Ism Hong [Thu, 1 Jun 2023 09:53:55 +0000 (17:53 +0800)]
riscv: perf: Fix callchain parse error with kernel tracepoint events
For RISC-V, when tracing with tracepoint events, the IP and status are
set to 0, preventing the perf code parsing the callchain and resolving
the symbols correctly.
The fix is to implement perf_arch_fetch_caller_regs for riscv, which
fills several necessary registers used for callchain unwinding,
including epc, sp, s0 and status. It's similar to commit b3eac0265bf6
("arm: perf: Fix callchain parse error with kernel tracepoint events")
and commit 5b09a094f2fb ("arm64: perf: Fix callchain parse error with
kernel tracepoint events").
With this patch, callchain can be parsed correctly as:
Paolo Abeni [Wed, 31 May 2023 19:37:08 +0000 (12:37 -0700)]
mptcp: fix active subflow finalization
Active subflow are inserted into the connection list at creation time.
When the MPJ handshake completes successfully, a new subflow creation
netlink event is generated correctly, but the current code wrongly
avoid initializing a couple of subflow data.
The above will cause misbehavior on a few exceptional events: unneeded
mptcp-level retransmission on msk-level sequence wrap-around and infinite
mapping fallback even when a MPJ socket is present.
Address the issue factoring out the needed initialization in a new helper
and invoking the latter from __mptcp_finish_join() time for passive
subflow and from mptcp_finish_join() for active ones.
Paolo Abeni [Wed, 31 May 2023 19:37:07 +0000 (12:37 -0700)]
mptcp: add annotations around sk->sk_shutdown accesses
Christoph reported the mptcp variant of a recently addressed plain
TCP issue. Similar to commit e14cadfd80d7 ("tcp: add annotations around
sk->sk_shutdown accesses") add READ/WRITE ONCE annotations to silence
KCSAN reports around lockless sk_shutdown access.
Fixes: 71ba088ce0aa ("mptcp: cleanup accept and poll") Reported-by: Christoph Paasch <[email protected]> Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/401 Reviewed-by: Mat Martineau <[email protected]> Signed-off-by: Paolo Abeni <[email protected]> Signed-off-by: Mat Martineau <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
Paolo Abeni [Wed, 31 May 2023 19:37:06 +0000 (12:37 -0700)]
mptcp: fix data race around msk->first access
The first subflow socket is accessed outside the msk socket lock
by mptcp_subflow_fail(), we need to annotate each write access
with WRITE_ONCE, but a few spots still lacks it.
When the msk socket is cloned at MPC handshake time, a few
fields are initialized in a racy way outside mptcp_sk_clone()
and the msk socket lock.
The above is due historical reasons: before commit a88d0092b24b
("mptcp: simplify subflow_syn_recv_sock()") as the first subflow socket
carrying all the needed date was not available yet at msk creation
time
We can now refactor the code moving the missing initialization bit
under the socket lock, removing the init race and avoiding some
code duplication.
This will also simplify the next patch, as all msk->first write
access are now under the msk socket lock.
Paolo Abeni [Wed, 31 May 2023 19:37:04 +0000 (12:37 -0700)]
mptcp: add annotations around msk->subflow accesses
The MPTCP can access the first subflow socket in a few spots
outside the socket lock scope. That is actually safe, as MPTCP
will delete the socket itself only after the msk sock close().
Still the such accesses causes a few KCSAN splats, as reported
by Christoph. Silence the harmless warning adding a few annotation
around the relevant accesses.
Fixes: 71ba088ce0aa ("mptcp: cleanup accept and poll") Reported-by: Christoph Paasch <[email protected]> Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/402 Reviewed-by: Mat Martineau <[email protected]> Signed-off-by: Paolo Abeni <[email protected]> Signed-off-by: Mat Martineau <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
Paolo Abeni [Wed, 31 May 2023 19:37:03 +0000 (12:37 -0700)]
mptcp: fix connect timeout handling
Ondrej reported a functional issue WRT timeout handling on connect
with a nice reproducer.
The problem is that the current mptcp connect waits for both the
MPTCP socket level timeout, and the first subflow socket timeout.
The latter is not influenced/touched by the exposed setsockopt().
Overall the above makes the SO_SNDTIMEO a no-op on connect.
Since mptcp_connect is invoked via inet_stream_connect and the
latter properly handle the MPTCP level timeout, we can address the
issue making the nested subflow level connect always unblocking.
This also allow simplifying a bit the code, dropping an ugly hack
to handle the fastopen and custom proto_ops connect.
The issues predates the blamed commit below, but the current resolution
requires the infrastructure introduced there.
Xin Long [Wed, 31 May 2023 16:01:44 +0000 (12:01 -0400)]
rtnetlink: add the missing IFLA_GRO_ tb check in validate_linkmsg
This fixes the issue that dev gro_max_size and gso_ipv4_max_size
can be set to a huge value:
# ip link add dummy1 type dummy
# ip link set dummy1 gro_max_size 4294967295
# ip -d link show dummy1
dummy addrgenmode eui64 ... gro_max_size 4294967295
Fixes: 0fe79f28bfaf ("net: allow gro_max_size to exceed 65536") Fixes: 9eefedd58ae1 ("net: add gso_ipv4_max_size and gro_ipv4_max_size per device") Reported-by: Xiumei Mu <[email protected]> Signed-off-by: Xin Long <[email protected]> Reviewed-by: Simon Horman <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
Xin Long [Wed, 31 May 2023 16:01:43 +0000 (12:01 -0400)]
rtnetlink: move IFLA_GSO_ tb check to validate_linkmsg
These IFLA_GSO_* tb check should also be done for the new created link,
otherwise, they can be set to a huge value when creating links:
# ip link add dummy1 gso_max_size 4294967295 type dummy
# ip -d link show dummy1
dummy addrgenmode eui64 ... gso_max_size 4294967295
Fixes: 46e6b992c250 ("rtnetlink: allow GSO maximums to be set on device creation") Fixes: 9eefedd58ae1 ("net: add gso_ipv4_max_size and gro_ipv4_max_size per device") Signed-off-by: Xin Long <[email protected]> Reviewed-by: Simon Horman <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
Xin Long [Wed, 31 May 2023 16:01:42 +0000 (12:01 -0400)]
rtnetlink: call validate_linkmsg in rtnl_create_link
validate_linkmsg() was introduced by commit 1840bb13c22f5b ("[RTNL]:
Validate hardware and broadcast address attribute for RTM_NEWLINK")
to validate tb[IFLA_ADDRESS/BROADCAST] for existing links. The same
check should also be done for newly created links.
