[ Upstream commit 64057f051f20c2a2184b9db7f8037d928d68a4f4 ]
Serdev recv_buf() callback is supposed to return the amount of bytes
consumed, therefore an int in between 0 and count.
Do not return negative number in case of issue, just print an error and
return count. This fixes a WARN in ttyport_receive_buf().
Link: https://lore.kernel.org/all/087be419-ec6b-47ad-851a-5e1e3ea5cfcc@kernel.org/
Fixes: 7237c4c9ec ("Bluetooth: mediatek: Add protocol support for MediaTek serial devices")
Signed-off-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit d03376c185926098cb4d668d6458801eb785c0a5 ]
This reverts 19f8def031
"Bluetooth: Fix auth_complete_evt for legacy units" which seems to be
working around a bug on a broken controller rather then any limitation
imposed by the Bluetooth spec, in fact if there ws not possible to
re-auth the command shall fail not succeed.
Fixes: 19f8def031 ("Bluetooth: Fix auth_complete_evt for legacy units")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 08e4c8c5919fd405a4d709b4ba43d836894a26eb ]
If a transaction is aborted, we should mark the to-be-released NEWSET dead,
just like commit path does for DEL and DESTROYSET commands.
In both cases all remaining elements will be released via
set->ops->destroy().
The existing abort code does NOT post the actual release to the work queue.
Also the entire __nf_tables_abort() function is wrapped in gc_seq
begin/end pair.
Therefore, async gc worker will never try to release the pending set
elements, as gc sequence is always stale.
It might be possible to speed up transaction aborts via work queue too,
this would result in a race and a possible use-after-free.
So fix this before it becomes an issue.
Fixes: 5f68718b34 ("netfilter: nf_tables: GC transaction API to avoid race with control plane")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 52221dfddbbfb5b4e029bb2efe9bb7da33ec1e46 ]
There are many same functions like _rtl88e_phy_calculate_bit_shift(),
_rtl92c_phy_calculate_bit_shift() and so on. And these functions can
cause undefined bitwise shift behavior. Add calculate_bit_shift() to
replace them and fix undefined behavior in subsequent patches.
Signed-off-by: Su Hui <suhui@nfschina.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20231219065739.1895666-2-suhui@nfschina.com
Stable-dep-of: 969bc926f04b ("wifi: rtlwifi: rtl8188ee: phy: using calculate_bit_shift()")
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 08f399a818 ]
This is not a fix on its own but more a cleanup. Phy qmp pcie driver
currently have a workaround to handle pcs_misc not declared and add
0x400 offset to the pcs reg if pcs_misc is not declared.
Correctly declare pcs_misc reg and reduce PCS size to the common value
of 0x1f0 as done for every other qmp based pcie phy device.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221103212125.17156-2-ansuelsmth@gmail.com
Stable-dep-of: 5c0dbe8b0584 ("arm64: dts: qcom: ipq6018: fix clock rates for GCC_USB0_MOCK_UTMI_CLK")
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 4c434392c4777881d01beada6701eff8c76b43fe ]
'first_minor' represents the starting minor number of disks, and
'minors' represents the number of partitions in the device. Neither
of them can be greater than MINORMASK + 1.
Commit e338924bd0 ("block: check minor range in device_add_disk()")
only added the check of 'first_minor + minors'. However, their sum might
be less than MINORMASK but their values are wrong. Complete the checks now.
Fixes: e338924bd0 ("block: check minor range in device_add_disk()")
Signed-off-by: Li Nan <linan122@huawei.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20231219075942.840255-1-linan666@huaweicloud.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a509adf05b2aac31b22781f5aa09e4768a5b6c39 ]
The SM8150-HDK uses two different regulators to power up SuperSpeed USB
PHYs. The L5A regulator is used for the second USB host, while the first
(OTG) USB host uses different regulator, L18A. Fix the regulator for the
usb_1 QMPPHY and (to remove possible confusion) drop the
usb_ss_dp_core_1/_2 labels.
Fixes: 0ab1b2d10a ("arm64: dts: qcom: add sm8150 hdk dts")
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20231215174152.315403-4-dmitry.baryshkov@linaro.org
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit eed6e57e9f3e2beac37563eb6a0129549daa330e ]
Commit c14e64b469 ("soc: qcom: llcc: Support chipsets that can
write to llcc") add the support for chipset where capacity based
allocation and retention through power collapse can be programmed
based on content of SCT table mentioned in the llcc driver where
the target like sdm845 where the entire programming related to it
is controlled in firmware. However, the commit introduces a bug
where capacity/retention register get overwritten each time it
gets programmed for each slice and that results in misconfiguration
of the register based on SCT table and that is not expected
behaviour instead it should be read modify write to retain the
configuration of other slices.
