce BCH_BTR_CHKTHREAD_MAX from 64 to 12 which is enough indeed.
- Increase check_state->started to count created kernel thread after it
succeeds to create.
- When wait for all checking kernel threads to finish, use wait_event()
to replace wait_event_interruptible().
With this change, the code is more clear, and some potential error
conditions are avoided.
Fixes: 8e7102273f59 ("bcache: make bch_btree_check() to be multithreaded") Signed-off-by: Coly Li <colyli@suse.de> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20220524102336.10684-2-colyli@suse.de Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The two bugs are here:
if (encoder) {
if (bridge && bridge->timings)
The list iterator value 'encoder/bridge' will *always* be set and
non-NULL by drm_for_each_encoder()/list_for_each_entry(), so it is
incorrect to assume that the iterator value will be NULL if the
list is empty or no element is found.
To fix the bug, use a new variable '*_iter' as the list iterator,
while use the old variable 'encoder/bridge' as a dedicated pointer
to point to the found element.
If the previous list_for_each_entry_continue_rcu() don't exit early
(no goto hit inside the loop), the iterator 'cvif' after the loop
will be a bogus pointer to an invalid structure object containing
the HEAD (&ar->vif_list). As a result, the use of 'cvif' after that
will lead to a invalid memory access (i.e., 'cvif->id': the invalid
pointer dereference when return back to/after the callsite in the
carl9170_update_beacon()).
The original intention should have been to return the valid 'cvif'
when found in list, NULL otherwise. So just return NULL when no
entry found, to fix this bug.
Cc: stable@vger.kernel.org Fixes: 1f1d9654e183c ("carl9170: refactor carl9170_update_beacon") Signed-off-by: Xiaomeng Tong <xiam0nd.tong@gmail.com> Acked-by: Christian Lamparter <chunkeey@gmail.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://lore.kernel.org/r/20220328122820.1004-1-xiam0nd.tong@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The driver has a custom put function for "DSP Voice Wake Up" which does
not generate event notifications on change, instead returning 0. Since we
already exit early in the case that there is no change this can be fixed
by unconditionally returning 1 at the end of the function.
Using not existing queues can panic the kernel with rtl8180/rtl8185 cards.
Ignore the skb priority for those cards, they only have one tx queue. Pierre
Asselin (pa@panix.com) reported the kernel crash in the Gentoo forum:
He also confirmed that this patch fixes the issue. In summary this happened:
After updating wpa_supplicant from 2.9 to 2.10 the kernel crashed with a
"divide error: 0000" when connecting to an AP. Control port tx now tries to
use IEEE80211_AC_VO for the priority, which wpa_supplicants starts to use in
2.10.
Since only the rtl8187se part of the driver supports QoS, the priority
of the skb is set to IEEE80211_AC_BE (2) by mac80211 for rtl8180/rtl8185
cards.
rtl8180 is then unconditionally reading out the priority and finally crashes on
drivers/net/wireless/realtek/rtl818x/rtl8180/dev.c line 544 without this
patch:
idx = (ring->idx + skb_queue_len(&ring->queue)) % ring->entries
"ring->entries" is zero for rtl8180/rtl8185 cards, tx_ring[2] never got
initialized.
Cc: stable@vger.kernel.org Reported-by: pa@panix.com Tested-by: pa@panix.com Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://lore.kernel.org/r/20220422145228.7567-1-alexander@wetzel-home.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The commit a69755b18774 ("xtensa simdisk: switch to proc_create_data()")
split read operation into two parts, first retrieving the path when it's
non-null and second retrieving the trailing '\n'. However when the path
is non-null the first simple_read_from_buffer updates ppos, and the
second simple_read_from_buffer returns 0 if ppos is greater than 1 (i.e.
almost always). As a result reading from that proc file is almost always
empty.
Fix it by making a temporary copy of the path with the trailing '\n' and
using simple_read_from_buffer on that copy.
Cc: stable@vger.kernel.org Fixes: a69755b18774 ("xtensa simdisk: switch to proc_create_data()") Signed-off-by: Yi Yang <yiyang13@huawei.com> Signed-off-by: Max Filippov <jcmvbkbc@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We forget to call untrack_pfn() to pair with track_pfn_remap() when range
is not allowed to hotplug. Fix it by jump err_kasan.
