If the sqpoll thread has died, the out condition doesn't remove the
waiting task from the waitqueue. The goto and check are not needed, just
make it a break condition after setting the error value. That ensures
that we always remove ourselves from sqo_sq_wait waitqueue.
The iterator can not be greater than ATC_MAX_DSCR_TRIALS, as the for loop
will stop when i == ATC_MAX_DSCR_TRIALS. While here, use the common "i"
name for the iterator.
at_hdmac uses __raw_writel for register writes. In the absence of a
barrier, the CPU may reorder the register operations.
Introduce a write memory barrier so that the CPU does not reorder the
channel enable, thus the start of the transfer, without making sure that
all the pre-required register fields are already written.
In case the controller detected an error, the code took the chance to move
all the queued (submitted) descriptors to the active (issued) list. This
was wrong as if there were any descriptors in the submitted list they were
moved to the issued list without actually issuing them to the controller,
thus a completion could be raised without even fireing the descriptor.
As it was before, the descriptor was issued to the hardware without adding
it to the active (issued) list. This could result in a completion of other
descriptor, or/and in the descriptor never being completed.
The tasklet (atc_advance_work()) did not held the channel lock when
retrieving the first active descriptor, causing concurrency problems if
issue_pending() was called in between. If issue_pending() was called
exactly after the lock was released in the tasklet (atc_advance_work()),
atc_chain_complete() could complete a descriptor for which the controller
has not yet raised an interrupt.
There's no need to hold the channel lock when freeing the memset buf, as
the operation has already completed. Free the memset buf without holding
the channel lock.
The descriptor was added to the free_list before calling the callback,
which could result in reissuing of the same descriptor and calling of a
single callback for both. Move the decriptor to the free list after the
callback is invoked.
atc_complete_all() had concurrency bugs, thus remove it:
1/ atc_complete_all() in its entirety was buggy, as when the atchan->queue
list (the one that contains descriptors that are not yet issued to the
hardware) contained descriptors, it fired just the first from the
atchan->queue, but moved all the desc from atchan->queue to
atchan->active_list and considered them all as fired. This could result in
calling the completion of a descriptor that was not yet issued to the
hardware.
2/ when in tasklet at atc_advance_work() time, atchan->active_list was
queried without holding the lock of the chan. This can result in
atchan->active_list concurrency problems between the tasklet and
issue_pending().
Now that the complete callback call was removed from
device_terminate_all(), we can protect the atchan->status with the channel
lock. The atomic bitops on atchan->status do not substitute proper locking
on the status, as one could still modify the status after the lock was
dropped in atc_terminate_all() but before the atomic bitops were executed.
The method was wrong because it violated the dmaengine API. For aborted
transfers the complete callback should not be called. Fix the behavior and
do not call the complete callback on device_terminate_all.
Multiple calls to atc_issue_pending() could result in a premature
completion of a descriptor from the atchan->active list, as the method
always completed the first active descriptor from the list. Instead,
issue_pending() should just take the first transaction descriptor from the
pending queue, move it to active_list and start the transfer.
Cyclic channels must too call issue_pending in order to start a transfer.
Start the transfer in issue_pending regardless of the type of channel.
This wrongly worked before, because in the past the transfer was started
at tx_submit level when only a desc in the transfer list.
tx_submit is supposed to push the current transaction descriptor to a
pending queue, waiting for issue_pending() to be called. issue_pending()
must start the transfer, not tx_submit(), thus remove atc_dostart() from
atc_tx_submit(). Clients of at_xdmac that assume that tx_submit() starts
the transfer must be updated and call dma_async_issue_pending() if they
miss to call it.
The vdbg print was moved to after the lock is released. It is desirable to
do the prints without the lock held if possible, and because the if
statement disappears there's no reason why to do the print while holding
the lock.
Those hardware registers are all of 32 bits, while dma_addr_t ca be of
type u64 or u32 depending on CONFIG_ARCH_DMA_ADDR_T_64BIT. Force u32 to
comply with what the hardware expects.
OpenSSL 3.0 deprecated the OpenSSL's ENGINE API. That is as may be, but
the kernel build host tools still use it. Disable the warning about
deprecated declarations until somebody who cares fixes it.
The read access to struct canxl_frame::len inside of a j1939 created
skbuff revealed a missing initialization of reserved and later filled
elements in struct can_frame.
This patch initializes the 8 byte CAN header with zero.
virtio_pmem use devm_memremap_pages() to map the device memory. By
default this memory is mapped as encrypted with SEV. Guest reboot changes
the current encryption key and guest no longer properly decrypts the FSDAX
device meta data.
Mark the corresponding device memory region for FSDAX devices (mapped with
memremap_pages) as decrypted to retain the persistent memory property.
Link: https://lkml.kernel.org/r/20221102160728.3184016-1-pankaj.gupta@amd.com Fixes: b7b3c01b19159 ("mm/memremap_pages: support multiple ranges per invocation") Signed-off-by: Pankaj Gupta <pankaj.gupta@amd.com> Cc: Dan Williams <dan.j.williams@intel.com> Cc: Tom Lendacky <thomas.lendacky@amd.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
loop0: detected capacity change from 0 to 2048
==================================================================
BUG: KASAN: slab-out-of-bounds in udf_find_entry+0x8a5/0x14f0
fs/udf/namei.c:253
Write of size 105 at addr ffff8880123ff896 by task syz-executor323/3610
The buggy address belongs to the object at ffff8880123ff800
which belongs to the cache kmalloc-256 of size 256
The buggy address is located 150 bytes inside of
256-byte region [ffff8880123ff800, ffff8880123ff900)
The buggy address belongs to the physical page:
page:ffffea000048ff80 refcount:1 mapcount:0 mapping:0000000000000000
index:0x0 pfn:0x123fe
head:ffffea000048ff80 order:1 compound_mapcount:0 compound_pincount:0
flags: 0xfff00000010200(slab|head|node=0|zone=1|lastcpupid=0x7ff)
raw: 00fff00000010200ffffea00004b8500dead000000000003ffff888012041b40
raw: 0000000000000000000000008010001000000001ffffffff0000000000000000
page dumped because: kasan: bad access detected
page_owner tracks the page as allocated
page last allocated via order 0, migratetype Unmovable, gfp_mask 0x0(),
pid 1, tgid 1 (swapper/0), ts 1841222404, free_ts 0
create_dummy_stack mm/page_owner.c:67 [inline]
register_early_stack+0x77/0xd0 mm/page_owner.c:83
init_page_owner+0x3a/0x731 mm/page_owner.c:93
kernel_init_freeable+0x41c/0x5d5 init/main.c:1629
kernel_init+0x19/0x2b0 init/main.c:1519
page_owner free stack trace missing
Memory state around the buggy address: ffff8880123ff780: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc ffff8880123ff800: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>ffff8880123ff880: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 06
^ ffff8880123ff900: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc ffff8880123ff980: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
==================================================================
Fix this by changing the memory size allocated for copy_name from
UDF_NAME_LEN(254) to UDF_NAME_LEN_CS0(255), because the total length
(lfi) of subsequent memcpy can be up to 255.