This issue is totally caught from code review and programming test
and not through any power/perf numbers so, it is not known what
impact this could make if we don't have this change however,
this feature are for these targets and they should have been
programmed accordingly as per their configuration mentioned in
SCT table like others bits information.
This change brings one difference where it keeps capacity/retention
bits of the slices that are not mentioned in SCT table in unknown
state where as earlier it was initialized to zero.
Fixes: c14e64b469 ("soc: qcom: llcc: Support chipsets that can write to llcc")
Signed-off-by: Atul Dhudase <quic_adhudase@quicinc.com>
Signed-off-by: Mukesh Ojha <quic_mojha@quicinc.com>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Link: https://lore.kernel.org/r/1701876771-10695-1-git-send-email-quic_mojha@quicinc.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit b07bc2347672cc8c7293c64499f1488278c5ca3d ]
Reproduced with below sequence:
dma_declare_coherent_memory()->dma_release_coherent_memory()
->dma_declare_coherent_memory()->"return -EBUSY" error
It will return -EBUSY from the dma_assign_coherent_memory()
in dma_declare_coherent_memory(), the reason is that dev->dma_mem
pointer has not been set to NULL after it's freed.
Fixes: cf65a0f6f6 ("dma-mapping: move all DMA mapping code to kernel/dma")
Signed-off-by: Joakim Zhang <joakim.zhang@cixtech.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 93b80887668226180ea5f5349cc728ca6dc700ab ]
Add one more condition for sending credit update during dequeue from
stream socket: when number of bytes in the rx queue is smaller than
SO_RCVLOWAT value of the socket. This is actual for non-default value
of SO_RCVLOWAT (e.g. not 1) - idea is to "kick" peer to continue data
transmission, because we need at least SO_RCVLOWAT bytes in our rx
queue to wake up user for reading data (in corner case it is also
possible to stuck both tx and rx sides, this is why 'Fixes' is used).
Fixes: b89d882dc9 ("vsock/virtio: reduce credit update messages")
Signed-off-by: Arseniy Krasnov <avkrasnov@salutedevices.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit bb7403655b3c3eb245d0ee330047cd3e20b3c4af ]
In order to support IP_PKTINFO on those packets, we need to call
ipv4_pktinfo_prepare.
When sending mrouted/pimd daemons a cache report IGMP msg, it is
unnecessary to set dst on the newly created skb.
It used to be necessary on older versions until
commit d826eb14ec ("ipv4: PKTINFO doesnt need dst reference") which
changed the way IP_PKTINFO struct is been retrieved.
Changes from v1:
1. Undo changes in ipv4_pktinfo_prepare function. use it directly
and copy the control block.
Fixes: d826eb14ec ("ipv4: PKTINFO doesnt need dst reference")
Signed-off-by: Leone Fernando <leone4fernando@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a33e9da3470499e9ff476138f271fb52d6bfe767 ]
When running fib_nexthop_multiprefix test I saw all IPv6 test failed.
e.g.
]# ./fib_nexthop_multiprefix.sh
TEST: IPv4: host 0 to host 1, mtu 1300 [ OK ]
TEST: IPv6: host 0 to host 1, mtu 1300 [FAIL]
With -v it shows
COMMAND: ip netns exec h0 /usr/sbin/ping6 -s 1350 -c5 -w5 2001:db8:101::1
PING 2001:db8:101::1(2001:db8:101::1) 1350 data bytes
From 2001:db8:100::64 icmp_seq=1 Packet too big: mtu=1300
--- 2001:db8:101::1 ping statistics ---
1 packets transmitted, 0 received, +1 errors, 100% packet loss, time 0ms
Route get
2001:db8:101::1 via 2001:db8:100::64 dev eth0 src 2001:db8:100::1 metric 1024 expires 599sec mtu 1300 pref medium
Searching for:
2001:db8:101::1 from :: via 2001:db8:100::64 dev eth0 src 2001:db8:100::1 .* mtu 1300
The reason is when CONFIG_IPV6_SUBTREES is not enabled, rt6_fill_node() will
not put RTA_SRC info. After fix:
]# ./fib_nexthop_multiprefix.sh
TEST: IPv4: host 0 to host 1, mtu 1300 [ OK ]
TEST: IPv6: host 0 to host 1, mtu 1300 [ OK ]
Fixes: 735ab2f65d ("selftests: Add test with multiple prefixes using single nexthop")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Link: https://lore.kernel.org/r/20231213060856.4030084-7-liuhangbin@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit bfc3c6743de0ecb169026c36cbdbc0d12d22a528 ]
The binding erroneously decreed that the SCMI variants of the ST
evaluation kits are compatible with the non-SCMI variants.