Link: https://lkml.kernel.org/r/20220531122643.25249-1-linmiaohe@huawei.com Fixes: bca3feaa0764 ("mm/memory_hotplug: prevalidate the address range being added with platform") Signed-off-by: Miaohe Lin <linmiaohe@huawei.com> Reviewed-by: David Hildenbrand <david@redhat.com> Acked-by: Muchun Song <songmuchun@bytedance.com> Cc: Anshuman Khandual <anshuman.khandual@arm.com> Cc: Oscar Salvador <osalvador@suse.de> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The routine huge_pmd_unshare() is passed a pointer to an address
associated with an area which may be unshared. If unshare is successful
this address is updated to 'optimize' callers iterating over huge page
addresses. For the optimization to work correctly, address should be
updated to the last huge page in the unmapped/unshared area. However, in
the common case where the passed address is PUD_SIZE aligned, the address
is incorrectly updated to the address of the preceding huge page. That
wastes CPU cycles as the unmapped/unshared range is scanned twice.
Link: https://lkml.kernel.org/r/20220524205003.126184-1-mike.kravetz@oracle.com Fixes: 39dde65c9940 ("shared page table for hugetlb page") Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com> Acked-by: Muchun Song <songmuchun@bytedance.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
With gcc version 12.0.1 20220401 (Red Hat 12.0.1-0), building with
defconfig results in the following compilation error:
| CC mm/swapfile.o
| mm/swapfile.c: In function `setup_swap_info':
| mm/swapfile.c:2291:47: error: array subscript -1 is below array bounds
| of `struct plist_node[]' [-Werror=array-bounds]
| 2291 | p->avail_lists[i].prio = 1;
| | ~~~~~~~~~~~~~~^~~
| In file included from mm/swapfile.c:16:
| ./include/linux/swap.h:292:27: note: while referencing `avail_lists'
| 292 | struct plist_node avail_lists[]; /*
| | ^~~~~~~~~~~
This is due to the compiler detecting that the mask in
node_states[__state] could theoretically be zero, which would lead to
first_node() returning -1 through find_first_bit.
I believe that the warning/error is legitimate. I first tried adding a
test to check that the node mask is not emtpy, since a similar test exists
in the case where MAX_NUMNODES == 1.
However, adding the if statement causes other warnings to appear in
for_each_cpu_node_but, because it introduces a dangling else ambiguity.
And unfortunately, GCC is not smart enough to detect that the added test
makes the case where (node) == -1 impossible, so it still complains with
the same message.
This is why I settled on replacing that with a harmless, but relatively
useless (node) >= 0 test. Based on the warning for the dangling else, I
also decided to fix the case where MAX_NUMNODES == 1 by moving the
condition inside the for loop. It will still only be tested once. This
ensures that the meaning of an else following for_each_node_mask or
derivatives would not silently have a different meaning depending on the
configuration.
Link: https://lkml.kernel.org/r/20220414150855.2407137-3-dinechin@redhat.com Signed-off-by: Christophe de Dinechin <christophe@dinechin.org> Signed-off-by: Christophe de Dinechin <dinechin@redhat.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Ben Segall <bsegall@google.com> Cc: "Michael S. Tsirkin" <mst@redhat.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Dietmar Eggemann <dietmar.eggemann@arm.com> Cc: Vincent Guittot <vincent.guittot@linaro.org> Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Daniel Bristot de Oliveira <bristot@redhat.com> Cc: Jason Wang <jasowang@redhat.com> Cc: Zhen Lei <thunder.leizhen@huawei.com> Cc: Juri Lelli <juri.lelli@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Peter Pavlisko reported the following problem on kernel bugzilla 216007.
When I try to extract an uncompressed tar archive (2.6 milion
files, 760.3 GiB in size) on newly created (empty) XFS file system,
after first low tens of gigabytes extracted the process hangs in
iowait indefinitely. One CPU core is 100% occupied with iowait,
the other CPU core is idle (on 2-core Intel Celeron G1610T).
It was bisected to c9fa563072e1 ("xfs: use alloc_pages_bulk_array() for
buffers") but XFS is only the messenger. The problem is that nothing is
waking kswapd to reclaim some pages at a time the PCP lists cannot be
refilled until some reclaim happens. The bulk allocator checks that there
are some pages in the array and the original intent was that a bulk
allocator did not necessarily need all the requested pages and it was best
to return as quickly as possible.