CC: stable@vger.kernel.org Reported-by: syzbot+69c9fdccc6dd08961d34@syzkaller.appspotmail.com Fixes: 066b9cded00b ("udf: Use separate buffer for copying split names") Signed-off-by: ZhangPeng <zhangpeng362@huawei.com> Signed-off-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20221109013542.442790-1-zhangpeng362@huawei.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[[ NOTE: this is completely untested by the author, but included solely
because, as noted in commit df57d73276b8 ("mmc: sdhci-pci: Fix
SDHCI_RESET_ALL for CQHCI for Intel GLK-based controllers"), "other
drivers using CQHCI might benefit from a similar change, if they
also have CQHCI reset by SDHCI_RESET_ALL." We've now seen the same
bug on at least MSM, Arasan, and Intel hardware. ]]
SDHCI_RESET_ALL resets will reset the hardware CQE state, but we aren't
tracking that properly in software. When out of sync, we may trigger
various timeouts.
It's not typical to perform resets while CQE is enabled, but this may
occur in some suspend or error recovery scenarios.
Include this fix by way of the new sdhci_and_cqhci_reset() helper.
This patch depends on (and should not compile without) the patch
entitled "mmc: cqhci: Provide helper for resetting both SDHCI and
CQHCI".
After upgrading BIOS to U82 01.02.01 Rev.A, the console is flooded
strange char "^@" which printed out every second and makes login
nearly impossible. Also the below messages were shown both in console
and journal/dmesg every second:
usb 1-3: Device not responding to setup address.
usb 1-3: device not accepting address 4, error -71
usb 1-3: device descriptor read/all, error -71
usb usb1-port3: unable to enumerate USB device
Wifi is soft blocked by checking rfkill. When unblocked manually,
after few seconds it would be soft blocked again. So I was suspecting
something triggered rfkill to soft block wifi. At the end it was
fixed by removing hp_wmi module.
The root cause is the way hp-wmi driver handles command 1B on
post-2009 BIOS. In pre-2009 BIOS, command 1Bh return 0x4 to indicate
that BIOS no longer controls the power for the wireless devices.
We need to iterate over the original entries here for the sg_table,
pulling out the struct page for each one, to be remapped. However
currently this incorrectly iterates over the final dma mapped entries,
which is likely just one gigantic sg entry if the iommu is enabled,
leading to us only mapping the first struct page (and any physically
contiguous pages following it), even if there is potentially lots more
data to follow.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/7306 Fixes: 1286ff739773 ("i915: add dmabuf/prime buffer sharing support.") Signed-off-by: Matthew Auld <matthew.auld@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Michael J. Ruhl <michael.j.ruhl@intel.com> Cc: <stable@vger.kernel.org> # v3.5+ Reviewed-by: Michael J. Ruhl <michael.j.ruhl@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20221028155029.494736-1-matthew.auld@intel.com
(cherry picked from commit 28d52f99bbca7227008cf580c9194c9b3516968e) Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If a nilfs2 filesystem is downgraded to read-only due to metadata
corruption on disk and is remounted read/write, or if emergency read-only
remount is performed, detaching a log writer and synchronizing the
filesystem can be done at the same time.
In these cases, use-after-free of the log writer (hereinafter
nilfs->ns_writer) can happen as shown in the scenario below:
While Task1 is sleeping, nilfs->ns_writer is freed by Task2. After Task1
waked up, Task1 accesses nilfs->ns_writer which is already freed. This
scenario diagram is based on the Shigeru Yoshida's post [1].
This patch fixes the issue by not detaching nilfs->ns_writer on remount so
that this UAF race doesn't happen. Along with this change, this patch
also inserts a few necessary read-only checks with superblock instance
where only the ns_writer pointer was used to check if the filesystem is
read-only.
A semaphore deadlock can occur if nilfs_get_block() detects metadata
corruption while locating data blocks and a superblock writeback occurs at
the same time:
task 1 task 2
------ ------
* A file operation *
nilfs_truncate()
nilfs_get_block()
down_read(rwsem A) <--
nilfs_bmap_lookup_contig()
... generic_shutdown_super()
nilfs_put_super()
* Prepare to write superblock *
down_write(rwsem B) <--
nilfs_cleanup_super()
* Detect b-tree corruption * nilfs_set_log_cursor()
nilfs_bmap_convert_error() nilfs_count_free_blocks()
__nilfs_error() down_read(rwsem A) <--
nilfs_set_error()
down_write(rwsem B) <--
*** DEADLOCK ***
Here, nilfs_get_block() readlocks rwsem A (= NILFS_MDT(dat_inode)->mi_sem)
and then calls nilfs_bmap_lookup_contig(), but if it fails due to metadata
corruption, __nilfs_error() is called from nilfs_bmap_convert_error()
inside the lock section.
Since __nilfs_error() calls nilfs_set_error() unless the filesystem is
read-only and nilfs_set_error() attempts to writelock rwsem B (=
nilfs->ns_sem) to write back superblock exclusively, hierarchical lock
acquisition occurs in the order rwsem A -> rwsem B.
Now, if another task starts updating the superblock, it may writelock
rwsem B during the lock sequence above, and can deadlock trying to
readlock rwsem A in nilfs_count_free_blocks().
However, there is actually no need to take rwsem A in
nilfs_count_free_blocks() because it, within the lock section, only reads
a single integer data on a shared struct with
nilfs_sufile_get_ncleansegs(). This has been the case after commit aa474a220180 ("nilfs2: add local variable to cache the number of clean
segments"), that is, even before this bug was introduced.
So, this resolves the deadlock problem by just not taking the semaphore in
nilfs_count_free_blocks().
SAT SCSI/ATA Translation specification requires SCSI SYNCHRONIZE CACHE
(10) and (16) commands both shall be translated to ATA flush command.
Also, ZBC Zoned Block Commands specification mandates SYNCHRONIZE CACHE
(16) command support. However, libata translates only SYNCHRONIZE CACHE
(10). This results in SYNCHRONIZE CACHE (16) command failures on SATA
drives and then libata translation does not conform to ZBC. To avoid the
failure, add support for SYNCHRONIZE CACHE (16).
Signed-off-by: Shin'ichiro Kawasaki <shinichiro.kawasaki@wdc.com> Cc: stable@vger.kernel.org Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit d4c639990036 ("vmlinux.lds.h: Avoid orphan section with !SMP")
fixed an orphan section warning by adding the '.data..decrypted' section
to the linker script under the PERCPU_DECRYPTED_SECTION define but that
placement introduced a panic with !SMP, as the percpu sections are not
instantiated with that configuration so attempting to access variables
defined with DEFINE_PER_CPU_DECRYPTED() will result in a page fault.
Move the '.data..decrypted' section to the DATA_MAIN define so that the
variables in it are properly instantiated at boot time with
CONFIG_SMP=n.
Accuphase DAC-60 option card supports native DSD up to DSD256,
but doesn't have support for auto-detection. Explicitly enable
DSD support for the correct altsetting.