This patch adds validate_linkmsg() call in rtnl_create_link(), to
avoid the invalid address set when creating some devices like:
# ip link add dummy0 type dummy
# ip link add link dummy0 name mac0 address 01:02 type macsec
Fixes: 0e06877c6fdb ("[RTNETLINK]: rtnl_link: allow specifying initial device address") Signed-off-by: Xin Long <[email protected]> Reviewed-by: Simon Horman <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
ice: recycle/free all of the fragments from multi-buffer frame
The ice driver caches next_to_clean value at the beginning of
ice_clean_rx_irq() in order to remember the first buffer that has to be
freed/recycled after main Rx processing loop. The end boundary is
indicated by first descriptor of frame that Rx processing loop has ended
its duties. Note that if mentioned loop ended in the middle of gathering
multi-buffer frame, next_to_clean would be pointing to the descriptor in
the middle of the frame BUT freeing/recycling stage will stop at the
first descriptor. This means that next iteration of ice_clean_rx_irq()
will miss the (first_desc, next_to_clean - 1) entries.
When running various 9K MTU workloads, such splats were observed:
Li Zhijian [Tue, 25 Apr 2023 01:02:43 +0000 (01:02 +0000)]
RDMA/rtrs: Fix rxe_dealloc_pd warning
In current design:
1. PD and clt_path->s.dev are shared among connections.
2. every con[n]'s cleanup phase will call destroy_con_cq_qp()
3. clt_path->s.dev will be always decreased in destroy_con_cq_qp(), and
when clt_path->s.dev become zero, it will destroy PD.
4. when con[1] failed to create, con[1] will not take clt_path->s.dev,
but it try to decreased clt_path->s.dev
So, in case create_cm(con[0]) succeeds but create_cm(con[1]) fails,
destroy_con_cq_qp(con[1]) will be called first which will destroy the PD
while this PD is still taken by con[0].
Here, we refactor the error path of create_cm() and init_conns(), so that
we do the cleanup in the order they are created.
The warning occurs when destroying RXE PD whose reference count is not
zero.
Chris Packham [Thu, 25 May 2023 00:31:53 +0000 (12:31 +1200)]
mtd: rawnand: marvell: don't set the NAND frequency select
marvell_nfc_setup_interface() uses the frequency retrieved from the
clock associated with the nand interface to determine the timings that
will be used. By changing the NAND frequency select without reflecting
this in the clock configuration this means that the timings calculated
don't correctly meet the requirements of the NAND chip. This hasn't been
an issue up to now because of a different bug that was stopping the
timings being updated after they were initially set.
Chris Packham [Thu, 25 May 2023 00:31:52 +0000 (12:31 +1200)]
mtd: rawnand: marvell: ensure timing values are written
When new timing values are calculated in marvell_nfc_setup_interface()
ensure that they will be applied in marvell_nfc_select_target() by
clearing the selected_chip pointer.
The addition of the mtdchar_read_ioctl() function caused the stack usage
of mtdchar_ioctl() to grow beyond the warning limit on 32-bit architectures
with gcc-13:
drivers/mtd/mtdchar.c: In function 'mtdchar_ioctl':
drivers/mtd/mtdchar.c:1229:1: error: the frame size of 1488 bytes is larger than 1024 bytes [-Werror=frame-larger-than=]
Mark both the read and write portions as noinline_for_stack to ensure
they don't get inlined and use separate stack slots to reduce the
maximum usage, both in the mtdchar_ioctl() and combined with any
of its callees.
Michal Simek [Fri, 26 May 2023 17:41:54 +0000 (19:41 +0200)]
MAINTAINERS: Add myself as reviewer instead of Naga
Naga no longer works for AMD/Xilinx and there is no activity from him to
continue to maintain Xilinx related drivers. Add myself instead to be kept
in loop if there is any need for testing.
These patches concern modifications made in mt8186[1]. The clock
unregistration mechanism used in mt8188 and mt8195 is similar with
mt8186, resulting in the same problem existing within the driver.
Therefore, the solution has also been applied to these two platforms.
Linus Torvalds [Thu, 1 Jun 2023 15:18:20 +0000 (11:18 -0400)]
Merge tag 'firewire-fixes-6.4-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394
Pull firewire fix from Takashi Sakamoto:
"A single patch to use a flexible array rather than a zero-length one"
* tag 'firewire-fixes-6.4-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394:
firewire: Replace zero-length array with flexible-array member
Linus Torvalds [Thu, 1 Jun 2023 15:13:10 +0000 (11:13 -0400)]
Merge tag 'mailbox-fixes-6.4-rc5' of git://git.linaro.org/landing-teams/working/fujitsu/integration
Pull mailbox fix from Jassi Brar:
"Fix missing mutex unlock in mailbox-test"
* tag 'mailbox-fixes-6.4-rc5' of git://git.linaro.org/landing-teams/working/fujitsu/integration:
mailbox: mailbox-test: fix a locking issue in mbox_test_message_write()
Avoid linker error for randomly generated config file that
has CONFIG_BRANCH_PROFILE_NONE enabled and make it similar
to riscv, x86 and also to commit 4bf3ec384edf ("s390: disable
branch profiling for vdso").
Andreas Svensson [Tue, 30 May 2023 14:52:23 +0000 (16:52 +0200)]
net: dsa: mv88e6xxx: Increase wait after reset deactivation
A switch held in reset by default needs to wait longer until we can
reliably detect it.
An issue was observed when testing on the Marvell 88E6393X (Link Street).
The driver failed to detect the switch on some upstarts. Increasing the
wait time after reset deactivation solves this issue.
The updated wait time is now also the same as the wait time in the
mv88e6xxx_hardware_reset function.
firewire: Replace zero-length array with flexible-array member
Zero-length and one-element arrays are deprecated, and we are moving
towards adopting C99 flexible-array members, instead.
Address the following warnings found with GCC-13 and
-fstrict-flex-arrays=3 enabled:
sound/firewire/amdtp-stream.c: In function ‘build_it_pkt_header’:
sound/firewire/amdtp-stream.c:694:17: warning: ‘generate_cip_header’ accessing 8 bytes in a region of size 0 [-Wstringop-overflow=]
694 | generate_cip_header(s, cip_header, data_block_counter, syt);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
sound/firewire/amdtp-stream.c:694:17: note: referencing argument 2 of type ‘__be32[2]’ {aka ‘unsigned int[2]’}
sound/firewire/amdtp-stream.c:667:13: note: in a call to function ‘generate_cip_header’
667 | static void generate_cip_header(struct amdtp_stream *s, __be32 cip_header[2],
| ^~~~~~~~~~~~~~~~~~~
This helps with the ongoing efforts to tighten the FORTIFY_SOURCE
routines on memcpy() and help us make progress towards globally
enabling -fstrict-flex-arrays=3 [1].
Ping-Ke Shih [Sat, 27 May 2023 08:29:39 +0000 (16:29 +0800)]
wifi: rtw89: remove redundant check of entering LPS
Originally, add this check rule to prevent entering LPS if more than one
vif (in station mode) connect to AP. Since we have checked this by previous
commit, remove this redundant check.