This was fine for the first user of the API but both NFS and XFS require
the requested number of pages be available before making progress. Both
could be adjusted to call the page allocator directly if a bulk allocation
fails but it puts a burden on users of the API. Adjust the semantics to
attempt at least one allocation via __alloc_pages() before returning so
kswapd is woken if necessary.
It was reported via bugzilla that the patch addressed the problem and that
the tar extraction completed successfully. This may also address bug
215975 but has yet to be confirmed.
BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=216007 BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=215975 Link: https://lkml.kernel.org/r/20220526091210.GC3441@techsingularity.net Fixes: 387ba26fb1cb ("mm/page_alloc: add a bulk page allocator") Signed-off-by: Mel Gorman <mgorman@techsingularity.net> Cc: "Darrick J. Wong" <djwong@kernel.org> Cc: Dave Chinner <dchinner@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Jesper Dangaard Brouer <brouer@redhat.com> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: <stable@vger.kernel.org> [5.13+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This reverts commit a4efc174b382fcdb which introduced a regression issue
that when there're multiple processes allocating dma memory in parallel by
calling dma_alloc_coherent(), it may fail sometimes as follows:
Error log:
cma: cma_alloc: linux,cma: alloc failed, req-size: 148 pages, ret: -16
cma: number of available pages:
3@125+20@172+12@236+4@380+32@736+17@2287+23@2473+20@36076+99@40477+108@40852+44@41108+20@41196+108@41364+108@41620+
108@42900+108@43156+483@44061+1763@45341+1440@47712+20@49324+20@49388+5076@49452+2304@55040+35@58141+20@58220+20@58284+
7188@58348+84@66220+7276@66452+227@74525+6371@75549=> 33161 free of 81920 total pages
When issue happened, we saw there were still 33161 pages (129M) free CMA
memory and a lot available free slots for 148 pages in CMA bitmap that we
want to allocate.
When dumping memory info, we found that there was also ~342M normal
memory, but only 1352K CMA memory left in buddy system while a lot of
pageblocks were isolated.
The root cause of this issue is that since commit a4efc174b382 ("mm/cma.c:
remove redundant cma_mutex lock"), CMA supports concurrent memory
allocation. It's possible that the memory range process A trying to alloc
has already been isolated by the allocation of process B during memory
migration.
The problem here is that the memory range isolated during one allocation
by start_isolate_page_range() could be much bigger than the real size we
want to alloc due to the range is aligned to MAX_ORDER_NR_PAGES.
Taking an ARMv7 platform with 1G memory as an example, when
MAX_ORDER_NR_PAGES is big (e.g. 32M with max_order 14) and CMA memory is
relatively small (e.g. 128M), there're only 4 MAX_ORDER slot, then it's
very easy that all CMA memory may have already been isolated by other
processes when one trying to allocate memory using dma_alloc_coherent().
Since current CMA code will only scan one time of whole available CMA
memory, then dma_alloc_coherent() may easy fail due to contention with
other processes.
This patch simply falls back to the original method that using cma_mutex
to make alloc_contig_range() run sequentially to avoid the issue.
The data type of the return value of the iommu_map_sg_atomic
is ssize_t, but the data type of iova size is size_t,
e.g. one is int while the other is unsigned int.
When iommu_map_sg_atomic return value is compared with iova size,
it will force the signed int to be converted to unsigned int, if
iova map fails and iommu_map_sg_atomic return error code is less
than 0, then (ret < iova_len) is false, which will to cause not
do free iova, and the master can still successfully get the iova
of map fail, which is not expected.
Therefore, we need to check the return value of iommu_map_sg_atomic
in two cases according to whether it is less than 0.
Fixes: ad8f36e4b6b1 ("iommu: return full error code from iommu_map_sg[_atomic]()") Signed-off-by: Yunfei Wang <yf.wang@mediatek.com> Cc: <stable@vger.kernel.org> # 5.15.* Reviewed-by: Robin Murphy <robin.murphy@arm.com> Reviewed-by: Miles Chen <miles.chen@mediatek.com> Link: https://lore.kernel.org/r/20220507085204.16914-1-yf.wang@mediatek.com Signed-off-by: Joerg Roedel <jroedel@suse.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The bug is here:
if (!iommu || iommu->dev->of_node != spec->np) {
The list iterator value 'iommu' will *always* be set and non-NULL by
list_for_each_entry(), so it is incorrect to assume that the iterator
value will be NULL if the list is empty or no element is found (in fact,
it will point to a invalid structure object containing HEAD).
