[ Upstream commit 1ef94169db ]
[TEST FAILURE WITH EXPERIMENTAL FEATURES]
When running test case generic/508, the test case will fail with the new
btrfs shutdown support:
generic/508 - output mismatch (see /home/adam/xfstests/results//generic/508.out.bad)
# --- tests/generic/508.out 2022-05-11 11:25:30.806666664 +0930
# +++ /home/adam/xfstests/results//generic/508.out.bad 2025-07-02 14:53:22.401824212 +0930
# @@ -1,2 +1,6 @@
# QA output created by 508
# Silence is golden
# +Before:
# +After : stat.btime = Thu Jan 1 09:30:00 1970
# +Before:
# +After : stat.btime = Wed Jul 2 14:53:22 2025
# ...
# (Run 'diff -u /home/adam/xfstests/tests/generic/508.out /home/adam/xfstests/results//generic/508.out.bad' to see the entire diff)
Ran: generic/508
Failures: generic/508
Failed 1 of 1 tests
Please note that the test case requires shutdown support, thus the test
case will be skipped using the current upstream kernel, as it doesn't
have shutdown ioctl support.
[CAUSE]
The direct cause the 0 time stamp in the log tree:
leaf 30507008 items 2 free space 16057 generation 9 owner TREE_LOG
leaf 30507008 flags 0x1(WRITTEN) backref revision 1
checksum stored e522548d
checksum calced e522548d
fs uuid 57d45451-481e-43e4-aa93-289ad707a3a0
chunk uuid d52bd3fd-5163-4337-98a7-7986993ad398
item 0 key (257 INODE_ITEM 0) itemoff 16123 itemsize 160
generation 9 transid 9 size 0 nbytes 0
block group 0 mode 100644 links 1 uid 0 gid 0 rdev 0
sequence 1 flags 0x0(none)
atime 1751432947.492000000 (2025-07-02 14:39:07)
ctime 1751432947.492000000 (2025-07-02 14:39:07)
mtime 1751432947.492000000 (2025-07-02 14:39:07)
otime 0.0 (1970-01-01 09:30:00) <<<
But the old fs tree has all the correct time stamp:
btrfs-progs v6.12
fs tree key (FS_TREE ROOT_ITEM 0)
leaf 30425088 items 2 free space 16061 generation 5 owner FS_TREE
leaf 30425088 flags 0x1(WRITTEN) backref revision 1
checksum stored 48f6c57e
checksum calced 48f6c57e
fs uuid 57d45451-481e-43e4-aa93-289ad707a3a0
chunk uuid d52bd3fd-5163-4337-98a7-7986993ad398
item 0 key (256 INODE_ITEM 0) itemoff 16123 itemsize 160
generation 3 transid 0 size 0 nbytes 16384
block group 0 mode 40755 links 1 uid 0 gid 0 rdev 0
sequence 0 flags 0x0(none)
atime 1751432947.0 (2025-07-02 14:39:07)
ctime 1751432947.0 (2025-07-02 14:39:07)
mtime 1751432947.0 (2025-07-02 14:39:07)
otime 1751432947.0 (2025-07-02 14:39:07) <<<
The root cause is that fill_inode_item() in tree-log.c is only
populating a/c/m time, not the otime (or btime in statx output).
Part of the reason is that, the vfs inode only has a/c/m time, no native
btime support yet.
[FIX]
Thankfully btrfs has its otime stored in btrfs_inode::i_otime_sec and
btrfs_inode::i_otime_nsec.
So what we really need is just fill the otime time stamp in
fill_inode_item() of tree-log.c
There is another fill_inode_item() in inode.c, which is doing the proper
otime population.
Fixes: 94edf4ae43 ("Btrfs: don't bother committing delayed inode updates when fsyncing")
CC: stable@vger.kernel.org
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
[ timespec changes in older tree ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit a0ee1d5faf ]
Ensure the shadow VMCS cache is evicted during an emergency reboot to
prevent potential memory corruption if the cache is evicted after reboot.
This issue was identified through code inspection, as __loaded_vmcs_clear()
flushes both the normal VMCS and the shadow VMCS.
Avoid checking the "launched" state during an emergency reboot, unlike the
behavior in __loaded_vmcs_clear(). This is important because reboot NMIs
can interfere with operations like copy_shadow_to_vmcs12(), where shadow
VMCSes are loaded directly using VMPTRLD. In such cases, if NMIs occur
right after the VMCS load, the shadow VMCSes will be active but the
"launched" state may not be set.
Fixes: 16f5b9034b ("KVM: nVMX: Copy processor-specific shadow-vmcs to VMCS12")
Cc: stable@vger.kernel.org
Signed-off-by: Chao Gao <chao.gao@intel.com>
Reviewed-by: Kai Huang <kai.huang@intel.com>
Link: https://lore.kernel.org/r/20250324140849.2099723-1-chao.gao@intel.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
[ adjusted context ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit c5f1dde7f7 ]
Instead of relying on RTNL, ets_dump() can use READ_ONCE()
annotations, paired with WRITE_ONCE() ones in ets_change().
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Stable-dep-of: 87c6efc5ce ("net/sched: ets: use old 'nbands' while purging unused classes")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 454d3e1ae0 ]
leaf classes of ETS qdiscs are served in strict priority or deficit round
robin (DRR), depending on the value of 'nstrict'. Since this value can be
changed while traffic is running, we need to be sure that the active list
of DRR classes can be updated at any time, so:
1) call INIT_LIST_HEAD(&alist) on all leaf classes in .init(), before the
first packet hits any of them.
2) ensure that 'alist' is not overwritten with zeros when a leaf class is
no more strict priority nor DRR (i.e. array elements beyond 'nbands').