Ping-Ke Shih [Sat, 27 May 2023 08:29:38 +0000 (16:29 +0800)]
wifi: rtw89: correct PS calculation for SUPPORTS_DYNAMIC_PS
This driver relies on IEEE80211_CONF_PS of hw->conf.flags to turn off PS or
turn on dynamic PS controlled by driver and firmware. Though this would be
incorrect, it did work before because the flag is always recalculated until
the commit 28977e790b5d ("wifi: mac80211: skip powersave recalc if driver SUPPORTS_DYNAMIC_PS")
is introduced by kernel 5.20 to skip to recalculate IEEE80211_CONF_PS
of hw->conf.flags if driver sets SUPPORTS_DYNAMIC_PS.
Correct this by doing recalculation while BSS_CHANGED_PS is changed and
interface is added or removed. For now, it is allowed to enter PS only if
single one station vif is working, and it could possible to have PS per
vif after firmware can support it. Without this fix, driver doesn't
enter PS anymore that causes higher power consumption.
Ping-Ke Shih [Sat, 27 May 2023 08:29:37 +0000 (16:29 +0800)]
wifi: rtw88: correct PS calculation for SUPPORTS_DYNAMIC_PS
This driver relies on IEEE80211_CONF_PS of hw->conf.flags to turn off PS or
turn on dynamic PS controlled by driver and firmware. Though this would be
incorrect, it did work before because the flag is always recalculated until
the commit 28977e790b5d ("wifi: mac80211: skip powersave recalc if driver SUPPORTS_DYNAMIC_PS")
is introduced by kernel 5.20 to skip to recalculate IEEE80211_CONF_PS
of hw->conf.flags if driver sets SUPPORTS_DYNAMIC_PS.
Correct this by doing recalculation while BSS_CHANGED_PS is changed and
interface is added or removed. It is allowed to enter PS only if single
one station vif is working. Without this fix, driver doesn't enter PS
anymore that causes higher power consumption.
Qu Wenruo [Thu, 1 Jun 2023 10:51:34 +0000 (18:51 +0800)]
btrfs: zoned: fix dev-replace after the scrub rework
[BUG]
After commit e02ee89baa66 ("btrfs: scrub: switch scrub_simple_mirror()
to scrub_stripe infrastructure"), scrub no longer works for zoned device
at all.
Even an empty zoned btrfs cannot be replaced:
# mkfs.btrfs -f /dev/nvme0n1
# mount /dev/nvme0n1 /mnt/btrfs
# btrfs replace start -Bf 1 /dev/nvme0n2 /mnt/btrfs
Resetting device zones /dev/nvme1n1 (160 zones) ...
ERROR: ioctl(DEV_REPLACE_START) failed on "/mnt/btrfs/": Input/output error
And we can hit kernel crash related to that:
BTRFS info (device nvme1n1): host-managed zoned block device /dev/nvme3n1, 160 zones of 134217728 bytes
BTRFS info (device nvme1n1): dev_replace from /dev/nvme2n1 (devid 2) to /dev/nvme3n1 started
nvme3n1: Zone Management Append(0x7d) @ LBA 65536, 4 blocks, Zone Is Full (sct 0x1 / sc 0xb9) DNR
I/O error, dev nvme3n1, sector 786432 op 0xd:(ZONE_APPEND) flags 0x4000 phys_seg 3 prio class 2
BTRFS error (device nvme1n1): bdev /dev/nvme3n1 errs: wr 1, rd 0, flush 0, corrupt 0, gen 0
BUG: kernel NULL pointer dereference, address: 00000000000000a8
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
RIP: 0010:_raw_spin_lock_irqsave+0x1e/0x40
Call Trace:
<IRQ>
btrfs_lookup_ordered_extent+0x31/0x190
btrfs_record_physical_zoned+0x18/0x40
btrfs_simple_end_io+0xaf/0xc0
blk_update_request+0x153/0x4c0
blk_mq_end_request+0x15/0xd0
nvme_poll_cq+0x1d3/0x360
nvme_irq+0x39/0x80
__handle_irq_event_percpu+0x3b/0x190
handle_irq_event+0x2f/0x70
handle_edge_irq+0x7c/0x210
__common_interrupt+0x34/0xa0
common_interrupt+0x7d/0xa0
</IRQ>
<TASK>
asm_common_interrupt+0x22/0x40
[CAUSE]
Dev-replace reuses scrub code to iterate all extents and write the
existing content back to the new device.
And for zoned devices, we call fill_writer_pointer_gap() to make sure
all the writes into the zoned device is sequential, even if there may be
some gaps between the writes.
However we have several different bugs all related to zoned dev-replace:
- We are using ZONE_APPEND operation for metadata style write back
For zoned devices, btrfs has two ways to write data:
* ZONE_APPEND for data
This allows higher queue depth, but will not be able to know where
the write would land.
Thus needs to grab the real on-disk physical location in it's endio.
* WRITE for metadata
This requires single queue depth (new writes can only be submitted
after previous one finished), and all writes must be sequential.
For scrub, we go single queue depth, but still goes with ZONE_APPEND,
which requires btrfs_bio::inode being populated.
This is the cause of that crash.
- No correct tracing of write_pointer
After a write finished, we should forward sctx->write_pointer, or
fill_writer_pointer_gap() would not work properly and cause more
than necessary zero out, and fill the whole zone prematurely.
- Incorrect physical bytenr passed to fill_writer_pointer_gap()
In scrub_write_sectors(), one call site passes logical address, which
is completely wrong.
The other call site passes physical address of current sector, but
we should pass the physical address of the btrfs_bio we're submitting.
This is the cause of the -EIO errors.
[FIX]
- Do not use ZONE_APPEND for btrfs_submit_repair_write().
- Manually forward sctx->write_pointer after successful writeback
- Use the physical address of the to-be-submitted btrfs_bio for
fill_writer_pointer_gap()
Now zoned device replace would work as expected.
Reported-by: Christoph Hellwig <[email protected]> Fixes: e02ee89baa66 ("btrfs: scrub: switch scrub_simple_mirror() to scrub_stripe infrastructure") Reviewed-by: Christoph Hellwig <[email protected]> Signed-off-by: Qu Wenruo <[email protected]> Signed-off-by: David Sterba <[email protected]>
Linus Torvalds [Thu, 1 Jun 2023 13:02:04 +0000 (09:02 -0400)]
Merge tag 'for-linus-2023060101' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid
Pull HID fixes from Jiri Kosina:
- Regression fix for overlong long timeouts during initialization on
some Logitech Unifying devices (Bastien Nocera)
- error handling and overflow fixes for Wacom driver (Denis Arefev,
Jason Gerecke, Nikita Zhandarovich)
* tag 'for-linus-2023060101' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid:
HID: logitech-hidpp: Handle timeout differently from busy
HID: wacom: Add error check to wacom_parse_and_register()
HID: google: add jewel USB id
HID: wacom: avoid integer overflow in wacom_intuos_inout()
HID: wacom: Check for string overflow from strscpy calls
gfs2: Don't get stuck writing page onto itself under direct I/O
When a direct I/O write is performed, iomap_dio_rw() invalidates the
part of the page cache which the write is going to before carrying out
the write. In the odd case, the direct I/O write will be reading from
the same page it is writing to. gfs2 carries out writes with page
faults disabled, so it should have been obvious that this page
invalidation can cause iomap_dio_rw() to never make any progress.