To fix the bug, use a new value 'iter' as the list iterator, while use
the old value 'iommu' as a dedicated variable to point to the found one,
and remove the unneeded check for 'iommu->dev->of_node != spec->np'
outside the loop.
outstanding credits must be initialized to 0,
because it means the sum of credits consumed by
in-flight requests.
And outstanding credits must be compared with
total credits in smb2_validate_credit_charge(),
because total credits are the sum of credits
granted by ksmbd.
This patch fix the following error,
while frametest with Windows clients:
Limits exceeding the maximum allowable outstanding requests,
given : 128, pending : 8065
Fixes: b589f5db6d4a ("ksmbd: limits exceeding the maximum allowable outstanding requests") Cc: stable@vger.kernel.org Signed-off-by: Hyunchul Lee <hyc.lee@gmail.com> Reported-by: Yufan Chen <wiz.chen@gmail.com> Tested-by: Yufan Chen <wiz.chen@gmail.com> Acked-by: Namjae Jeon <linkinjeon@kernel.org> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
One way to trigger this is:
1. load a livepatch that patches kernel function xxx;
2. run bpftrace -e 'kfunc:xxx {}', this will fail (expected for now);
3. repeat #2 => gpf.
This is because the entry is added to direct_functions, but not removed.
Fix this by remove the entry from direct_functions when
register_ftrace_direct fails.
Also remove the last trailing space from ftrace.c, so we don't have to
worry about it anymore.
Link: https://lkml.kernel.org/r/20220524170839.900849-1-song@kernel.org Cc: stable@vger.kernel.org Fixes: 763e34e74bb7 ("ftrace: Add register_ftrace_direct()") Signed-off-by: Song Liu <song@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since commit d1bcae833b32f1 ("ELF: Don't generate unused section
symbols") [1], binutils (v2.36+) started dropping section symbols that
it thought were unused. This isn't an issue in general, but with
kexec_file.c, gcc is placing kexec_arch_apply_relocations[_add] into a
separate .text.unlikely section and the section symbol ".text.unlikely"
is being dropped. Due to this, recordmcount is unable to find a non-weak
symbol in .text.unlikely to generate a relocation record against.
Address this by dropping the weak attribute from these functions.
Instead, follow the existing pattern of having architectures #define the
name of the function they want to override in their headers.
syscall_stub_data() expects the data_count parameter to be the number of
longs, not bytes.
==================================================================
BUG: KASAN: stack-out-of-bounds in syscall_stub_data+0x70/0xe0
Read of size 128 at addr 000000006411f6f0 by task swapper/1
The previous fix here was only partially correct, it did
result in returning a proper error value in case of error,
but it also clobbered the pid that we need to return from
this function (not just zero for success).
As a result, it returned 0 here, but later this is treated
as a pid and used to kill the process, but since it's now
0 we kill(0, SIGKILL), which makes UML kill itself rather
than just the helper thread.
Fix that and make it more obvious by using a separate
variable for the pid.
Fixes: ccf1236ecac4 ("um: fix error return code in winch_tramp()") Reported-and-tested-by: Nathan Chancellor <nathan@kernel.org> Signed-off-by: Johannes Berg <johannes.berg@intel.com> Cc: stable@vger.kernel.org Signed-off-by: Richard Weinberger <richard@nod.at> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If DMA (PCI over virtio) is enabled, then some drivers may
enable CONFIG_DMA_OPS as well, and then we pull in the x86
definition of get_arch_dma_ops(), which uses the dma_ops
symbol, which isn't defined.
Since we don't have real DMA ops nor any kind of IOMMU fix
this in the simplest possible way: pull in the asm-generic
file instead of inheriting the x86 one. It's not clear why
those drivers that do (e.g. VDPA) "select DMA_OPS", and if
they'd even work with this, but chances are nobody will be
wanting to do that anyway, so fixing the build failure is
good enough.