Link: https://lore.kernel.org/netdev/YS%2FoZ+f0Nr8eQkzH@dcaratti.users.ipa.redhat.com
Suggested-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Stable-dep-of: 87c6efc5ce ("net/sched: ets: use old 'nbands' while purging unused classes")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 961a325bec ]
Lockdep reports a bogus possible deadlock on MT8192 Chromebooks due to
the following lock sequences:
1. lock(i2c_register_adapter) [1]; lock(&ec_dev->lock)
2. lock(&ec_dev->lock); lock(prepare_lock);
The actual dependency chains are much longer. The shortened version
looks somewhat like:
1. cros-ec-rpmsg on mtk-scp
ec_dev->lock -> prepare_lock
2. In rt5682_i2c_probe() on native I2C bus:
prepare_lock -> regmap->lock -> (possibly) i2c_adapter->bus_lock
3. In rt5682_i2c_probe() on native I2C bus:
regmap->lock -> i2c_adapter->bus_lock
4. In sbs_probe() on i2c-cros-ec-tunnel I2C bus attached on cros-ec:
i2c_adapter->bus_lock -> ec_dev->lock
While lockdep is correct that the shared lockdep classes have a circular
dependency, it is bogus because
a) 2+3 happen on a native I2C bus
b) 4 happens on the actual EC on ChromeOS devices
c) 1 happens on the SCP coprocessor on MediaTek Chromebooks that just
happens to expose a cros-ec interface, but does not have an
i2c-cros-ec-tunnel I2C bus
In short, the "dependencies" are actually on different devices.
Setup a per-device lockdep key for cros_ec devices so lockdep can tell
the two instances apart. This helps with getting rid of the bogus
lockdep warning. For ChromeOS devices that only have one cros-ec
instance this doesn't change anything.
Also add a missing mutex_destroy, just to make the teardown complete.
[1] This is likely the per I2C bus lock with shared lockdep class
Signed-off-by: Chen-Yu Tsai <wenst@chromium.org>
Signed-off-by: Tzung-Bi Shih <tzungbi@kernel.org>
Link: https://lore.kernel.org/r/20230111074146.2624496-1-wenst@chromium.org
Stable-dep-of: e237495346 ("platform/chrome: cros_ec: Unregister notifier in cros_ec_unregister()")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 5bd398e20f ]
When a remote device sends a completion event to the host, it contains a
pointer to the consumed TRE. The host uses this pointer to process all of
the TREs between it and the host's local copy of the ring's read pointer.
This works when processing completion for chained transactions, but can
lead to nasty results if the device sends an event for a single-element
transaction with a read pointer that is multiple elements ahead of the
host's read pointer.
For instance, if the host accesses an event ring while the device is
updating it, the pointer inside of the event might still point to an old
TRE. If the host uses the channel's xfer_cb() to directly free the buffer
pointed to by the TRE, the buffer will be double-freed.
This behavior was observed on an ep that used upstream EP stack without
'commit 6f18d174b7 ("bus: mhi: ep: Update read pointer only after buffer
is written")'. Where the device updated the events ring pointer before
updating the event contents, so it left a window where the host was able to
access the stale data the event pointed to, before the device had the
chance to update them. The usual pattern was that the host received an
event pointing to a TRE that is not immediately after the last processed
one, so it got treated as if it was a chained transaction, processing all
of the TREs in between the two read pointers.
This commit aims to harden the host by ensuring transactions where the
event points to a TRE that isn't local_rp + 1 are chained.
Fixes: 1d3173a3ba ("bus: mhi: core: Add support for processing events from client device")
Signed-off-by: Youssef Samir <quic_yabdulra@quicinc.com>
[mani: added stable tag and reworded commit message]
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20250714163039.3438985-1-quic_yabdulra@quicinc.com
[ Replaced missing MHI_TRE_DATA_GET_CHAIN macro with direct bit 0 check on dword[1]. ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit ed62a62a18 ]
Improve the description of the possible default SATA link power
management policies and add the missing description for policy 5.
No functional changes.
Fixes: a5ec5a7bfd ("ata: ahci: Support state with min power but Partial low power state")
Cc: stable@vger.kernel.org
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
[ Adjust context ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit cb020bf522 ]
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/20230405141009.3400693-8-u.kleine-koenig@pengutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Stable-dep-of: 1473e9e767 ("usb: musb: omap2430: fix device leak at unbind")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 59305202c6 ]
Memory hot remove unmaps and tears down various kernel page table regions
as required. The ptdump code can race with concurrent modifications of
the kernel page tables. When leaf entries are modified concurrently, the
dump code may log stale or inconsistent information for a VA range, but
this is otherwise not harmful.
But when intermediate levels of kernel page table are freed, the dump code
will continue to use memory that has been freed and potentially
reallocated for another purpose. In such cases, the ptdump code may
dereference bogus addresses, leading to a number of potential problems.
To avoid the above mentioned race condition, platforms such as arm64,
riscv and s390 take memory hotplug lock, while dumping kernel page table
via the sysfs interface /sys/kernel/debug/kernel_page_tables.
Similar race condition exists while checking for pages that might have
been marked W+X via /sys/kernel/debug/kernel_page_tables/check_wx_pages
which in turn calls ptdump_check_wx(). Instead of solving this race
condition again, let's just move the memory hotplug lock inside generic
ptdump_check_wx() which will benefit both the scenarios.
Drop get_online_mems() and put_online_mems() combination from all existing
platform ptdump code paths.