Currently, gfs2 will end up in an endless retry loop in
gfs2_file_direct_write() instead, though.
Break this endless loop by limiting the number of retries and falling
back to buffered I/O after that.
Also simplify should_fault_in_pages() sightly and add a comment to make
the above case easier to understand.
Linus Torvalds [Thu, 1 Jun 2023 12:41:33 +0000 (08:41 -0400)]
Merge tag 'ata-6.4-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/libata
Pull ata fix from Damien Le Moal:
- Fix ata_find_dev() use of the device number to find a struct
ata_device for a port. This addresses issues with some passthrough
commands with libsas managed devices.
* tag 'ata-6.4-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/libata:
ata: libata-scsi: Use correct device no in ata_find_dev()
Linus Torvalds [Thu, 1 Jun 2023 12:27:34 +0000 (08:27 -0400)]
Merge tag '6.4-rc4-smb3-server-fixes' of git://git.samba.org/ksmbd
Pull smb server fixes from Steve French:
"Eight server fixes (most also for stable):
- Two fixes for uninitialized pointer reads (rename and link)
- Fix potential UAF in oplock break
- Two fixes for potential out of bound reads in negotiate
- Fix crediting bug
- Two fixes for xfstests (allocation size fix for test 694 and lookup
issue shown by test 464)"
* tag '6.4-rc4-smb3-server-fixes' of git://git.samba.org/ksmbd:
ksmbd: call putname after using the last component
ksmbd: fix incorrect AllocationSize set in smb2_get_info
ksmbd: fix UAF issue from opinfo->conn
ksmbd: fix multiple out-of-bounds read during context decoding
ksmbd: fix slab-out-of-bounds read in smb2_handle_negotiate
ksmbd: fix credit count leakage
ksmbd: fix uninitialized pointer read in smb2_create_link()
ksmbd: fix uninitialized pointer read in ksmbd_vfs_rename()
Trevor Wu [Thu, 1 Jun 2023 03:33:18 +0000 (11:33 +0800)]
ASoC: mediatek: mt8195: fix use-after-free in driver remove path
During mt8195_afe_init_clock(), mt8195_audsys_clk_register() was called
followed by several other devm functions. At mt8195_afe_deinit_clock()
located at mt8195_afe_pcm_dev_remove(), mt8195_audsys_clk_unregister()
was called.
However, there was an issue with the order in which these functions were
called. Specifically, the remove callback of platform_driver was called
before devres released the resource, resulting in a use-after-free issue
during remove time.
At probe time, the order of calls was:
1. mt8195_audsys_clk_register
2. afe_priv->clk = devm_kcalloc
3. afe_priv->clk[i] = devm_clk_get
At remove time, the order of calls was:
1. mt8195_audsys_clk_unregister
3. free afe_priv->clk[i]
2. free afe_priv->clk
To resolve the problem, we can utilize devm_add_action_or_reset() in
mt8195_audsys_clk_register() so that the remove order can be changed to
3->2->1.
Trevor Wu [Thu, 1 Jun 2023 03:33:17 +0000 (11:33 +0800)]
ASoC: mediatek: mt8188: fix use-after-free in driver remove path
During mt8188_afe_init_clock(), mt8188_audsys_clk_register() was called
followed by several other devm functions. The caller of
mt8188_afe_init_clock() utilized devm_add_action_or_reset() to call
mt8188_afe_deinit_clock(). However, the order was incorrect, causing a
use-after-free issue during remove time.
At probe time, the order of calls was:
1. mt8188_audsys_clk_register
2. afe_priv->clk = devm_kcalloc
3. afe_priv->clk[i] = devm_clk_get
At remove time, the order of calls was:
1. mt8188_audsys_clk_unregister
3. free afe_priv->clk[i]
2. free afe_priv->clk
To resolve the problem, it's necessary to move devm_add_action_or_reset()
to the appropriate position so that the remove order can be 3->2->1.
Bert Karwatzki [Wed, 31 May 2023 10:36:19 +0000 (12:36 +0200)]
net: ipa: Use correct value for IPA_STATUS_SIZE
IPA_STATUS_SIZE was introduced in commit b8dc7d0eea5a as a replacement
for the size of the removed struct ipa_status which had size
sizeof(__le32[8]). Use this value as IPA_STATUS_SIZE.
fuyuanli [Wed, 31 May 2023 08:01:50 +0000 (16:01 +0800)]
tcp: fix mishandling when the sack compression is deferred.
In this patch, we mainly try to handle sending a compressed ack
correctly if it's deferred.
Here are more details in the old logic:
When sack compression is triggered in the tcp_compressed_ack_kick(),
if the sock is owned by user, it will set TCP_DELACK_TIMER_DEFERRED
and then defer to the release cb phrase. Later once user releases
the sock, tcp_delack_timer_handler() should send a ack as expected,
which, however, cannot happen due to lack of ICSK_ACK_TIMER flag.
Therefore, the receiver would not sent an ack until the sender's
retransmission timeout. It definitely increases unnecessary latency.
Hangyu Hua [Wed, 31 May 2023 10:28:04 +0000 (18:28 +0800)]
net/sched: flower: fix possible OOB write in fl_set_geneve_opt()
If we send two TCA_FLOWER_KEY_ENC_OPTS_GENEVE packets and their total
size is 252 bytes(key->enc_opts.len = 252) then
key->enc_opts.len = opt->length = data_len / 4 = 0 when the third
TCA_FLOWER_KEY_ENC_OPTS_GENEVE packet enters fl_set_geneve_opt. This
bypasses the next bounds check and results in an out-of-bounds.
Chen-Yu Tsai [Fri, 26 May 2023 08:53:59 +0000 (16:53 +0800)]
iommu/mediatek: Flush IOTLB completely only if domain has been attached
If an IOMMU domain was never attached, it lacks any linkage to the
actual IOMMU hardware. Attempting to do flush_iotlb_all() on it will
result in a NULL pointer dereference. This seems to happen after the
recent IOMMU core rework in v6.4-rc1.
Unable to handle kernel read from unreadable memory at virtual address 0000000000000018
Call trace:
mtk_iommu_flush_iotlb_all+0x20/0x80
iommu_create_device_direct_mappings.part.0+0x13c/0x230
iommu_setup_default_domain+0x29c/0x4d0
iommu_probe_device+0x12c/0x190
of_iommu_configure+0x140/0x208
of_dma_configure_id+0x19c/0x3c0
platform_dma_configure+0x38/0x88
really_probe+0x78/0x2c0
Check if the "bank" field has been filled in before actually attempting
the IOTLB flush to avoid it. The IOTLB is also flushed when the device
comes out of runtime suspend, so it should have a clean initial state.