M-Audio Micro (0762:201a) defines the descriptor as vendor-specific,
while the content seems class-compliant. Just overriding the probe
makes the device working.
As 'kobject_add' may allocated memory for 'kobject->name' when return error.
And in this function, if call 'kobject_add' failed didn't free kobject.
So call 'kobject_put' to recycling resources.
The Z390 DARK mainboard uses a CA0132 audio controller. The quirk is
needed to enable surround sound and 3.5mm headphone jack handling in
the front audio connector as well as in the rear of the board when in
stereo mode.
Page 97 of the linked manual contains instructions to setup the
controller.
[[ NOTE: this is completely untested by the author, but included solely
because, as noted in commit df57d73276b8 ("mmc: sdhci-pci: Fix
SDHCI_RESET_ALL for CQHCI for Intel GLK-based controllers"), "other
drivers using CQHCI might benefit from a similar change, if they
also have CQHCI reset by SDHCI_RESET_ALL." We've now seen the same
bug on at least MSM, Arasan, and Intel hardware. ]]
SDHCI_RESET_ALL resets will reset the hardware CQE state, but we aren't
tracking that properly in software. When out of sync, we may trigger
various timeouts.
It's not typical to perform resets while CQE is enabled, but this may
occur in some suspend or error recovery scenarios.
Include this fix by way of the new sdhci_and_cqhci_reset() helper.
This patch depends on (and should not compile without) the patch
entitled "mmc: cqhci: Provide helper for resetting both SDHCI and
CQHCI".
Fixes: 3c4019f97978 ("mmc: tegra: HW Command Queue Support for Tegra SDMMC") Signed-off-by: Brian Norris <briannorris@chromium.org> Acked-by: Adrian Hunter <adrian.hunter@intel.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20221026124150.v4.5.I418c9eaaf754880fcd2698113e8c3ef821a944d7@changeid Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[[ NOTE: this is completely untested by the author, but included solely
because, as noted in commit df57d73276b8 ("mmc: sdhci-pci: Fix
SDHCI_RESET_ALL for CQHCI for Intel GLK-based controllers"), "other
drivers using CQHCI might benefit from a similar change, if they
also have CQHCI reset by SDHCI_RESET_ALL." We've now seen the same
bug on at least MSM, Arasan, and Intel hardware. ]]
SDHCI_RESET_ALL resets will reset the hardware CQE state, but we aren't
tracking that properly in software. When out of sync, we may trigger
various timeouts.
It's not typical to perform resets while CQE is enabled, but this may
occur in some suspend or error recovery scenarios.
Include this fix by way of the new sdhci_and_cqhci_reset() helper.
This patch depends on (and should not compile without) the patch
entitled "mmc: cqhci: Provide helper for resetting both SDHCI and
CQHCI".
Fixes: f545702b74f9 ("mmc: sdhci_am654: Add Support for Command Queuing Engine to J721E") Signed-off-by: Brian Norris <briannorris@chromium.org> Acked-by: Adrian Hunter <adrian.hunter@intel.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20221026124150.v4.6.I35ca9d6220ba48304438b992a76647ca8e5b126f@changeid Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
SDHCI_RESET_ALL resets will reset the hardware CQE state, but we aren't
tracking that properly in software. When out of sync, we may trigger
various timeouts.
It's not typical to perform resets while CQE is enabled, but one
particular case I hit commonly enough: mmc_suspend() -> mmc_power_off().
Typically we will eventually deactivate CQE (cqhci_suspend() ->
cqhci_deactivate()), but that's not guaranteed -- in particular, if
we perform a partial (e.g., interrupted) system suspend.
The same bug was already found and fixed for two other drivers, in v5.7
and v5.9:
5cf583f1fb9c ("mmc: sdhci-msm: Deactivate CQE during SDHC reset") df57d73276b8 ("mmc: sdhci-pci: Fix SDHCI_RESET_ALL for CQHCI for Intel
GLK-based controllers")
The latter is especially prescient, saying "other drivers using CQHCI
might benefit from a similar change, if they also have CQHCI reset by
SDHCI_RESET_ALL."
So like these other patches, deactivate CQHCI when resetting the
controller. Do this via the new sdhci_and_cqhci_reset() helper.
This patch depends on (and should not compile without) the patch
entitled "mmc: cqhci: Provide helper for resetting both SDHCI and
CQHCI".
Fixes: 84362d79f436 ("mmc: sdhci-of-arasan: Add CQHCI support for arasan,sdhci-5.1") Cc: <stable@vger.kernel.org> Signed-off-by: Brian Norris <briannorris@chromium.org> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Acked-by: Adrian Hunter <adrian.hunter@intel.com> Link: https://lore.kernel.org/r/20221026124150.v4.2.I29f6a2189e84e35ad89c1833793dca9e36c64297@changeid Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Several SDHCI drivers need to deactivate command queueing in their reset
hook (see sdhci_cqhci_reset() / sdhci-pci-core.c, for example), and
several more are coming.
Those reset implementations have some small subtleties (e.g., ordering
of initialization of SDHCI vs. CQHCI might leave us resetting with a
NULL ->cqe_private), and are often identical across different host
drivers.
We also don't want to force a dependency between SDHCI and CQHCI, or
vice versa; non-SDHCI drivers use CQHCI, and SDHCI drivers might support
command queueing through some other means.
So, implement a small helper, to avoid repeating the same mistakes in
different drivers. Simply stick it in a header, because it's so small it
doesn't deserve its own module right now, and inlining to each driver is
pretty reasonable.
This is marked for -stable, as it is an important prerequisite patch for
several SDHCI controller bugfixes that follow.
Currently, when mapping the EFI runtime regions in the EFI page tables,
we complain about misaligned regions in a rather noisy way, using
WARN().
Not only does this produce a lot of irrelevant clutter in the log, it is
factually incorrect, as misaligned runtime regions are actually allowed
by the EFI spec as long as they don't require conflicting memory types
within the same 64k page.
So let's drop the warning, and tweak the code so that we
- take both the start and end of the region into account when checking
for misalignment
- only revert to RWX mappings for non-code regions if misaligned code
regions are also known to exist.
Currently, RISC-V sets up reserved memory using the "early" copy of the
device tree. As a result, when trying to get a reserved memory region
using of_reserved_mem_lookup(), the pointer to reserved memory regions
is using the early, pre-virtual-memory address which causes a kernel
panic when trying to use the buffer's name:
early_init_fdt_scan_reserved_mem() takes no arguments as it operates on
initial_boot_params, which is populated by early_init_dt_verify(). On
RISC-V, early_init_dt_verify() is called twice. Once, directly, in
setup_arch() if CONFIG_BUILTIN_DTB is not enabled and once indirectly,
very early in the boot process, by parse_dtb() when it calls
early_init_dt_scan_nodes().
This first call uses dtb_early_va to set initial_boot_params, which is
not usable later in the boot process when
early_init_fdt_scan_reserved_mem() is called. On arm64 for example, the
corresponding call to early_init_dt_scan_nodes() uses fixmap addresses
and doesn't suffer the same fate.