Link: https://lkml.kernel.org/r/20250620052427.2092093-1-anshuman.khandual@arm.com
Fixes: bbd6ec605c ("arm64/mm: Enable memory hot remove")
Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Dev Jain <dev.jain@arm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com> [s390]
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: Gerald Schaefer <gerald.schaefer@linux.ibm.com>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Christian Borntraeger <borntraeger@linux.ibm.com>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 3f66ccbaae ]
REQ_OP_ZONE_FINISH is defined as "12", which makes
op_is_write(REQ_OP_ZONE_FINISH) return false, despite the fact that a
zone finish operation is an operation that modifies a zone (transition
it to full) and so should be considered as a write operation (albeit
one that does not transfer any data to the device).
Fix this by redefining REQ_OP_ZONE_FINISH to be an odd number (13), and
redefine REQ_OP_ZONE_RESET and REQ_OP_ZONE_RESET_ALL using sequential
odd numbers from that new value.
Fixes: 6c1b1da58f ("block: add zone open, close and finish operations")
Cc: stable@vger.kernel.org
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20250625093327.548866-2-dlemoal@kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
[ More renumbering... ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 6cff20ce3b ]
pci_bridge_d3_possible() is called from both pcie_portdrv_probe() and
pcie_portdrv_remove() to determine whether runtime power management shall
be enabled (on probe) or disabled (on remove) on a PCIe port.
The underlying assumption is that pci_bridge_d3_possible() always returns
the same value, else a runtime PM reference imbalance would occur. That
assumption is not given if the PCIe port is inaccessible on remove due to
hot-unplug: pci_bridge_d3_possible() calls pciehp_is_native(), which
accesses Config Space to determine whether the port is Hot-Plug Capable.
An inaccessible port returns "all ones", which is converted to "all
zeroes" by pcie_capability_read_dword(). Hence the port no longer seems
Hot-Plug Capable on remove even though it was on probe.
The resulting runtime PM ref imbalance causes warning messages such as:
pcieport 0000:02:04.0: Runtime PM usage count underflow!
Avoid the Config Space access (and thus the runtime PM ref imbalance) by
caching the Hot-Plug Capable bit in struct pci_dev.
The struct already contains an "is_hotplug_bridge" flag, which however is
not only set on Hot-Plug Capable PCIe ports, but also Conventional PCI
Hot-Plug bridges and ACPI slots. The flag identifies bridges which are
allocated additional MMIO and bus number resources to allow for hierarchy
expansion.
The kernel is somewhat sloppily using "is_hotplug_bridge" in a number of
places to identify Hot-Plug Capable PCIe ports, even though the flag
encompasses other devices. Subsequent commits replace these occurrences
with the new flag to clearly delineate Hot-Plug Capable PCIe ports from
other kinds of hotplug bridges.
Document the existing "is_hotplug_bridge" and the new "is_pciehp" flag
and document the (non-obvious) requirement that pci_bridge_d3_possible()
always returns the same value across the entire lifetime of a bridge,
including its hot-removal.
Fixes: 5352a44a56 ("PCI: pciehp: Make pciehp_is_native() stricter")
Reported-by: Laurent Bigonville <bigon@bigon.be>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=220216
Reported-by: Mario Limonciello <mario.limonciello@amd.com>
Closes: https://lore.kernel.org/r/20250609020223.269407-3-superm1@kernel.org/
Link: https://lore.kernel.org/all/20250620025535.3425049-3-superm1@kernel.org/T/#u
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Rafael J. Wysocki <rafael@kernel.org>
Cc: stable@vger.kernel.org # v4.18+
Link: https://patch.msgid.link/fe5dcc3b2e62ee1df7905d746bde161eb1b3291c.1752390101.git.lukas@wunner.de
[ changed "recent enough PCIe ports" comment to "some PCIe ports" ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 1e61f6ab08 ]
This patch fixes a race condition communication error, which ends up in
PD hard resets when losing the race. Some systems, like the Radxa ROCK
5B are powered through USB-C without any backup power source and use a
FUSB302 chip to do the PD negotiation. This means it is quite important
to avoid hard resets, since that effectively kills the system's
power-supply.
I've found the following race condition while debugging unplanned power
loss during booting the board every now and then:
1. lots of TCPM/FUSB302/PD initialization stuff
2. TCPM ends up in SNK_WAIT_CAPABILITIES (tcpm_set_pd_rx is enabled here)
3. the remote PD source does not send anything, so TCPM does a SOFT RESET
4. TCPM ends up in SNK_WAIT_CAPABILITIES for the second time
(tcpm_set_pd_rx is enabled again, even though it is still on)
At this point I've seen broken CRC good messages being send by the
FUSB302 with a logic analyzer sniffing the CC lines. Also it looks like
messages are being lost and things generally going haywire with one of
the two sides doing a hard reset once a broken CRC good message was send
to the bus.
I think the system is running into a race condition, that the FIFOs are
being cleared and/or the automatic good CRC message generation flag is
being updated while a message is already arriving.
Let's avoid this by caching the PD RX enabled state, as we have already
processed anything in the FIFOs and are in a good state. As a side
effect that this also optimizes I2C bus usage :)
As far as I can tell the problem theoretically also exists when TCPM
enters SNK_WAIT_CAPABILITIES the first time, but I believe this is less
critical for the following reason:
On devices like the ROCK 5B, which are powered through a TCPM backed
USB-C port, the bootloader must have done some prior PD communication
(initial communication must happen within 5 seconds after plugging the
USB-C plug). This means the first time the kernel TCPM state machine
reaches SNK_WAIT_CAPABILITIES, the remote side is not sending messages
actively. On other devices a hard reset simply adds some extra delay and
things should be good afterwards.