Ashutosh Dixit [Tue, 23 May 2023 20:40:42 +0000 (13:40 -0700)]
drm/i915/perf: Clear out entire reports after reading if not power of 2 size
Clearing out report id and timestamp as means to detect unlanded reports
only works if report size is power of 2. That is, only when report size is
a sub-multiple of the OA buffer size can we be certain that reports will
land at the same place each time in the OA buffer (after rewind). If report
size is not a power of 2, we need to zero out the entire report to be able
to detect unlanded reports reliably.
Edward Cree [Tue, 30 May 2023 20:25:27 +0000 (21:25 +0100)]
sfc: fix error unwinds in TC offload
Failure ladders weren't exactly unwinding what the function had done up
to that point; most seriously, when we encountered an already offloaded
rule, the failure path tried to remove the new rule from the hashtable,
which would in fact remove the already-present 'old' rule (since it has
the same key) from the table, and leak its resources.
net/mlx5: Read embedded cpu after init bit cleared
During driver load it reads embedded_cpu bit from initialization
segment, but the initialization segment is readable only after
initialization bit is cleared.
Move the call to mlx5_read_embedded_cpu() right after initialization bit
cleared.
The crash is due to matrix_alloc_area() attempting to access per-CPU
memory for CPUs that are not present on the system. The CPU mask
passed into reserve_managed_vector() via it's @irqd parameter is
corrupted because it contains uninitialized stack data.
Fixes: bbac70c74183 ("net/mlx5: Use newer affinity descriptor") Reviewed-by: Thomas Gleixner <[email protected]> Signed-off-by: Chuck Lever <[email protected]> Signed-off-by: Saeed Mahameed <[email protected]>
Niklas Schnelle [Wed, 31 May 2023 08:48:56 +0000 (10:48 +0200)]
net/mlx5: Fix setting of irq->map.index for static IRQ case
When dynamic IRQ allocation is not supported all IRQs are allocated up
front in mlx5_irq_table_create() instead of dynamically as part of
mlx5_irq_alloc(). In the latter dynamic case irq->map.index is set
via the mapping returned by pci_msix_alloc_irq_at(). In the static case
and prior to commit 1da438c0ae02 ("net/mlx5: Fix indexing of mlx5_irq")
irq->map.index was set in mlx5_irq_alloc() twice once initially to 0 and
then to the requested index before storing in the xarray. After this
commit it is only set to 0 which breaks all other IRQ mappings.
Fix this by setting irq->map.index to the requested index together with
irq->map.virq and improve the related comment to make it clearer which
cases it deals with.
Shay Drory [Tue, 30 May 2023 08:59:34 +0000 (11:59 +0300)]
net/mlx5: Remove rmap also in case dynamic MSIX not supported
mlx5 add IRQs to rmap upon MSIX request, and mlx5 remove rmap from
MSIX only if msi_map.index is populated. However, msi_map.index is
populated only when dynamic MSIX is supported. This results in freeing
IRQs without removing them from rmap, which triggers the bellow
WARN_ON[1].
rmap is a feature which have no relation to dynamic MSIX.
Hence, remove the check of msi_map.index when removing IRQ from rmap.
Linus Torvalds [Wed, 31 May 2023 23:24:01 +0000 (19:24 -0400)]
Merge tag '6.4-rc4-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6
Pull smb client fixes from Steve French:
"Four small smb3 client fixes:
- two small fixes suggested by kernel test robot
- small cleanup fix
- update Paulo's email address in the maintainer file"
* tag '6.4-rc4-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
cifs: address unused variable warning
smb: delete an unnecessary statement
smb3: missing null check in SMB2_change_notify
smb3: update a reviewer email in MAINTAINERS file
Samsung Odyssey Neo G9, running at 5120x1440@240/VRR, connected to Navi
21 via DisplayPort, blanks and the GPU hangs while starting the Steam
game Assetto Corsa Competizione (via Proton 7.0).
Tim Huang [Mon, 22 May 2023 15:17:28 +0000 (23:17 +0800)]
drm/amd/pm: reverse mclk and fclk clocks levels for renoir
This patch reverses the DPM clocks levels output of pp_dpm_mclk
and pp_dpm_fclk for renoir.
On dGPUs and older APUs we expose the levels from lowest clocks
to highest clocks. But for some APUs, the clocks levels are
given the reversed orders by PMFW. Like the memory DPM clocks
that are exposed by pp_dpm_mclk.
It's not intuitive that they are reversed on these APUs. All tools
and software that talks to the driver then has to know different ways
to interpret the data depending on the asic.
So we need to reverse them to expose the clocks levels from the
driver consistently.
Tim Huang [Sun, 21 May 2023 03:10:19 +0000 (11:10 +0800)]
drm/amd/pm: reverse mclk and fclk clocks levels for vangogh
This patch reverses the DPM clocks levels output of pp_dpm_mclk
and pp_dpm_fclk.
On dGPUs and older APUs we expose the levels from lowest clocks
to highest clocks. But for some APUs, the clocks levels that from
the DFPstateTable are given the reversed orders by PMFW. Like the
memory DPM clocks that are exposed by pp_dpm_mclk.
It's not intuitive that they are reversed on these APUs. All tools
and software that talks to the driver then has to know different ways
to interpret the data depending on the asic.
So we need to reverse them to expose the clocks levels from the
driver consistently.
Tim Huang [Sun, 21 May 2023 02:35:59 +0000 (10:35 +0800)]
drm/amd/pm: reverse mclk and fclk clocks levels for yellow carp
This patch reverses the DPM clocks levels output of pp_dpm_mclk
and pp_dpm_fclk.
On dGPUs and older APUs we expose the levels from lowest clocks
to highest clocks. But for some APUs, the clocks levels that from
the DFPstateTable are given the reversed orders by PMFW. Like the
memory DPM clocks that are exposed by pp_dpm_mclk.
It's not intuitive that they are reversed on these APUs. All tools
and software that talks to the driver then has to know different ways
to interpret the data depending on the asic.
So we need to reverse them to expose the clocks levels from the
driver consistently.
Tim Huang [Sun, 21 May 2023 02:28:05 +0000 (10:28 +0800)]
drm/amd/pm: reverse mclk clocks levels for SMU v13.0.5
This patch reverses the DPM clocks levels output of pp_dpm_mclk.
On dGPUs and older APUs we expose the levels from lowest clocks
to highest clocks. But for some APUs, the clocks levels that from
the DFPstateTable are given the reversed orders by PMFW. Like the
memory DPM clocks that are exposed by pp_dpm_mclk.