Move early_init_fdt_scan_reserved_mem() further along the boot sequence,
after the direct call to early_init_dt_verify() in setup_arch() so that
the names use the correct virtual memory addresses. The above supposed
that CONFIG_BUILTIN_DTB was not set, but should work equally in the case
where it is - unflatted_and_copy_device_tree() also updates
initial_boot_params.
Currently, we perform some memory init functions in paging init. But,
that will be an issue for NUMA support where DT needs to be flattened
before numa initialization and memblock_present can only be called
after numa initialization.
Move memory initialization related functions to a separate function.
Even after commit 89fd4a1df829 ("riscv: jump_label: mark arguments as
const to satisfy asm constraints"), building with CC_OPTIMIZE_FOR_SIZE
+ LLVM=1 can reproduce below build error:
CC arch/riscv/kernel/vdso/vgettimeofday.o
In file included from <built-in>:4:
In file included from lib/vdso/gettimeofday.c:5:
In file included from include/vdso/datapage.h:17:
In file included from include/vdso/processor.h:10:
In file included from arch/riscv/include/asm/vdso/processor.h:7:
In file included from include/linux/jump_label.h:112:
arch/riscv/include/asm/jump_label.h:42:3: error:
invalid operand for inline asm constraint 'i'
" .option push \n\t"
^
1 error generated.
I think the problem is when "-Os" is passed as CFLAGS, it's removed by
"CFLAGS_REMOVE_vgettimeofday.o = $(CC_FLAGS_FTRACE) -Os" which is
introduced in commit e05d57dcb8c7 ("riscv: Fixup __vdso_gettimeofday
broke dynamic ftrace"), thus no optimization at all for vgettimeofday.c
arm64 does remove "-Os" as well, but it forces "-O2" after removing
"-Os".
I compared the generated vgettimeofday.o with "-O2" and "-Os",
I think no big performance difference. So let's tell the kbuild not
to remove "-Os" rather than follow arm64 style.
vdso related performance can be improved a lot when building kernel with
CC_OPTIMIZE_FOR_SIZE after this commit, ("-Os" VS no optimization)
thread_struct's s[12] may contain random kernel memory content, which
may be finally leaked to userspace. This is a security hole. Fix it
by clearing the s[12] array in thread_struct when fork.
As for kthread case, it's better to clear the s[12] array as well.
In the scenario where the macvlan mode is configured as 'source',
macvlan_changelink_sources() will be execured to reconfigure list of
remote source mac addresses, at the same time, if register_netdevice()
return an error, the resource generated by macvlan_changelink_sources()
is not cleaned up.
Using this patch, in the case of an error, it will execute
macvlan_flush_sources() to ensure that the resource is cleaned up.
Fixes: aa5fd0fb7748 ("driver: macvlan: Destroy new macvlan port if macvlan_common_newlink failed.") Signed-off-by: Chuang Wang <nashuiliang@gmail.com> Link: https://lore.kernel.org/r/20221109090735.690500-1-nashuiliang@gmail.com Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When failed to init rxq or txq in mv643xx_eth_open() for opening device,
napi isn't disabled. When open mv643xx_eth device next time, it will
trigger a BUG_ON() in napi_enable(). Compile tested only.
Fixes: 2257e05c1705 ("mv643xx_eth: get rid of receive-side locking") Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com> Link: https://lore.kernel.org/r/20221109025432.80900-1-shaozhengchao@huawei.com Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When failed to start nic or add interrupt service routine in
s2io_card_up() for opening device, napi isn't disabled. When open
s2io device next time, it will trigger a BUG_ON()in napi_enable().
Compile tested only.
Fixes: 5f490c968056 ("S2io: Fixed synchronization between scheduling of napi with card reset and close") Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com> Link: https://lore.kernel.org/r/20221109023741.131552-1-shaozhengchao@huawei.com Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit aaab73f8fba4 ("macsec: clear encryption keys from the stack after
setting up offload") made sure to clean encryption keys from the stack
after setting up offloading, but the atlantic driver made a copy and did
not clear it. Fix this.
[4 Fixes tags below, all part of the same series, no need to split this]
Commit aaab73f8fba4 ("macsec: clear encryption keys from the stack after
setting up offload") made sure to clean encryption keys from the stack
after setting up offloading, but the MSCC PHY driver made a copy, kept
it in the flow data and did not clear it when freeing a flow. Fix this.
Fixes: 28c5107aa904 ("net: phy: mscc: macsec support") Signed-off-by: Antoine Tenart <atenart@kernel.org> Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When t4vf_update_port_info() failed in cxgb4vf_open(), resources applied
during adapter goes up are not cleared. Fix it. Only be compiled, not be
tested.
Fixes: 18d79f721e0a ("cxgb4vf: Update port information in cxgb4vf_open()") Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com> Link: https://lore.kernel.org/r/20221109012100.99132-1-shaozhengchao@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When failed to bind qsets in cxgb_up() for opening device, napi isn't
disabled. When open cxgb3 device next time, it will trigger a BUG_ON()
in napi_enable(). Compile tested only.
When failed to create xdp rxqs or fill rx channels in cpsw_ndo_open() for
opening device, napi isn't disabled. When open cpsw device next time, it
will report a invalid opcode issue. Compiled tested only.
The pkt_reformat pointer being saved under flow_act and not
dest attribute in the termination table instance.
Fix the comparison pointers.
Also fix returning success if one pkt_reformat pointer is null
and the other is not.
Fixes: 249ccc3c95bd ("net/mlx5e: Add support for offloading traffic from uplink to uplink") Signed-off-by: Roi Dayan <roid@nvidia.com> Reviewed-by: Chris Mi <cmi@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
For a single CPU system, the kernel thread executing mlx5_cmd_flush()
never releases the CPU but calls down_trylock(&cmd→sem) in a busy loop.
On a single processor system, this leads to a deadlock as the kernel
thread which executes mlx5_cmd_invoke() never gets scheduled. Fix this,
by adding the cond_resched() call to the loop, allow the command
completion kernel thread to execute.
Fixes: 8e715cd613a1 ("net/mlx5: Set command entry semaphore up once got index free") Signed-off-by: Alexander Schmidt <alexschm@de.ibm.com> Signed-off-by: Roy Novich <royno@nvidia.com> Reviewed-by: Moshe Shemesh <moshe@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When failed to enable interrupts in nixge_open() for opening device,
napi isn't disabled. When open nixge device next time, it will reports
a invalid opcode issue. Fix it. Only be compiled, not be tested.