Fixes: c034a43e72 ("staging: typec: Fairchild FUSB302 Type-c chip driver")
Cc: stable <stable@kernel.org>
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Reviewed-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Link: https://lore.kernel.org/r/20250704-fusb302-race-condition-fix-v1-1-239012c0e27a@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ replaced str_on_off(on) with ternary operator ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 8e7d178d06 ]
In ksmbd_extract_shortname(), strscpy() is incorrectly called with the
length of the source string (excluding the NUL terminator) rather than
the size of the destination buffer. This results in "__" being copied
to 'extension' rather than "___" (two underscores instead of three).
Use the destination buffer size instead to ensure that the string "___"
(three underscores) is copied correctly.
Cc: stable@vger.kernel.org
Fixes: e2f34481b2 ("cifsd: add server-side procedures for SMB3")
Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 8a15ca0ca5 ]
During communication with Focusrite Scarlett Gen 2/3/4 USB audio
interfaces, -EPROTO is sometimes returned from scarlett2_usb_tx(),
snd_usb_ctl_msg() which can cause initialisation and control
operations to fail intermittently.
This patch adds up to 5 retries in scarlett2_usb(), with a delay
starting at 5ms and doubling each time. This follows the same approach
as the fix for usb_set_interface() in endpoint.c (commit f406005e16
("ALSA: usb-audio: Add retry on -EPROTO from usb_set_interface()")),
which resolved similar -EPROTO issues during device initialisation,
and is the same approach as in fcp.c:fcp_usb().
Fixes: 9e4d5c1be2 ("ALSA: usb-audio: Scarlett Gen 2 mixer interface")
Closes: https://github.com/geoffreybennett/linux-fcp/issues/41
Cc: stable@vger.kernel.org
Signed-off-by: Geoffrey D. Bennett <g@b4.vu>
Link: https://patch.msgid.link/aIdDO6ld50WQwNim@m.b4.vu
Signed-off-by: Takashi Iwai <tiwai@suse.de>
[ Applied patch to the older file name mixer_scarlett_gen2.c instead of mixer_scarlett2.c. ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 1cec9ac2d0 ]
Right now, if XRSTOR fails a console message like this is be printed:
Bad FPU state detected at restore_fpregs_from_fpstate+0x9a/0x170, reinitializing FPU registers.
However, the text location (...+0x9a in this case) is the instruction
*AFTER* the XRSTOR. The highlighted instruction in the "Code:" dump
also points one instruction late.
The reason is that the "fixup" moves RIP up to pass the bad XRSTOR and
keep on running after returning from the #GP handler. But it does this
fixup before warning.
The resulting warning output is nonsensical because it looks like the
non-FPU-related instruction is #GP'ing.
Do not fix up RIP until after printing the warning. Do this by using
the more generic and standard ex_handler_default().
Fixes: d5c8028b47 ("x86/fpu: Reinitialize FPU registers if restoring FPU state fails")
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Chao Gao <chao.gao@intel.com>
Acked-by: Alison Schofield <alison.schofield@intel.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc:stable@vger.kernel.org
Link: https://lore.kernel.org/all/20250624210148.97126F9E%40davehans-spike.ostc.intel.com
[ Replace fpu_reset_from_exception_fixup() with __restore_fpregs_from_fpstate() ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit b220bed633 ]
The CVE-2024-50047 fix removed asynchronous crypto handling from
crypt_message(), assuming all crypto operations are synchronous.
However, when hardware crypto accelerators are used, this can cause
use-after-free crashes:
crypt_message()
// Allocate the creq buffer containing the req
creq = smb2_get_aead_req(..., &req);
// Async encryption returns -EINPROGRESS immediately
rc = enc ? crypto_aead_encrypt(req) : crypto_aead_decrypt(req);
// Free creq while async operation is still in progress
kvfree_sensitive(creq, ...);
Hardware crypto modules often implement async AEAD operations for
performance. When crypto_aead_encrypt/decrypt() returns -EINPROGRESS,
the operation completes asynchronously. Without crypto_wait_req(),
the function immediately frees the request buffer, leading to crashes
when the driver later accesses the freed memory.
This results in a use-after-free condition when the hardware crypto
driver later accesses the freed request structure, leading to kernel
crashes with NULL pointer dereferences.
The issue occurs because crypto_alloc_aead() with mask=0 doesn't
guarantee synchronous operation. Even without CRYPTO_ALG_ASYNC in
the mask, async implementations can be selected.
Fix by restoring the async crypto handling:
- DECLARE_CRYPTO_WAIT(wait) for completion tracking
- aead_request_set_callback() for async completion notification
- crypto_wait_req() to wait for operation completion
This ensures the request buffer isn't freed until the crypto operation
completes, whether synchronous or asynchronous, while preserving the
CVE-2024-50047 fix.
Fixes: b0abcd65ec ("smb: client: fix UAF in async decryption")
Link: https://lore.kernel.org/all/8b784a13-87b0-4131-9ff9-7a8993538749@huaweicloud.com/
Cc: stable@vger.kernel.org
Reviewed-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Signed-off-by: Wang Zhaolong <wangzhaolong@huaweicloud.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 2521106fc7 ]
Hub driver warm-resets ports in SS.Inactive or Compliance mode to
recover a possible connected device. The port reset code correctly
detects if a connection is lost during reset, but hub driver
port_event() fails to take this into account in some cases.
port_event() ends up using stale values and assumes there is a
connected device, and will try all means to recover it, including
power-cycling the port.
Details:
This case was triggered when xHC host was suspended with DbC (Debug
Capability) enabled and connected. DbC turns one xHC port into a simple
usb debug device, allowing debugging a system with an A-to-A USB debug
cable.
xhci DbC code disables DbC when xHC is system suspended to D3, and
enables it back during resume.
We essentially end up with two hosts connected to each other during
suspend, and, for a short while during resume, until DbC is enabled back.