It's not intuitive that they are reversed on these APUs. All tools
and software that talks to the driver then has to know different ways
to interpret the data depending on the asic.
So we need to reverse them to expose the clocks levels from the
driver consistently.
Tim Huang [Sun, 21 May 2023 01:24:00 +0000 (09:24 +0800)]
drm/amd/pm: reverse mclk and fclk clocks levels for SMU v13.0.4
This patch reverses the DPM clocks levels output of pp_dpm_mclk
and pp_dpm_fclk.
On dGPUs and older APUs we expose the levels from lowest clocks
to highest clocks. But for some APUs, the clocks levels that from
the DFPstateTable are given the reversed orders by PMFW. Like the
memory DPM clocks that are exposed by pp_dpm_mclk.
It's not intuitive that they are reversed on these APUs. All tools
and software that talks to the driver then has to know different ways
to interpret the data depending on the asic.
So we need to reverse them to expose the clocks levels from the
driver consistently.
K Prateek Nayak [Mon, 8 May 2023 08:41:15 +0000 (14:11 +0530)]
drivers: base: cacheinfo: Update cpu_map_populated during CPU Hotplug
Until commit 5c2712387d48 ("cacheinfo: Fix LLC is not exported through
sysfs"), cacheinfo called populate_cache_leaves() for CPU coming online
which let the arch specific functions handle (at least on x86)
populating the shared_cpu_map. However, with the changes in the
aforementioned commit, populate_cache_leaves() is not called when a CPU
comes online as a result of hotplug since last_level_cache_is_valid()
returns true as the cacheinfo data is not discarded. The CPU coming
online is not present in shared_cpu_map, however, it will not be added
since the cpu_cacheinfo->cpu_map_populated flag is set (it is set in
populate_cache_leaves() when cacheinfo is first populated for x86)
This can lead to inconsistencies in the shared_cpu_map when an offlined
CPU comes online again. Example below depicts the inconsistency in the
shared_cpu_list in cacheinfo when CPU8 is offlined and onlined again on
a 3rd Generation EPYC processor:
# for i in /sys/devices/system/cpu/cpu8/cache/index*/shared_cpu_list; do echo -n "$i: "; cat $i; done
/sys/devices/system/cpu/cpu8/cache/index0/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index1/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index2/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index3/shared_cpu_list: 8-15,136-143
Clear the flag when the CPU is removed from shared_cpu_map when
cache_shared_cpu_map_remove() is called during CPU hotplug. This will
allow cache_shared_cpu_map_setup() to add the CPU coming back online in
the shared_cpu_map. Set the flag again when the shared_cpu_map is setup.
Following are results of performing the same test as described above with
the changes:
# for i in /sys/devices/system/cpu/cpu8/cache/index*/shared_cpu_list; do echo -n "$i: "; cat $i; done
/sys/devices/system/cpu/cpu8/cache/index0/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index1/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index2/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index3/shared_cpu_list: 8-15,136-143
K Prateek Nayak [Mon, 8 May 2023 08:41:14 +0000 (14:11 +0530)]
drivers: base: cacheinfo: Fix shared_cpu_map changes in event of CPU hotplug
While building the shared_cpu_map, check if the cache level and cache
type matches. On certain systems that build the cache topology based on
the instance ID, there are cases where the same ID may repeat across
multiple cache levels, leading inaccurate topology.
In event of CPU offlining, the cache_shared_cpu_map_remove() does not
consider if IDs at same level are being compared. As a result, when same
IDs repeat across different cache levels, the CPU going offline is not
removed from all the shared_cpu_map.
Below is the output of cache topology of CPU8 and it's SMT sibling after
CPU8 is offlined on a dual socket 3rd Generation AMD EPYC processor
(2 x 64C/128T) running kernel release v6.3:
# for i in /sys/devices/system/cpu/cpu8/cache/index*/shared_cpu_list; do echo -n "$i: "; cat $i; done
/sys/devices/system/cpu/cpu8/cache/index0/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index1/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index2/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index3/shared_cpu_list: 8-15,136-143
# echo 0 > /sys/devices/system/cpu/cpu8/online
# for i in /sys/devices/system/cpu/cpu136/cache/index*/shared_cpu_list; do echo -n "$i: "; cat $i; done
/sys/devices/system/cpu/cpu136/cache/index0/shared_cpu_list: 136
/sys/devices/system/cpu/cpu136/cache/index1/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu136/cache/index2/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu136/cache/index3/shared_cpu_list: 9-15,136-143
CPU8 is removed from index0 (L1i) but remains in the shared_cpu_list of
index1 (L1d) and index2 (L2). Since L1i, L1d, and L2 are shared by the
SMT siblings, and they have the same cache instance ID, CPU 2 is only
removed from the first index with matching ID which is index1 (L1i) in
this case. With this fix, the results are as expected when performing
the same experiment on the same system:
# for i in /sys/devices/system/cpu/cpu8/cache/index*/shared_cpu_list; do echo -n "$i: "; cat $i; done
/sys/devices/system/cpu/cpu8/cache/index0/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index1/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index2/shared_cpu_list: 8,136
/sys/devices/system/cpu/cpu8/cache/index3/shared_cpu_list: 8-15,136-143
# echo 0 > /sys/devices/system/cpu/cpu8/online
# for i in /sys/devices/system/cpu/cpu136/cache/index*/shared_cpu_list; do echo -n "$i: "; cat $i; done
/sys/devices/system/cpu/cpu136/cache/index0/shared_cpu_list: 136
/sys/devices/system/cpu/cpu136/cache/index1/shared_cpu_list: 136
/sys/devices/system/cpu/cpu136/cache/index2/shared_cpu_list: 136
/sys/devices/system/cpu/cpu136/cache/index3/shared_cpu_list: 9-15,136-143
When rebuilding topology, the same problem appears as
cache_shared_cpu_map_setup() implements a similar logic. Consider the
same 3rd Generation EPYC processor: CPUs in Core 1, that share the L1
and L2 caches, have L1 and L2 instance ID as 1. For all the CPUs on
the second chiplet, the L3 ID is also 1 leading to grouping on CPUs from
Core 1 (1, 17) and the entire second chiplet (8-15, 24-31) as CPUs
sharing one cache domain. This went undetected since x86 processors
depended on arch specific populate_cache_leaves() method to repopulate
the shared_cpus_map when CPU came back online until kernel release
v6.3-rc5.
Note that the size 1024 corresponds to the size of the test firmware
buffer. The actual number of the buffers leaked is around 70-110,
depending on the test run.
The cause of the leak is the following:
request_partial_firmware_into_buf() and request_firmware_into_buf()
provided firmware buffer isn't released on release_firmware(), we
have allocated it and we are responsible for deallocating it manually.
This is introduced in a number of context where previously only
release_firmware() was called, which was insufficient.
Dan Carpenter spotted that test_fw_config->reqs will be leaked if
trigger_batched_requests_store() is called two or more times.