Reported-by: Randy Dunlap <rdunlap@infradead.org> Fixes: 68f5d3f3b654 ("um: add PCI over virtio emulation driver") Signed-off-by: Johannes Berg <johannes.berg@intel.com> Tested-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Randy Dunlap <rdunlap@infradead.org> Cc: stable@vger.kernel.org Signed-off-by: Richard Weinberger <richard@nod.at> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If a device implementation crashes, virtio_uml will mark it
as dead by calling virtio_break_device() and scheduling the
work that will remove it.
This still seems like the right thing to do, but it's done
directly while reading the message, and if time-travel is
used, this is in the time-travel handler, outside of the
normal Linux machinery. Therefore, we cannot acquire locks
or do normal "linux-y" things because e.g. lockdep will be
confused about the context.
Move handling this situation out of the read function and
into the actual IRQ handler and response handling instead,
so that in the case of time-travel we don't call it in the
wrong context.
Chances are the system will still crash immediately, since
the device implementation crashing may also cause the time-
travel controller to go down, but at least all of that now
happens without strange warnings from lockdep.
Fixes: c8177aba37ca ("um: time-travel: rework interrupt handling in ext mode") Cc: stable@vger.kernel.org Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Richard Weinberger <richard@nod.at> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In client mode, we can't connect to hidden SSID APs or SSIDs not advertised
in beacons on DFS channels, since we're forced to passive scan. Fix this by
sending out a probe request immediately after the first beacon, if active
scan was requested by the user.
Cc: stable@vger.kernel.org Reported-by: Catrinel Catrinescu <cc@80211.de> Signed-off-by: Felix Fietkau <nbd@nbd.name> Link: https://lore.kernel.org/r/20220420104907.36275-1-nbd@nbd.name Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fixes an issue caught by KASAN about use-after-free in mt76_txq_schedule
by protecting mtxq->wcid with rcu_lock between mt76_txq_schedule and
sta_info_[alloc, free].
[18853.876689] ==================================================================
[18853.876751] BUG: KASAN: use-after-free in mt76_txq_schedule+0x204/0xaf8 [mt76]
[18853.876773] Read of size 8 at addr ffffffaf989a2138 by task mt76-tx phy0/883
[18853.876786]
[18853.876810] CPU: 5 PID: 883 Comm: mt76-tx phy0 Not tainted 5.10.100-fix-510-56778d365941-kasan #5 0b01fbbcf41a530f52043508fec2e31a4215
Cc: stable@vger.kernel.org Reported-by: Sean Wang <sean.wang@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix access illegal address problem in following condition:
There are multiple devfreq cooling devices in system, some of them has
EM model but others do not. Energy model ops such as state2power will
append to global devfreq_cooling_ops when the cooling device with
EM model is registered. It makes the cooling device without EM model
also use devfreq_cooling_ops after appending when registered later by
of_devfreq_cooling_register_power() or of_devfreq_cooling_register().
The IPA governor regards the cooling devices without EM model as a power
actor, because they also have energy model ops, and will access illegal
address at dfc->em_pd when execute cdev->ops->get_requested_power,
cdev->ops->state2power or cdev->ops->power2state.
Fixes: 615510fe13bd2 ("thermal: devfreq_cooling: remove old power model and use EM") Cc: 5.13+ <stable@vger.kernel.org> # 5.13+ Signed-off-by: Kant Fan <kant@allwinnertech.com> Reviewed-by: Lukasz Luba <lukasz.luba@arm.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When irq-xtensa-mx chip is used in non-SMP configuration its
irq_set_affinity callback is not called leaving IRQ affinity set empty.
As a result IRQ delivery does not work in that configuration.
Initialize IRQ affinity of the xtensa MX interrupt distributor to CPU 0
for all external IRQ lines.
Cc: stable@vger.kernel.org Signed-off-by: Max Filippov <jcmvbkbc@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Register ARMADA_370_XP_INT_FABRIC_MASK_OFFS is Armada 370 and XP specific
and on new Armada platforms it has different meaning. It does not configure
Performance Counter Overflow interrupt masking. So do not touch this
register on non-A370/XP platforms (A375, A38x and A39x).