The suspended xHC host notices some activity on the roothub port, but
can't train the link due to being suspended, so xHC hardware sets a CAS
(Cold Attach Status) flag for this port to inform xhci host driver that
the port needs to be warm reset once xHC resumes.
CAS is xHCI specific, and not part of USB specification, so xhci driver
tells usb core that the port has a connection and link is in compliance
mode. Recovery from complinace mode is similar to CAS recovery.
xhci CAS driver support that fakes a compliance mode connection was added
in commit 8bea2bd37d ("usb: Add support for root hub port status CAS")
Once xHCI resumes and DbC is enabled back, all activity on the xHC
roothub host side port disappears. The hub driver will anyway think
port has a connection and link is in compliance mode, and hub driver
will try to recover it.
The port power-cycle during recovery seems to cause issues to the active
DbC connection.
Fix this by clearing connect_change flag if hub_port_reset() returns
-ENOTCONN, thus avoiding the whole unnecessary port recovery and
initialization attempt.
Cc: stable@vger.kernel.org
Fixes: 8bea2bd37d ("usb: Add support for root hub port status CAS")
Tested-by: Łukasz Bartosik <ukaszb@chromium.org>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Link: https://lore.kernel.org/r/20250623133947.3144608-1-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit f59f93cd1d ]
During disconnect USB-3 ports often go via SS.Inactive link error state
before the missing terminations are noticed, and link finally goes to
RxDetect state
Avoid immediately warm-resetting ports in SS.Inactive state.
Let ports settle for a while and re-read the link status a few times 20ms
apart to see if the ports transitions out of SS.Inactive.
According to USB 3.x spec 7.5.2, a port in SS.Inactive should
automatically check for missing far-end receiver termination every
12 ms (SSInactiveQuietTimeout)
The futile multiple warm reset retries of a disconnected device takes
a lot of time, also the resetting of a removed devices has caused cases
where the reset bit got stuck for a long time on xHCI roothub.
This lead to issues in detecting new devices connected to the same port
shortly after.
Tested-by: Mark Pearson <markpearson@lenovo.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20211210111653.1378381-1-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Stable-dep-of: 2521106fc7 ("usb: hub: Don't try to recover devices lost during warm reset.")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit d66e1e90b1 ]
Ensure that sysfs init doesn't fail for new/unrecognized bank types or if
a bank has additional blocks available.
Most MCA banks have a single thresholding block, so the block takes the same
name as the bank.
Unified Memory Controllers (UMCs) are a special case where there are two
blocks and each has a unique name.
However, the microarchitecture allows for five blocks. Any new MCA bank types
with more than one block will be missing names for the extra blocks. The MCE
sysfs will fail to initialize in this case.
Fixes: 87a6d4091b ("x86/mce/AMD: Update sysfs bank names for SMCA systems")
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/20250624-wip-mca-updates-v4-3-236dd74f645f@amd.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 87c4e1459e ]
After commit d5c8d6e0fa ("kbuild: Update assembler calls to use proper
flags and language target"), which updated as-instr to use the
'assembler-with-cpp' language option, the Kbuild version of as-instr
always fails internally for arch/arm with
<command-line>: fatal error: asm/unified.h: No such file or directory
compilation terminated.
because '-include' flags are now taken into account by the compiler
driver and as-instr does not have '$(LINUXINCLUDE)', so unified.h is not
found.
This went unnoticed at the time of the Kbuild change because the last
use of as-instr in Kbuild that arch/arm could reach was removed in 5.7
by commit 541ad0150c ("arm: Remove 32bit KVM host support") but a
stable backport of the Kbuild change to before that point exposed this
potential issue if one were to be reintroduced.
Follow the general pattern of '-include' paths throughout the tree and
make unified.h absolute using '$(srctree)' to ensure KBUILD_AFLAGS can
be used independently.
Closes: https://lore.kernel.org/CACo-S-1qbCX4WAVFA63dWfHtrRHZBTyyr2js8Lx=Az03XHTTHg@mail.gmail.com/
Cc: stable@vger.kernel.org
Fixes: d5c8d6e0fa ("kbuild: Update assembler calls to use proper flags and language target")
Reported-by: KernelCI bot <bot@kernelci.org>
Reviewed-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
[ adapted to missing -Wa ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit d42e6c20de ]
`cpu_switch_to()` and `call_on_irq_stack()` manipulate SP to change
to different stacks along with the Shadow Call Stack if it is enabled.
Those two stack changes cannot be done atomically and both functions
can be interrupted by SErrors or Debug Exceptions which, though unlikely,
is very much broken : if interrupted, we can end up with mismatched stacks
and Shadow Call Stack leading to clobbered stacks.
In `cpu_switch_to()`, it can happen when SP_EL0 points to the new task,
but x18 stills points to the old task's SCS. When the interrupt handler
tries to save the task's SCS pointer, it will save the old task
SCS pointer (x18) into the new task struct (pointed to by SP_EL0),
clobbering it.
In `call_on_irq_stack()`, it can happen when switching from the task stack
to the IRQ stack and when switching back. In both cases, we can be
interrupted when the SCS pointer points to the IRQ SCS, but SP points to
the task stack. The nested interrupt handler pushes its return addresses
on the IRQ SCS. It then detects that SP points to the task stack,
calls `call_on_irq_stack()` and clobbers the task SCS pointer with
the IRQ SCS pointer, which it will also use !
This leads to tasks returning to addresses on the wrong SCS,
or even on the IRQ SCS, triggering kernel panics via CONFIG_VMAP_STACK
or FPAC if enabled.
This is possible on a default config, but unlikely.
However, when enabling CONFIG_ARM64_PSEUDO_NMI, DAIF is unmasked and
instead the GIC is responsible for filtering what interrupts the CPU
should receive based on priority.