The same appears with trigger_batched_requests_async_store().
This bug wasn't trigger by the tests, but observed by Dan's visual
inspection of the code.
The recommended workaround was to return -EBUSY if test_fw_config->reqs
is already allocated.
The function test_dev_config_update_u8() is called from both the locked
and the unlocked context, function config_num_requests_store() and
config_read_fw_idx_store() which can both be called asynchronously as
they are driver's methods, while test_dev_config_update_u8() and siblings
change their argument pointed to by u8 *cfg or similar pointer.
To avoid deadlock on test_fw_mutex, the lock is dropped before calling
test_dev_config_update_u8() and re-acquired within test_dev_config_update_u8()
itself, but alas this creates a race condition.
Having two locks wouldn't assure a race-proof mutual exclusion.
This situation is best avoided by the introduction of a new, unlocked
function __test_dev_config_update_u8() which can be called from the locked
context and reducing test_dev_config_update_u8() to:
static int test_dev_config_update_u8(const char *buf, size_t size, u8 *cfg)
{
int ret;
mutex_lock(&test_fw_mutex);
ret = __test_dev_config_update_u8(buf, size, cfg);
mutex_unlock(&test_fw_mutex);
return ret;
}
doing the locking and calling the unlocked primitive, which enables both
locked and unlocked versions without duplication of code.
The similar approach was applied to all functions called from the locked
and the unlocked context, which safely mitigates both deadlocks and race
conditions in the driver.
__test_dev_config_update_bool(), __test_dev_config_update_u8() and
__test_dev_config_update_size_t() unlocked versions of the functions
were introduced to be called from the locked contexts as a workaround
without releasing the main driver's lock and thereof causing a race
condition.
The test_dev_config_update_bool(), test_dev_config_update_u8() and
test_dev_config_update_size_t() locked versions of the functions
are being called from driver methods without the unnecessary multiplying
of the locking and unlocking code for each method, and complicating
the code with saving of the return value across lock.
Linus Torvalds [Wed, 31 May 2023 18:13:09 +0000 (14:13 -0400)]
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
Pull rdma fixes from Jason Gunthorpe:
- Fix 64K ARM page size support in bnxt_re and efa
- bnxt_re fixes for a memory leak, incorrect error handling and a
remove a bogus FW failure when running on a VF
- Update MAINTAINERS for hns and efa
- Fix two rxe regressions added this merge window in error unwind and
incorrect spinlock primitives
- hns gets a better algorithm for allocating page tables to avoid
running out of resources, and a timeout adjustment
- Fix a text case failure in hns
- Use after free in irdma and fix incorrect construction of a WQE
causing mis-execution
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
RDMA/irdma: Fix Local Invalidate fencing
RDMA/irdma: Prevent QP use after free
MAINTAINERS: Update maintainer of Amazon EFA driver
RDMA/bnxt_re: Do not enable congestion control on VFs
RDMA/bnxt_re: Fix return value of bnxt_re_process_raw_qp_pkt_rx
RDMA/bnxt_re: Fix a possible memory leak
RDMA/hns: Modify the value of long message loopback slice
RDMA/hns: Fix base address table allocation
RDMA/hns: Fix timeout attr in query qp for HIP08
RDMA/efa: Fix unsupported page sizes in device
RDMA/rxe: Convert spin_{lock_bh,unlock_bh} to spin_{lock_irqsave,unlock_irqrestore}
RDMA/rxe: Fix double unlock in rxe_qp.c
MAINTAINERS: Update maintainers of HiSilicon RoCE
RDMA/bnxt_re: Fix the page_size used during the MR creation
Linus Torvalds [Wed, 31 May 2023 18:06:01 +0000 (14:06 -0400)]
Merge tag 'ext4_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4
Pull ext4 fixes from Ted Ts'o:
"Fix two regressions in ext4 and a number of issues reported by syzbot"
* tag 'ext4_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
ext4: enable the lazy init thread when remounting read/write
ext4: fix fsync for non-directories
ext4: add lockdep annotations for i_data_sem for ea_inode's
ext4: disallow ea_inodes with extended attributes
ext4: set lockdep subclass for the ea_inode in ext4_xattr_inode_cache_find()
ext4: add EA_INODE checking to ext4_iget()
Bart Van Assche [Mon, 29 May 2023 19:50:34 +0000 (12:50 -0700)]
scsi: stex: Fix gcc 13 warnings
gcc 13 may assign another type to enumeration constants than gcc 12. Split
the large enum at the top of source file stex.c such that the type of the
constants used in time expressions is changed back to the same type chosen
by gcc 12. This patch suppresses compiler warnings like this one:
In file included from ./include/linux/bitops.h:7,
from ./include/linux/kernel.h:22,
from drivers/scsi/stex.c:13:
drivers/scsi/stex.c: In function ‘stex_common_handshake’:
./include/linux/typecheck.h:12:25: error: comparison of distinct pointer types lacks a cast [-Werror]
12 | (void)(&__dummy == &__dummy2); \
| ^~
./include/linux/jiffies.h:106:10: note: in expansion of macro ‘typecheck’
106 | typecheck(unsigned long, b) && \
| ^~~~~~~~~
drivers/scsi/stex.c:1035:29: note: in expansion of macro ‘time_after’
1035 | if (time_after(jiffies, before + MU_MAX_DELAY * HZ)) {
| ^~~~~~~~~~
See also https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107405.
Bastien Nocera [Wed, 31 May 2023 08:24:28 +0000 (10:24 +0200)]
HID: logitech-hidpp: Handle timeout differently from busy
If an attempt at contacting a receiver or a device fails because the
receiver or device never responds, don't restart the communication, only
restart it if the receiver or device answers that it's busy, as originally
intended.
This was the behaviour on communication timeout before commit 586e8fede795
("HID: logitech-hidpp: Retry commands when device is busy").
This fixes some overly long waits in a critical path on boot, when
checking whether the device is connected by getting its HID++ version.
Alexandre Ghiti [Fri, 26 May 2023 15:46:30 +0000 (17:46 +0200)]
riscv: Fix relocatable kernels with early alternatives using -fno-pie
Early alternatives are called with the mmu disabled, and then should not
access any global symbols through the GOT since it requires relocations,
relocations that we do before but *virtually*. So only use medany code
model for this early code.
Dan Carpenter [Mon, 29 May 2023 11:35:55 +0000 (14:35 +0300)]
nfsd: fix double fget() bug in __write_ports_addfd()
The bug here is that you cannot rely on getting the same socket
from multiple calls to fget() because userspace can influence
that. This is a kind of double fetch bug.
The fix is to delete the svc_alien_sock() function and instead do
the checking inside the svc_addsock() function.
Jammy Huang [Tue, 30 May 2023 04:12:40 +0000 (12:12 +0800)]
drm/ast: Fix long time waiting on s3/s4 resume
In resume, DP's launch function, ast_dp_launch, could wait at most 30
seconds before timeout to check if DP is enabled. It could lead to 'DPM
device timeout' and trigger unrecoverable kernel panic.