Fixes: 492caffa8a1a ("net: ethernet: nixge: Add support for National Instruments XGE netdev") Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com> Link: https://lore.kernel.org/r/20221107101443.120205-1-shaozhengchao@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
'perf stat' with CSV output option prints an extra empty string as first
field in metrics output line. Sample output below:
# ./perf stat -x, --per-socket -a -C 1 ls
S0,1,1.78,msec,cpu-clock,1785146,100.00,0.973,CPUs utilized
S0,1,26,,context-switches,1781750,100.00,0.015,M/sec
S0,1,1,,cpu-migrations,1780526,100.00,0.561,K/sec
S0,1,1,,page-faults,1779060,100.00,0.561,K/sec
S0,1,875807,,cycles,1769826,100.00,0.491,GHz
S0,1,85281,,stalled-cycles-frontend,1767512,100.00,9.74,frontend cycles idle
S0,1,576839,,stalled-cycles-backend,1766260,100.00,65.86,backend cycles idle
S0,1,288430,,instructions,1762246,100.00,0.33,insn per cycle
====> ,S0,1,,,,,,,2.00,stalled cycles per insn
The above command line uses field separator as "," via "-x," option and
per-socket option displays socket value as first field. But here the
last line for "stalled cycles per insn" has "," in the beginning.
Sample output using interval mode:
# ./perf stat -I 1000 -x, --per-socket -a -C 1 ls
0.001813453,S0,1,1.87,msec,cpu-clock,1872052,100.00,0.002,CPUs utilized
0.001813453,S0,1,2,,context-switches,1868028,100.00,1.070,K/sec
------
0.001813453,S0,1,85379,,instructions,1856754,100.00,0.32,insn per cycle
====> 0.001813453,,S0,1,,,,,,,1.34,stalled cycles per insn
Above result also has an extra CSV separator after
the timestamp. Patch addresses extra field separator
in the beginning of the metric output line.
The counter stats are displayed by function
"perf_stat__print_shadow_stats" in code
"util/stat-shadow.c". While printing the stats info
for "stalled cycles per insn", function "new_line_csv"
is used as new_line callback.
The new_line_csv function has check for "os->prefix"
and if prefix is not null, it will be printed along
with cvs separator.
Snippet from "new_line_csv":
if (os->prefix)
fprintf(os->fh, "%s%s", os->prefix, config->csv_sep);
Here os->prefix gets printed followed by ","
which is the cvs separator. The os->prefix is
used in interval mode option ( -I ), to print
time stamp on every new line. But prefix is
already set to contain CSV separator when used
in interval mode for CSV option.
Also if prefix is not assigned (if not used with
-I option), it gets set to empty string.
Reference: function printout() in util/stat-display.c
Snippet:
.prefix = prefix ? prefix : "",
Since prefix already set to contain cvs_sep in interval
option, patch removes printing config->csv_sep in
new_line_csv function to avoid printing extra field.
After the patch:
# ./perf stat -x, --per-socket -a -C 1 ls
S0,1,2.04,msec,cpu-clock,2045202,100.00,1.013,CPUs utilized
S0,1,2,,context-switches,2041444,100.00,979.289,/sec
S0,1,0,,cpu-migrations,2040820,100.00,0.000,/sec
S0,1,2,,page-faults,2040288,100.00,979.289,/sec
S0,1,254589,,cycles,2036066,100.00,0.125,GHz
S0,1,82481,,stalled-cycles-frontend,2032420,100.00,32.40,frontend cycles idle
S0,1,113170,,stalled-cycles-backend,2031722,100.00,44.45,backend cycles idle
S0,1,88766,,instructions,2030942,100.00,0.35,insn per cycle
S0,1,,,,,,,1.27,stalled cycles per insn
Fixes: 92a61f6412d3a09d ("perf stat: Implement CSV metrics output") Reported-by: Disha Goel <disgoel@linux.vnet.ibm.com> Reviewed-By: Kajol Jain <kjain@linux.ibm.com> Signed-off-by: Athira Jajeev <atrajeev@linux.vnet.ibm.com> Tested-by: Disha Goel <disgoel@linux.vnet.ibm.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Ian Rogers <irogers@google.com> Cc: James Clark <james.clark@arm.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: linuxppc-dev@lists.ozlabs.org Cc: Madhavan Srinivasan <maddy@linux.vnet.ibm.com> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Nageswara R Sastry <rnsastry@linux.ibm.com> Cc: Namhyung Kim <namhyung@kernel.org> Link: https://lore.kernel.org/r/20221018085605.63834-1-atrajeev@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When failed to register irq in xgene_enet_open() for opening device,
napi isn't disabled. When open xgene device next time, it will reports
a invalid opcode issue. Fix it. Only be compiled, not be tested.
The first IRQ is required, but IRQs 1 through (nb_phy_chans - 1) are
optional, because on some platforms (e.g. PXA168) there is a single IRQ
shared between all channels.
This change inhibits a flood of "IRQ index # not found" messages at
startup. Tested on a PXA168-based device.
Fixes: 7723f4c5ecdb ("driver core: platform: Add an error message to platform_get_irq*()") Signed-off-by: Doug Brown <doug@schmorgal.com> Link: https://lore.kernel.org/r/20220906000709.52705-1-doug@schmorgal.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
This is a follow-up for commit 974cb0e3e7c9 ("tipc: fix uninit-value
in tipc_nl_compat_name_table_dump") where it should have type casted
sizeof(..) to int to work when TLV_GET_DATA_LEN() returns a negative
value.
While BCMGENET select BROADCOM_PHY as y, but PTP_1588_CLOCK_OPTIONAL is m,
kconfig warning and build errors:
WARNING: unmet direct dependencies detected for BROADCOM_PHY
Depends on [m]: NETDEVICES [=y] && PHYLIB [=y] && PTP_1588_CLOCK_OPTIONAL [=m]
Selected by [y]:
- BCMGENET [=y] && NETDEVICES [=y] && ETHERNET [=y] && NET_VENDOR_BROADCOM [=y] && HAS_IOMEM [=y] && ARCH_BCM2835 [=y]
drivers/net/phy/broadcom.o: In function `bcm54xx_suspend':
broadcom.c:(.text+0x6ac): undefined reference to `bcm_ptp_stop'
drivers/net/phy/broadcom.o: In function `bcm54xx_phy_probe':
broadcom.c:(.text+0x784): undefined reference to `bcm_ptp_probe'
drivers/net/phy/broadcom.o: In function `bcm54xx_config_init':
broadcom.c:(.text+0xd4c): undefined reference to `bcm_ptp_config_init'
There are two problems with meson8b_devm_clk_prepare_enable(),
introduced in commit a54dc4a49045 ("net: stmmac: dwmac-meson8b:
Make the clock enabling code re-usable"):
- It doesn't pass the clk argument, but instead always the
rgmii_tx_clk of the device.
- It silently ignores the return value of devm_add_action_or_reset().
The former didn't become an actual bug until another user showed up in
the next commit 9308c47640d5 ("net: stmmac: dwmac-meson8b: add support
for the RX delay configuration"). The latter means the callers could
end up with the clock not actually prepared/enabled.
Fixes: a54dc4a49045 ("net: stmmac: dwmac-meson8b: Make the clock enabling code re-usable") Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Link: https://lore.kernel.org/r/20221104083004.2212520-1-linux@rasmusvillemoes.dk Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
It causes NULL pointer dereference when testing as following:
(a) use syscall(__NR_socket, 0x10ul, 3ul, 0) to create netlink socket.