Given the goal of emulating NMIs, pseudo-NMIs can be received by the CPU
even in `cpu_switch_to()` and `call_on_irq_stack()`, possibly *very*
frequently depending on the system configuration and workload, leading
to unpredictable kernel panics.
Completely mask DAIF in `cpu_switch_to()` and restore it when returning.
Do the same in `call_on_irq_stack()`, but restore and mask around
the branch.
Mask DAIF even if CONFIG_SHADOW_CALL_STACK is not enabled for consistency
of behaviour between all configurations.
Introduce and use an assembly macro for saving and masking DAIF,
as the existing one saves but only masks IF.
Cc: <stable@vger.kernel.org>
Signed-off-by: Ada Couprie Diaz <ada.coupriediaz@arm.com>
Reported-by: Cristian Prundeanu <cpru@amazon.com>
Fixes: 59b37fe52f ("arm64: Stash shadow stack pointer in the task struct on interrupt")
Tested-by: Cristian Prundeanu <cpru@amazon.com>
Acked-by: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250718142814.133329-1-ada.coupriediaz@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
[ removed duplicate save_and_disable_daif macro ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 15f77764e9 ]
When application A submits jobs and application B submits a job with a
dependency on A's fence, the normal flow wakes up the scheduler after
processing each job. However, the optimization in
drm_sched_entity_add_dependency_cb() uses a callback that only clears
dependencies without waking up the scheduler.
When application A is killed before its jobs can run, the callback gets
triggered but only clears the dependency without waking up the scheduler,
causing the scheduler to enter sleep state and application B to hang.
Remove the optimization by deleting drm_sched_entity_clear_dep() and its
usage, ensuring the scheduler is always woken up when dependencies are
cleared.
Fixes: 777dbd458c ("drm/amdgpu: drop a dummy wakeup scheduler")
Cc: stable@vger.kernel.org # v4.6+
Signed-off-by: Lin.Cao <lincao12@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250717084453.921097-1-lincao12@amd.com
[ Changed drm_sched_wakeup_if_can_queue() calls to drm_sched_wakeup() ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 8ec396d05d ]
Patch series "mm: reinstate ability to map write-sealed memfd mappings
read-only".
In commit 158978945f ("mm: perform the mapping_map_writable() check
after call_mmap()") (and preceding changes in the same series) it became
possible to mmap() F_SEAL_WRITE sealed memfd mappings read-only.
Commit 5de195060b ("mm: resolve faulty mmap_region() error path
behaviour") unintentionally undid this logic by moving the
mapping_map_writable() check before the shmem_mmap() hook is invoked,
thereby regressing this change.
This series reworks how we both permit write-sealed mappings being mapped
read-only and disallow mprotect() from undoing the write-seal, fixing this
regression.
We also add a regression test to ensure that we do not accidentally
regress this in future.
Thanks to Julian Orth for reporting this regression.
This patch (of 2):
In commit 158978945f ("mm: perform the mapping_map_writable() check
after call_mmap()") (and preceding changes in the same series) it became
possible to mmap() F_SEAL_WRITE sealed memfd mappings read-only.
This was previously unnecessarily disallowed, despite the man page
documentation indicating that it would be, thereby limiting the usefulness
of F_SEAL_WRITE logic.
We fixed this by adapting logic that existed for the F_SEAL_FUTURE_WRITE
seal (one which disallows future writes to the memfd) to also be used for
F_SEAL_WRITE.
For background - the F_SEAL_FUTURE_WRITE seal clears VM_MAYWRITE for a
read-only mapping to disallow mprotect() from overriding the seal - an
operation performed by seal_check_write(), invoked from shmem_mmap(), the
f_op->mmap() hook used by shmem mappings.
By extending this to F_SEAL_WRITE and critically - checking
mapping_map_writable() to determine if we may map the memfd AFTER we
invoke shmem_mmap() - the desired logic becomes possible. This is because
mapping_map_writable() explicitly checks for VM_MAYWRITE, which we will
have cleared.
Commit 5de195060b ("mm: resolve faulty mmap_region() error path
behaviour") unintentionally undid this logic by moving the
mapping_map_writable() check before the shmem_mmap() hook is invoked,
thereby regressing this change.
We reinstate this functionality by moving the check out of shmem_mmap()
and instead performing it in do_mmap() at the point at which VMA flags are
being determined, which seems in any case to be a more appropriate place
in which to make this determination.
In order to achieve this we rework memfd seal logic to allow us access to
this information using existing logic and eliminate the clearing of
VM_MAYWRITE from seal_check_write() which we are performing in do_mmap()
instead.
Link: https://lkml.kernel.org/r/99fc35d2c62bd2e05571cf60d9f8b843c56069e0.1732804776.git.lorenzo.stoakes@oracle.com
Fixes: 5de195060b ("mm: resolve faulty mmap_region() error path behaviour")
Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Reported-by: Julian Orth <ju.orth@gmail.com>
Closes: https://lore.kernel.org/all/CAHijbEUMhvJTN9Xw1GmbM266FXXv=U7s4L_Jem5x3AaPZxrYpQ@mail.gmail.com/
Cc: Jann Horn <jannh@google.com>
Cc: Liam R. Howlett <Liam.Howlett@Oracle.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Isaac J. Manjarres <isaacmanjarres@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit e8e17ee90e ]
Patch series "permit write-sealed memfd read-only shared mappings", v4.
The man page for fcntl() describing memfd file seals states the following
about F_SEAL_WRITE:-
Furthermore, trying to create new shared, writable memory-mappings via
mmap(2) will also fail with EPERM.
With emphasis on 'writable'. In turns out in fact that currently the
kernel simply disallows all new shared memory mappings for a memfd with
F_SEAL_WRITE applied, rendering this documentation inaccurate.