To avoid this problem, we check if DP enable or not at driver probe only.
All callers of trace_probe_primary_from_call() check the return
value to be non NULL. However, the function returns
list_first_entry(&tpe->probes, ...) which can never be NULL.
Additionally, it does not check for the list being possibly empty,
possibly causing a type confusion on empty lists.
Use list_first_entry_or_null() which solves both problems.
Vladislav Efanov [Tue, 30 May 2023 11:39:41 +0000 (14:39 +0300)]
udp6: Fix race condition in udp6_sendmsg & connect
Syzkaller got the following report:
BUG: KASAN: use-after-free in sk_setup_caps+0x621/0x690 net/core/sock.c:2018
Read of size 8 at addr ffff888027f82780 by task syz-executor276/3255
The function sk_setup_caps (called by ip6_sk_dst_store_flow->
ip6_dst_store) referenced already freed memory as this memory was
freed by parallel task in udpv6_sendmsg->ip6_sk_dst_lookup_flow->
sk_dst_check.
Oliver Upton [Thu, 25 May 2023 21:27:22 +0000 (21:27 +0000)]
KVM: arm64: Document default vPMU behavior on heterogeneous systems
KVM maintains a mask of supported CPUs when a vPMU type is explicitly
selected by userspace and is used to reject any attempt to run the vCPU
on an unsupported CPU. This is great, but we're still beholden to the
default behavior where vCPUs can be scheduled anywhere and guest
counters may silently stop working.
Avoid confusing the next poor sod to look at this code and document the
intended behavior.
Oliver Upton [Thu, 25 May 2023 21:27:21 +0000 (21:27 +0000)]
KVM: arm64: Iterate arm_pmus list to probe for default PMU
To date KVM has relied on using a perf event to probe the core PMU at
the time of vPMU initialization. Behind the scenes perf_event_init()
would iteratively walk the PMUs of the system and return the PMU that
could handle the event. However, an upcoming change in perf core will
drop the iterative walk, thereby breaking the fragile dance we do on the
KVM side.
Avoid the problem altogether by iterating over the list of supported
PMUs maintained in KVM, returning the core PMU that matches the CPU
we were called on.
Mika Westerberg [Tue, 30 May 2023 05:48:29 +0000 (08:48 +0300)]
thunderbolt: Mask ring interrupt on Intel hardware as well
When resuming from system sleep states the driver issues following
warning on Intel hardware:
thunderbolt 0000:07:00.0: interrupt for TX ring 0 is already enabled
The reason for this is that the commit in question did not mask the ring
interrupt on Intel hardware leaving the interrupt active. Fix this by
masking it also in Intel hardware.
The current code for the length calculation wrongly truncates the reported
length of the groups array, causing an under report of the subscribed
groups. To fix this, use 'BITS_TO_BYTES()' which rounds up the
division by 8.
Oliver Upton [Tue, 30 May 2023 19:32:13 +0000 (19:32 +0000)]
KVM: arm64: Drop last page ref in kvm_pgtable_stage2_free_removed()
The reference count on page table allocations is increased for every
'counted' PTE (valid or donated) in the table in addition to the initial
reference from ->zalloc_page(). kvm_pgtable_stage2_free_removed() fails
to drop the last reference on the root of the table walk, meaning we
leak memory.
Fix it by dropping the last reference after the free walker returns,
at which point all references for 'counted' PTEs have been released.
Zhengchao Shao [Sat, 27 May 2023 09:37:47 +0000 (17:37 +0800)]
net: sched: fix NULL pointer dereference in mq_attach
When use the following command to test:
1)ip link add bond0 type bond
2)ip link set bond0 up
3)tc qdisc add dev bond0 root handle ffff: mq
4)tc qdisc replace dev bond0 parent ffff:fff1 handle ffff: mq
This is because when mq is added for the first time, qdiscs in mq is set
to NULL in mq_attach(). Therefore, when replacing mq after adding mq, we
need to initialize qdiscs in the mq before continuing to graft. Otherwise,
it will couse NULL pointer dereference issue in mq_attach(). And the same
issue will occur in the attach functions of mqprio, taprio and htb.
ffff:fff1 means that the repalce qdisc is ingress. Ingress does not allow
any qdisc to be attached. Therefore, ffff:fff1 is incorrectly used, and
the command should be dropped.
====================
net/sched: Fixes for sch_ingress and sch_clsact
These are v6 fixes for ingress and clsact Qdiscs, including only first 4
patches (already tested and reviewed) from v5. Patch 5 and 6 from
previous versions are still under discussion and will be sent separately.
Peilin Ye [Mon, 29 May 2023 19:54:26 +0000 (12:54 -0700)]
net/sched: Prohibit regrafting ingress or clsact Qdiscs
Currently, after creating an ingress (or clsact) Qdisc and grafting it
under TC_H_INGRESS (TC_H_CLSACT), it is possible to graft it again under
e.g. a TBF Qdisc:
$ ip link add ifb0 type ifb
$ tc qdisc add dev ifb0 handle 1: root tbf rate 20kbit buffer 1600 limit 3000
$ tc qdisc add dev ifb0 clsact
$ tc qdisc link dev ifb0 handle ffff: parent 1:1
$ tc qdisc show dev ifb0
qdisc tbf 1: root refcnt 2 rate 20Kbit burst 1600b lat 560.0ms
qdisc clsact ffff: parent ffff:fff1 refcnt 2
^^^^^^^^
clsact's refcount has increased: it is now grafted under both
TC_H_CLSACT and 1:1.
ingress and clsact Qdiscs should only be used under TC_H_INGRESS
(TC_H_CLSACT). Prohibit regrafting them.
Peilin Ye [Mon, 29 May 2023 19:54:03 +0000 (12:54 -0700)]
net/sched: Reserve TC_H_INGRESS (TC_H_CLSACT) for ingress (clsact) Qdiscs
Currently it is possible to add e.g. an HTB Qdisc under ffff:fff1
(TC_H_INGRESS, TC_H_CLSACT):
$ ip link add name ifb0 type ifb
$ tc qdisc add dev ifb0 parent ffff:fff1 htb
$ tc qdisc add dev ifb0 clsact
Error: Exclusivity flag on, cannot modify.
$ drgn
...
>>> ifb0 = netdev_get_by_name(prog, "ifb0")
>>> qdisc = ifb0.ingress_queue.qdisc_sleeping
>>> print(qdisc.ops.id.string_().decode())
htb
>>> qdisc.flags.value_() # TCQ_F_INGRESS
2
Only allow ingress and clsact Qdiscs under ffff:fff1. Return -EINVAL
for everything else. Make TCQ_F_INGRESS a static flag of ingress and
clsact Qdiscs.
projects / linux.git / log