(b) use syscall(__NR_sendmsg, ...) to create bond link device and vxcan
link device, and bind vxcan device to bond device (can also use
ifenslave command to bind vxcan device to bond device).
(c) use syscall(__NR_socket, 0x1dul, 3ul, 1) to create CAN socket.
(d) use syscall(__NR_bind, ...) to bind the bond device to CAN socket.
The bond device invokes the can-raw protocol registration interface to
receive CAN packets. However, ml_priv is not allocated to the dev,
dev_rcv_lists is assigned to NULL in can_rx_register(). In this case,
it will occur the NULL pointer dereference issue.
The following is the stack information:
BUG: kernel NULL pointer dereference, address: 0000000000000008
PGD 122a4067 P4D 122a4067 PUD 1223c067 PMD 0
Oops: 0000 [#1] PREEMPT SMP
RIP: 0010:can_rx_register+0x12d/0x1e0
Call Trace:
<TASK>
raw_enable_filters+0x8d/0x120
raw_enable_allfilters+0x3b/0x130
raw_bind+0x118/0x4f0
__sys_bind+0x163/0x1a0
__x64_sys_bind+0x1e/0x30
do_syscall_64+0x35/0x80
entry_SYSCALL_64_after_hwframe+0x63/0xcd
</TASK>
Fixes: 4e096a18867a ("net: introduce CAN specific pointer in the struct net_device") Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com> Reviewed-by: Marc Kleine-Budde <mkl@pengutronix.de> Link: https://lore.kernel.org/all/20221028085650.170470-1-shaozhengchao@huawei.com Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
This patch ensures that the reserved field is always initialized.
Reported-by: syzbot+3553517af6020c4f2813f1003fe76ef3cbffe98d@syzkaller.appspotmail.com Fixes: 2a8cc6c89039 ("[IPV6] ADDRCONF: Support RFC3484 configurable address selection policy table.") Signed-off-by: Alexander Potapenko <glider@google.com> Reviewed-by: David Ahern <dsahern@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
If setsockopt with option name of TCP_REPAIR_OPTIONS and opt_code
of TCPOPT_SACK_PERM is called to enable sack after data is sent
and dupacks are received , it will trigger a warning in function
tcp_verify_left_out() as follows:
The warning is caused in the following steps:
1. a socket named socketA is created
2. socketA enters repair mode without build a connection
3. socketA calls connect() and its state is changed to TCP_ESTABLISHED
directly
4. socketA leaves repair mode
5. socketA calls sendmsg() to send data, packets_out and sack_outs(dup
ack receives) increase
6. socketA enters repair mode again
7. socketA calls setsockopt with TCPOPT_SACK_PERM to enable sack
8. retransmit timer expires, it calls tcp_timeout_mark_lost(), lost_out
increases
9. sack_outs + lost_out > packets_out triggers since lost_out and
sack_outs increase repeatly
In function tcp_timeout_mark_lost(), tp->sacked_out will be cleared if
Step7 not happen and the warning will not be triggered. As suggested by
Denis and Eric, TCP_REPAIR_OPTIONS should be prohibited if data was
already sent.
socket-tcp tests in CRIU has been tested as follows:
$ sudo ./test/zdtm.py run -t zdtm/static/socket-tcp* --keep-going \
--ignore-taint
socket-tcp* represent all socket-tcp tests in test/zdtm/static/.
Fixes: b139ba4e90dc ("tcp: Repair connection-time negotiated parameters") Signed-off-by: Lu Wei <luwei32@huawei.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
A problem about modprobe vc4 failed is triggered with the following log
given:
[ 420.327987] Error: Driver 'vc4_hvs' is already registered, aborting...
[ 420.333904] failed to register platform driver vc4_hvs_driver [vc4]: -16
modprobe: ERROR: could not insert 'vc4': Device or resource busy
The reason is that vc4_drm_register() returns platform_driver_register()
directly without checking its return value, if platform_driver_register()
fails, it returns without unregistering all the vc4 drivers, resulting the
vc4 can never be installed later.
A simple call graph is shown as below:
vc4_drm_register()
platform_register_drivers() # all vc4 drivers are registered
platform_driver_register()
driver_register()
bus_add_driver()
priv = kzalloc(...) # OOM happened
# return without unregister drivers
Fixing this problem by checking the return value of
platform_driver_register() and do platform_unregister_drivers() if
error happened.
When following tests are performed, it will cause dev reference counting
leakage.
a)ip link add bond2 type bond mode balance-rr
b)ip link set bond2 up
c)ifenslave -f bond2 rose1
d)ip link del bond2
When new bond device is created, the default type of the bond device is
ether. And the bond device is up, bpq_device_event() receives the message
and creates a new bpq device. In this case, the reference count value of
dev is hold once. But after "ifenslave -f bond2 rose1" command is
executed, the type of the bond device is changed to rose. When the bond
device is unregistered, bpq_device_event() will not put the dev reference
count.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
When following tests are performed, it will cause dev reference counting
leakage.
a)ip link add bond2 type bond mode balance-rr
b)ip link set bond2 up
c)ifenslave -f bond2 rose1
d)ip link del bond2
When new bond device is created, the default type of the bond device is
ether. And the bond device is up, lapbeth_device_event() receives the
message and creates a new lapbeth device. In this case, the reference
count value of dev is hold once. But after "ifenslave -f bond2 rose1"
command is executed, the type of the bond device is changed to rose. When
the bond device is unregistered, lapbeth_device_event() will not put the
dev reference count.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
When running under PV, the guest's TOD clock is under control of the
ultravisor and the hypervisor isn't allowed to change it. Hence, don't
allow userspace to change the guest's TOD clock by returning
-EOPNOTSUPP.
When userspace changes the guest's TOD clock, KVM updates its
kvm.arch.epoch field and, in addition, the epoch field in all state
descriptions of all VCPUs.
But, under PV, the ultravisor will ignore the epoch field in the state
description and simply overwrite it on next SIE exit with the actual
guest epoch. This leads to KVM having an incorrect view of the guest's
TOD clock: it has updated its internal kvm.arch.epoch field, but the
ultravisor ignores the field in the state description.
Whenever a guest is now waiting for a clock comparator, KVM will
incorrectly calculate the time when the guest should wake up, possibly
causing the guest to sleep for much longer than expected.
With this change, kvm_s390_set_tod() will now take the kvm->lock to be
able to call kvm_s390_pv_is_protected(). Since kvm_s390_set_tod_clock()
also takes kvm->lock, use __kvm_s390_set_tod_clock() instead.
The function kvm_s390_set_tod_clock is now unused, hence remove it.
Update the documentation to indicate the TOD clock attr calls can now
return -EOPNOTSUPP.