This matters because users are therefore unable to obtain a shared mapping
to a memfd after write sealing altogether, which limits their usefulness.
This was reported in the discussion thread [1] originating from a bug
report [2].
This is a product of both using the struct address_space->i_mmap_writable
atomic counter to determine whether writing may be permitted, and the
kernel adjusting this counter when any VM_SHARED mapping is performed and
more generally implicitly assuming VM_SHARED implies writable.
It seems sensible that we should only update this mapping if VM_MAYWRITE
is specified, i.e. whether it is possible that this mapping could at any
point be written to.
If we do so then all we need to do to permit write seals to function as
documented is to clear VM_MAYWRITE when mapping read-only. It turns out
this functionality already exists for F_SEAL_FUTURE_WRITE - we can
therefore simply adapt this logic to do the same for F_SEAL_WRITE.
We then hit a chicken and egg situation in mmap_region() where the check
for VM_MAYWRITE occurs before we are able to clear this flag. To work
around this, perform this check after we invoke call_mmap(), with careful
consideration of error paths.
Thanks to Andy Lutomirski for the suggestion!
[1]:https://lore.kernel.org/all/20230324133646.16101dfa666f253c4715d965@linux-foundation.org/
[2]:https://bugzilla.kernel.org/show_bug.cgi?id=217238
This patch (of 3):
There is a general assumption that VMAs with the VM_SHARED flag set are
writable. If the VM_MAYWRITE flag is not set, then this is simply not the
case.
Update those checks which affect the struct address_space->i_mmap_writable
field to explicitly test for this by introducing
[vma_]is_shared_maywrite() helper functions.
This remains entirely conservative, as the lack of VM_MAYWRITE guarantees
that the VMA cannot be written to.
Link: https://lkml.kernel.org/r/cover.1697116581.git.lstoakes@gmail.com
Link: https://lkml.kernel.org/r/d978aefefa83ec42d18dfa964ad180dbcde34795.1697116581.git.lstoakes@gmail.com
Signed-off-by: Lorenzo Stoakes <lstoakes@gmail.com>
Suggested-by: Andy Lutomirski <luto@kernel.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Muchun Song <muchun.song@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: stable@vger.kernel.org
[isaacmanjarres: resolved merge conflicts due to
due to refactoring that happened in upstream commit
5de195060b ("mm: resolve faulty mmap_region() error path behaviour")]
Signed-off-by: Isaac J. Manjarres <isaacmanjarres@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit ccab044697 upstream.
When skb_ext_add(skb, SKB_EXT_MPTCP) fails in mptcp_incoming_options(),
we used to return true, letting the segment proceed through the TCP
receive path without a DSS mapping. Such segments can leave inconsistent
mapping state and trigger a mid-stream fallback to TCP, which in testing
collapsed (by artificially forcing failures in skb_ext_add) throughput
to zero.
Return false instead so the TCP input path drops the skb (see
tcp_data_queue() and step-7 processing). This is the safer choice
under memory pressure: it preserves MPTCP correctness and provides
backpressure to the sender.
Control packets remain unaffected: ACK updates and DATA_FIN handling
happen before attempting the extension allocation, and tcp_reset()
continues to ignore the return value.
With this change, MPTCP continues to work at high throughput if we
artificially inject failures into skb_ext_add.
Fixes: 6787b7e350 ("mptcp: avoid processing packet if a subflow reset")
Cc: stable@vger.kernel.org
Signed-off-by: Christoph Paasch <cpaasch@openai.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20250815-net-mptcp-misc-fixes-6-17-rc2-v1-1-521fe9957892@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 7af76e9d18 upstream.
Receiving HSR frame with insufficient space to hold HSR tag in the skb
can result in a crash (kernel BUG):
[ 45.390915] skbuff: skb_under_panic: text:ffffffff86f32cac len:26 put:14 head:ffff888042418000 data:ffff888042417ff4 tail:0xe end:0x180 dev:bridge_slave_1
[ 45.392559] ------------[ cut here ]------------
[ 45.392912] kernel BUG at net/core/skbuff.c:211!
[ 45.393276] Oops: invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN NOPTI
[ 45.393809] CPU: 1 UID: 0 PID: 2496 Comm: reproducer Not tainted 6.15.0 #12 PREEMPT(undef)
[ 45.394433] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.3-0-ga6ed6b701f0a-prebuilt.qemu.org 04/01/2014
[ 45.395273] RIP: 0010:skb_panic+0x15b/0x1d0
<snip registers, remove unreliable trace>
[ 45.402911] Call Trace:
[ 45.403105] <IRQ>
[ 45.404470] skb_push+0xcd/0xf0
[ 45.404726] br_dev_queue_push_xmit+0x7c/0x6c0
[ 45.406513] br_forward_finish+0x128/0x260
[ 45.408483] __br_forward+0x42d/0x590
[ 45.409464] maybe_deliver+0x2eb/0x420
[ 45.409763] br_flood+0x174/0x4a0
[ 45.410030] br_handle_frame_finish+0xc7c/0x1bc0
[ 45.411618] br_handle_frame+0xac3/0x1230
[ 45.413674] __netif_receive_skb_core.constprop.0+0x808/0x3df0
[ 45.422966] __netif_receive_skb_one_core+0xb4/0x1f0
[ 45.424478] __netif_receive_skb+0x22/0x170
[ 45.424806] process_backlog+0x242/0x6d0
[ 45.425116] __napi_poll+0xbb/0x630
[ 45.425394] net_rx_action+0x4d1/0xcc0
[ 45.427613] handle_softirqs+0x1a4/0x580
[ 45.427926] do_softirq+0x74/0x90
[ 45.428196] </IRQ>
This issue was found by syzkaller.