Signed-off-by: Pali Rohár <pali@kernel.org> Cc: stable@vger.kernel.org Fixes: 28da06dfd9e4 ("irqchip: armada-370-xp: Enable the PMU interrupts") Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Marc Zyngier <maz@kernel.org> Link: https://lore.kernel.org/r/20220425113706.29310-1-pali@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
These patch_text implementations are using stop_machine_cpuslocked
infrastructure with atomic cpu_count. The original idea: When the
master CPU patch_text, the others should wait for it. But current
implementation is using the first CPU as master, which couldn't
guarantee the remaining CPUs are waiting. This patch changes the
last CPU as the master to solve the potential risk.
Occasionally, user-land applications initiate longer timeout values for certain commands
through ioctl() system call. But so far we are still using a fixed timeout of 10 seconds
in mmc_poll_for_busy() on the ioctl() path, even if a custom timeout is specified in the
userspace application. This patch allows custom timeout values to override this default
timeout values on the ioctl path.
When multiplying of different types, an overflow is possible even when
storing the result in a larger type. This is because the conversion is
done after the multiplication. So arithmetic overflow and thus in
incorrect value is possible.
Correct an instance of this in the inter packet delay calculation. Fix by
ensuring one of the operands is u64 which will promote the other to u64 as
well ensuring no overflow.
Cc: stable@vger.kernel.org Fixes: 7724105686e7 ("IB/hfi1: add driver files") Link: https://lore.kernel.org/r/20220520183712.48973.29855.stgit@awfm-01.cornelisnetworks.com Reviewed-by: Mike Marciniszyn <mike.marciniszyn@cornelisnetworks.com> Signed-off-by: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add a config option to guard (future) usage of asm_volatile_goto() that
includes "tied outputs", i.e. "+" constraints that specify both an input
and output parameter. clang-13 has a bug[1] that causes compilation of
such inline asm to fail, and KVM wants to use a "+m" constraint to
implement a uaccess form of CMPXCHG[2]. E.g. the test code fails with
<stdin>:1:29: error: invalid operand in inline asm: '.long (${1:l}) - .'
int foo(int *x) { asm goto (".long (%l[bar]) - .\n": "+m"(*x) ::: bar); return *x; bar: return 0; }
^
<stdin>:1:29: error: unknown token in expression
<inline asm>:1:9: note: instantiated into assembly here
.long () - .
^
2 errors generated.
on clang-13, but passes on gcc (with appropriate asm goto support). The
bug is fixed in clang-14, but won't be backported to clang-13 as the
changes are too invasive/risky.
gcc also had a similar bug[3], fixed in gcc-11, where gcc failed to
account for its behavior of assigning two numbers to tied outputs (one
for input, one for output) when evaluating symbolic references.
The original 'ima' measurement list template contains a hash, defined
as 20 bytes, and a null terminated pathname, limited to 255
characters. Other measurement list templates permit both larger hashes
and longer pathnames. When the "ima" template is configured as the
default, a new measurement list template (ima_template=) must be
specified before specifying a larger hash algorithm (ima_hash=) on the
boot command line.
To avoid this boot command line ordering issue, remove the legacy "ima"
template configuration option, allowing it to still be specified on the
boot command line.
The root cause of this issue is that during the processing of ima_hash,
we would try to check whether the hash algorithm is compatible with the
template. If the template is not set at the moment we do the check, we
check the algorithm against the configured default template. If the
default template is "ima", then we reject any hash algorithm other than
sha1 and md5.
For example, if the compiled default template is "ima", and the default
algorithm is sha1 (which is the current default). In the cmdline, we put
in "ima_hash=sha256 ima_template=ima-ng". The expected behavior would be
that ima starts with ima-ng as the template and sha256 as the hash
algorithm. However, during the processing of "ima_hash=",
"ima_template=" has not been processed yet, and hash_setup would check
the configured hash algorithm against the compiled default: ima, and
reject sha256. So at the end, the hash algorithm that is actually used
will be sha1.
With template "ima" removed from the configured default, we ensure that
the default tempalte would at least be "ima-ng" which allows for
basically any hash algorithm.
This change would not break the algorithm compatibility checks for IMA.
Add H264 level 1.0, 4.1, 4.2 to the list of supported formats.
While the hardware does not fully support these levels, it does support
most of them. The constraints on frame size and pixel formats already
cover the limitation.
This fixes negotiation of level on GStreamer 1.17.1.
projects / thirdparty / kernel / stable.git / log