Fixes: 0f3035047140 ("KVM: s390: protvirt: Do only reset registers that are accessible") Reported-by: Marc Hartmayer <mhartmay@linux.ibm.com> Signed-off-by: Nico Boehr <nrb@linux.ibm.com> Reviewed-by: Claudio Imbrenda <imbrenda@linux.ibm.com> Reviewed-by: Janosch Frank <frankja@linux.ibm.com> Link: https://lore.kernel.org/r/20221011160712.928239-2-nrb@linux.ibm.com
Message-Id: <20221011160712.928239-2-nrb@linux.ibm.com> Signed-off-by: Janosch Frank <frankja@linux.ibm.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When handling the SCK instruction, the kvm lock is taken, even though
the vcpu lock is already being held. The normal locking order is kvm
lock first and then vcpu lock. This is can (and in some circumstances
does) lead to deadlocks.
The function kvm_s390_set_tod_clock is called both by the SCK handler
and by some IOCTLs to set the clock. The IOCTLs will not hold the vcpu
lock, so they can safely take the kvm lock. The SCK handler holds the
vcpu lock, but will also somehow need to acquire the kvm lock without
relinquishing the vcpu lock.
The solution is to factor out the code to set the clock, and provide
two wrappers. One is called like the original function and does the
locking, the other is called kvm_s390_try_set_tod_clock and uses
trylock to try to acquire the kvm lock. This new wrapper is then used
in the SCK handler. If locking fails, -EAGAIN is returned, which is
eventually propagated to userspace, thus also freeing the vcpu lock and
allowing for forward progress.
This is not the most efficient or elegant way to solve this issue, but
the SCK instruction is deprecated and its performance is not critical.
The goal of this patch is just to provide a simple but correct way to
fix the bug.
Fixes: 6a3f95a6b04c ("KVM: s390: Intercept SCK instruction") Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com> Reviewed-by: Christian Borntraeger <borntraeger@linux.ibm.com> Reviewed-by: Janis Schoetterl-Glausch <scgl@linux.ibm.com> Link: https://lore.kernel.org/r/20220301143340.111129-1-imbrenda@linux.ibm.com Cc: stable@vger.kernel.org Signed-off-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Stable-dep-of: 6973091d1b50 ("KVM: s390: pv: don't allow userspace to set the clock under PV") Signed-off-by: Sasha Levin <sashal@kernel.org>
Shifting signed 32-bit value by 31 bits is undefined, so changing
significant bit to unsigned. The UBSAN warning calltrace like below:
UBSAN: shift-out-of-bounds in security/commoncap.c:1252:2
left shift of 1 by 31 places cannot be represented in type 'int'
Call Trace:
<TASK>
dump_stack_lvl+0x7d/0xa5
dump_stack+0x15/0x1b
ubsan_epilogue+0xe/0x4e
__ubsan_handle_shift_out_of_bounds+0x1e7/0x20c
cap_task_prctl+0x561/0x6f0
security_task_prctl+0x5a/0xb0
__x64_sys_prctl+0x61/0x8f0
do_syscall_64+0x58/0x80
entry_SYSCALL_64_after_hwframe+0x63/0xcd
</TASK>
Fixes: e338d263a76a ("Add 64-bit capability support to the kernel") Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com> Acked-by: Andrew G. Morgan <morgan@kernel.org> Reviewed-by: Serge Hallyn <serge@hallyn.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When the mac device gets removed, it leaves behind the ethernet device.
This will result in a segfault next time the ethernet device accesses
mac_dev. Remove the ethernet device when we get removed to prevent
this. This is not completely reversible, since some resources aren't
cleaned up properly, but that can be addressed later.
In the bnxt_en driver ndo_rx_flow_steer returns '0' whenever an entry
that we are attempting to steer is already found. This is not the
correct behavior. The return code should be the value/index that
corresponds to the entry. Returning zero all the time causes the
RFS records to be incorrect unless entry '0' is the correct one. As
flows migrate to different cores this can create entries that are not
correct.
Fixes: c0c050c58d84 ("bnxt_en: New Broadcom ethernet driver.") Reported-by: Akshay Navgire <anavgire@purestorage.com> Signed-off-by: Alex Barba <alex.barba@broadcom.com> Signed-off-by: Andy Gospodarek <gospo@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
During the error recovery sequence, the rtnl_lock is not held for the
entire duration and some datastructures may be freed during the sequence.
Check for the BNXT_STATE_OPEN flag instead of netif_running() to ensure
that the device is fully operational before proceeding to reconfigure
the coalescing settings.
The issue occurs in the following scenarios:
tun_get_user()
napi_gro_frags()
napi_frags_finish()
case GRO_NORMAL:
gro_normal_one()
list_add_tail(&skb->list, &napi->rx_list);
<-- While napi->rx_count < READ_ONCE(gro_normal_batch),
<-- gro_normal_list() is not called, napi->rx_list is not empty
<-- not ask to complete the gro work, will cause memory leaks in
<-- following tun_napi_del()
...
tun_napi_del()
netif_napi_del()
__netif_napi_del()
<-- &napi->rx_list is not empty, which caused memory leaks
To fix, add napi_complete() after napi_gro_frags().
Fixes: 90e33d459407 ("tun: enable napi_gro_frags() for TUN/TAP driver") Signed-off-by: Wang Yufen <wangyufen@huawei.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
macsec_add_rxsa and macsec_add_txsa copy the key to an on-stack
offloading context to pass it to the drivers, but leaves it there when
it's done. Clear it with memzero_explicit as soon as it's not needed
anymore.
Fixes: 3cf3227a21d1 ("net: macsec: hardware offloading infrastructure") Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Reviewed-by: Antoine Tenart <atenart@kernel.org> Reviewed-by: Leon Romanovsky <leonro@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
macsec_is_configured incorrectly uses secy->n_rx_sc to check if some
RXSCs exist. secy->n_rx_sc only counts the number of active RXSCs, but
there can also be inactive SCs as well, which may be stored in the
driver (in case we're disabling offloading), or would have to be
pushed to the device (in case we're trying to enable offloading).
As long as RXSCs active on creation and never turned off, the issue is
not visible.
Fixes: dcb780fb2795 ("net: macsec: add nla support for changing the offloading selection") Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Reviewed-by: Antoine Tenart <atenart@kernel.org> Reviewed-by: Leon Romanovsky <leonro@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
secy->n_rx_sc is supposed to be the number of _active_ rxsc's within a
secy. This is then used by macsec_send_sci to help decide if we should
add the SCI to the header or not.
This logic is currently broken when we create a new RXSC and turn it
off at creation, as create_rx_sc always sets ->active to true (and
immediately uses that to increment n_rx_sc), and only later
macsec_add_rxsc sets rx_sc->active.
Fixes: c09440f7dcb3 ("macsec: introduce IEEE 802.1AE driver") Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Reviewed-by: Antoine Tenart <atenart@kernel.org> Reviewed-by: Leon Romanovsky <leonro@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
Since commit 3dcbdb134f32 ("net: gso: Fix skb_segment splat when
splitting gso_size mangled skb having linear-headed frag_list"), it is
allowed to change gso_size of a GRO packet. However, that commit assumes
that "checking the first list_skb member suffices; i.e if either of the
list_skb members have non head_frag head, then the first one has too".
It turns out this assumption does not hold. We've seen BUG_ON being hit
in skb_segment when skbs on the frag_list had differing head_frag with
the vmxnet3 driver. This happens because __netdev_alloc_skb and
__napi_alloc_skb can return a skb that is page backed or kmalloced
depending on the requested size. As the result, the last small skb in
the GRO packet can be kmalloced.