The panic happens in br_dev_queue_push_xmit() once it receives a
corrupted skb with ETH header already pushed in linear data. When it
attempts the skb_push() call, there's not enough headroom and
skb_push() panics.
The corrupted skb is put on the queue by HSR layer, which makes a
sequence of unintended transformations when it receives a specific
corrupted HSR frame (with incomplete TAG).
Fix it by dropping and consuming frames that are not long enough to
contain both ethernet and hsr headers.
Alternative fix would be to check for enough headroom before skb_push()
in br_dev_queue_push_xmit().
In the reproducer, this is injected via AF_PACKET, but I don't easily
see why it couldn't be sent over the wire from adjacent network.
Further Details:
In the reproducer, the following network interface chain is set up:
┌────────────────┐ ┌────────────────┐
│ veth0_to_hsr ├───┤ hsr_slave0 ┼───┐
└────────────────┘ └────────────────┘ │
│ ┌──────┐
├─┤ hsr0 ├───┐
│ └──────┘ │
┌────────────────┐ ┌────────────────┐ │ │┌────────┐
│ veth1_to_hsr ┼───┤ hsr_slave1 ├───┘ └┤ │
└────────────────┘ └────────────────┘ ┌┼ bridge │
││ │
│└────────┘
│
┌───────┐ │
│ ... ├──────┘
└───────┘
To trigger the events leading up to crash, reproducer sends a corrupted
HSR frame with incomplete TAG, via AF_PACKET socket on 'veth0_to_hsr'.
The first HSR-layer function to process this frame is
hsr_handle_frame(). It and then checks if the
protocol is ETH_P_PRP or ETH_P_HSR. If it is, it calls
skb_set_network_header(skb, ETH_HLEN + HSR_HLEN), without checking that
the skb is long enough. For the crashing frame it is not, and hence the
skb->network_header and skb->mac_len fields are set incorrectly,
pointing after the end of the linear buffer.
I will call this a BUG#1 and it is what is addressed by this patch. In
the crashing scenario before the fix, the skb continues to go down the
hsr path as follows.
hsr_handle_frame() then calls this sequence
hsr_forward_skb()
fill_frame_info()
hsr->proto_ops->fill_frame_info()
hsr_fill_frame_info()
hsr_fill_frame_info() contains a check that intends to check whether the
skb actually contains the HSR header. But the check relies on the
skb->mac_len field which was erroneously setup due to BUG#1, so the
check passes and the execution continues back in the hsr_forward_skb():
hsr_forward_skb()
hsr_forward_do()
hsr->proto_ops->get_untagged_frame()
hsr_get_untagged_frame()
create_stripped_skb_hsr()
In create_stripped_skb_hsr(), a copy of the skb is created and is
further corrupted by operation that attempts to strip the HSR tag in a
call to __pskb_copy().
The skb enters create_stripped_skb_hsr() with ethernet header pushed in
linear buffer. The skb_pull(skb_in, HSR_HLEN) thus pulls 6 bytes of
ethernet header into the headroom, creating skb_in with a headroom of
size 8. The subsequent __pskb_copy() then creates an skb with headroom
of just 2 and skb->len of just 12, this is how it looks after the copy:
gdb) p skb->len
$10 = 12
(gdb) p skb->data
$11 = (unsigned char *) 0xffff888041e45382 "\252\252\252\252\252!\210\373",
(gdb) p skb->head
$12 = (unsigned char *) 0xffff888041e45380 ""
It seems create_stripped_skb_hsr() assumes that ETH header is pulled
in the headroom when it's entered, because it just pulls HSR header on
top. But that is not the case in our code-path and we end up with the
corrupted skb instead. I will call this BUG#2
*I got confused here because it seems that under no conditions can
create_stripped_skb_hsr() work well, the assumption it makes is not true
during the processing of hsr frames - since the skb_push() in
hsr_handle_frame to skb_pull in hsr_deliver_master(). I wonder whether I
missed something here.*
Next, the execution arrives in hsr_deliver_master(). It calls
skb_pull(ETH_HLEN), which just returns NULL - the SKB does not have
enough space for the pull (as it only has 12 bytes in total at this
point).
*The skb_pull() here further suggests that ethernet header is meant
to be pushed through the whole hsr processing and
create_stripped_skb_hsr() should pull it before doing the HSR header
pull.*
hsr_deliver_master() then puts the corrupted skb on the queue, it is
then picked up from there by bridge frame handling layer and finally
lands in br_dev_queue_push_xmit where it panics.
Cc: stable@kernel.org
Fixes: 48b491a5cc ("net: hsr: fix mac_len checks")
Reported-by: syzbot+a81f2759d022496b40ab@syzkaller.appspotmail.com
Signed-off-by: Jakub Acs <acsjakub@amazon.de>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20250819082842.94378-1-acsjakub@amazon.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cong Wang
Qu Wenruo
Chao Gao
Davide Caratti
Eric Dumazet
Tzung-Bi Shih
Chen-Yu Tsai
Uwe Kleine-König
Johan Hovold
Youssef Samir
Damien Le Moal
Anshuman Khandual
Trond Myklebust
Anna Schumaker
Lukas Wunner
Sebastian Reichel
Haiyang Zhang
Thorsten Blum
Geoffrey D. Bennett
Dave Hansen
Wang Zhaolong
Mathias Nyman
Yazen Ghannam
Zhang Lixu
Chao Yu
Nathan Chancellor
Ada Couprie Diaz
Lin.Cao
Haoxiang Li
Lorenzo Stoakes
Lorenzo Stoakes
Matthieu Baerts (NGI0)
Christoph Paasch
Eric Biggers
Jakub Acs
Timur Kristóf