There are three different locations where this can be fixed:
(1) We could check head_frag in GRO and not allow GROing skbs with
different head_frag. However, that would lead to performance
regression on normal forward paths with unmodified gso_size, where
!head_frag in the last packet is not a problem.
(2) Set a flag in bpf_skb_net_grow and bpf_skb_net_shrink indicating
that NETIF_F_SG is undesirable. That would need to eat a bit in
sk_buff. Furthermore, that flag can be unset when all skbs on the
frag_list are page backed. To retain good performance,
bpf_skb_net_grow/shrink would have to walk the frag_list.
(3) Walk the frag_list in skb_segment when determining whether
NETIF_F_SG should be cleared. This of course slows things down.
This patch implements (3). To limit the performance impact in
skb_segment, the list is walked only for skbs with SKB_GSO_DODGY set
that have gso_size changed. Normal paths thus will not hit it.
We could check only the last skb but since we need to walk the whole
list anyway, let's stay on the safe side.
Fixes: 3dcbdb134f32 ("net: gso: Fix skb_segment splat when splitting gso_size mangled skb having linear-headed frag_list") Signed-off-by: Jiri Benc <jbenc@redhat.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Link: https://lore.kernel.org/r/e04426a6a91baf4d1081e1b478c82b5de25fdf21.1667407944.git.jbenc@redhat.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Some helper functions will allocate memory. To avoid memory leaks, the
verifier requires the eBPF program to release these memories by calling
the corresponding helper functions.
When a resource is released, all pointer registers corresponding to the
resource should be invalidated. The verifier use release_references() to
do this job, by apply __mark_reg_unknown() to each relevant register.
It will give these registers the type of SCALAR_VALUE. A register that
will contain a pointer value at runtime, but of type SCALAR_VALUE, which
may allow the unprivileged user to get a kernel pointer by storing this
register into a map.
Using __mark_reg_not_init() while NOT allow_ptr_leaks can mitigate this
problem.
Fixes: fd978bf7fd31 ("bpf: Add reference tracking to verifier") Signed-off-by: Youlin Li <liulin063@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Link: https://lore.kernel.org/bpf/20221103093440.3161-1-liulin063@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org>
For a lot of use cases in future patches, we will want to modify the
state of registers part of some same 'group' (e.g. same ref_obj_id). It
won't just be limited to releasing reference state, but setting a type
flag dynamically based on certain actions, etc.
Hence, we need a way to easily pass a callback to the function that
iterates over all registers in current bpf_verifier_state in all frames
upto (and including) the curframe.
While in C++ we would be able to easily use a lambda to pass state and
the callback together, sadly we aren't using C++ in the kernel. The next
best thing to avoid defining a function for each case seems like
statement expressions in GNU C. The kernel already uses them heavily,
hence they can passed to the macro in the style of a lambda. The
statement expression will then be substituted in the for loop bodies.
Variables __state and __reg are set to current bpf_func_state and reg
for each invocation of the expression inside the passed in verifier
state.
Then, convert mark_ptr_or_null_regs, clear_all_pkt_pointers,
release_reference, find_good_pkt_pointers, find_equal_scalars to
use bpf_for_each_reg_in_vstate.
This patch adds the verifier support to recognize inlined branch conditions.
The LLVM knows that the branch evaluates to the same value, but the verifier
couldn't track it. Hence causing valid programs to be rejected.
The potential LLVM workaround: https://reviews.llvm.org/D87428
can have undesired side effects, since LLVM doesn't know that
skb->data/data_end are being compared. LLVM has to introduce extra boolean
variable and use inline_asm trick to force easier for the verifier assembly.
Instead teach the verifier to recognize that
r1 = skb->data;
r1 += 10;
r2 = skb->data_end;
if (r1 > r2) {
here r1 points beyond packet_end and
subsequent
if (r1 > r2) // always evaluates to "true".
}
Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Tested-by: Jiri Olsa <jolsa@redhat.com> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20201111031213.25109-2-alexei.starovoitov@gmail.com
Stable-dep-of: f1db20814af5 ("bpf: Fix wrong reg type conversion in release_reference()") Signed-off-by: Sasha Levin <sashal@kernel.org>
When using bpftool to pin {PROG, MAP, LINK} without FILE,
segmentation fault will occur. The reson is that the lack
of FILE will cause strlen to trigger NULL pointer dereference.
The corresponding stacktrace is shown below:
The root case is in commit 84472b436e76 ("bpf, sockmap: Fix more uncharged
while msg has more_data"), where I used msg->sg.size to replace the tosend,
causing breakage:
if (msg->apply_bytes && msg->apply_bytes < tosend)
tosend = psock->apply_bytes;
Fixes: 84472b436e76 ("bpf, sockmap: Fix more uncharged while msg has more_data") Reported-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Wang Yufen <wangyufen@huawei.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Acked-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/1667266296-8794-1-git-send-email-wangyufen@huawei.com Signed-off-by: Sasha Levin <sashal@kernel.org>
In the function query_regdb_file() the alpha2 parameter is duplicated
using kmemdup() and subsequently freed in regdb_fw_cb(). However,
request_firmware_nowait() can fail without calling regdb_fw_cb() and
thus leak memory.
Fixes: 007f6c5e6eb4 ("cfg80211: support loading regulatory database as firmware file") Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
All we're going to do with this pointer is assign it to
another __rcu pointer, but sparse can't see that, so
use rcu_access_pointer() to silence the warning here.
Fixes: c90b93b5b782 ("wifi: cfg80211: update hidden BSSes to avoid WARN_ON") Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
There's a race in fuse's readdir cache that can result in an uninitilized
page being read. The page lock is supposed to prevent this from happening
but in the following case it doesn't:
Two fuse_add_dirent_to_cache() start out and get the same parameters
(size=0,offset=0). One of them wins the race to create and lock the page,
after which it fills in data, sets rdc.size and unlocks the page.
In the meantime the page gets evicted from the cache before the other
instance gets to run. That one also creates the page, but finds the
size to be mismatched, bails out and leaves the uninitialized page in the
cache.
Fix by marking a filled page uptodate and ignoring non-uptodate pages.
Linus proposes to revert an accounting for sops objects in
do_semtimedop() because it's really just a temporary buffer
for a single semtimedop() system call.
This object can consume up to 2 pages, syscall is sleeping
one, size and duration can be controlled by user, and this
allocation can be repeated by many thread at the same time.
However Shakeel Butt pointed that there are much more popular
objects with the same life time and similar memory
consumption, the accounting of which was decided to be
rejected for performance reasons.
Considering at least 2 pages for task_struct and 2 pages for
the kernel stack, a back of the envelope calculation gives a
footprint amplification of <1.5 so this temporal buffer can be
safely ignored.
The factor would IMO be interesting if it was >> 2 (from the
PoV of excessive (ab)use, fine-grained accounting seems to be
currently unfeasible due to performance impact).