This is not correct, as a kernel or bootloader compatible with the non-SCMI
variant is not necessarily able to function, when direct access
to resources is replaced by having to talk SCMI to the secure monitor.
The binding has been adjusted to reflect thus, so synchronize the device
trees now.
Fixes: 5b7e58313a ("ARM: dts: stm32: Add SCMI version of STM32 boards (DK1/DK2/ED1/EV1)")
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Alexandre Torgue <alexandre.torgue@foss.st.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 7ea3e7763c50b20a8bd25cf524ea0c6463de69be ]
We obtain the semaphore and set HISI_SAS_RESETTING_BIT in
hisi_sas_reset_prepare_v3_hw(), block the scsi host and set
HISI_SAS_REJECT_CMD_BIT in hisi_sas_controller_reset_prepare(), released
them in hisi_sas_controller_reset_done(). However, if the HW reset failure
in FLR results in early return, the semaphore and flag bits will not be
release.
Rollback some operations including clearing flags / releasing semaphore
when FLR is failed.
Fixes: e5ea48014a ("scsi: hisi_sas: Implement handlers of PCIe FLR for v3 hw")
Signed-off-by: Yihang Li <liyihang9@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Link: https://lore.kernel.org/r/1702525516-51258-5-git-send-email-chenxiang66@hisilicon.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit d34ee535705eb43885bc0f561c63046f697355ad ]
In function hisi_sas_controller_prereset(), -ENOSYS (Function not
implemented) should be returned if the driver does not support .soft_reset.
Returns -EPERM (Operation not permitted) if HISI_SAS_RESETTING_BIT is
already be set.
In function _suspend_v3_hw(), returns -EPERM (Operation not permitted) if
HISI_SAS_RESETTING_BIT is already be set.
Fixes: 4522204ab2 ("scsi: hisi_sas: tidy host controller reset function a bit")
Signed-off-by: Yihang Li <liyihang9@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Link: https://lore.kernel.org/r/1702525516-51258-3-git-send-email-chenxiang66@hisilicon.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 1f794d3eed5345413c2b0cf1bcccc92d77681220 ]
When the GPU nodes were added, the GPU_PLL_OUT was configured
for 1000MHz, but this requires the SoC to run in overdrive mode
which requires an elevated voltage operating point.
Since this may run some boards out of spec, the default clock
should be set to 800MHz for nominal operating mode. Boards
that run at the higher voltage can update their clocks
accordingly.
Fixes: 4523be8e46 ("arm64: dts: imx8mm: Add GPU nodes for 2D and 3D core")
Signed-off-by: Adam Ford <aford173@gmail.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 5fa3d1a00c2d4ba14f1300371ad39d5456e890d7 ]
On the error path of device_add_disk(), device's memalloc_noio flag was
set but not cleared. As the comment of pm_runtime_set_memalloc_noio(),
"The function should be called between device_add() and device_del()".
Clear this flag before device_del() now.
Fixes: 25e823c8c3 ("block/genhd.c: apply pm_runtime_set_memalloc_noio on block devices")
Signed-off-by: Li Nan <linan122@huawei.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20231211075356.1839282-1-linan666@huaweicloud.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit fa6ad88e023ddfa6c5dcdb466d159e89f451e305 ]
Due to the increase in the number of power tables for 6Ghz on CLC,
the variable nr_country is no longer sufficient to represent the
total quantity. Therefore, we have switched to calculating the
length of clc buf to obtain the correct power table. Additionally,
the version number has been incremented to 1.
Fixes: 23bdc5d8ca ("wifi: mt76: mt7921: introduce Country Location Control support")
Signed-off-by: Ming Yen Hsieh <mingyen.hsieh@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit f5f185bf7c42f6ca885202fefc40fc871d08a722 ]
The DMA-related nodes RDMA/WROT in MDP3 should be changed to generic names.
In addition, fix improper space indent in example.
Fixes: 4ad7b39623 ("media: dt-binding: mediatek: add bindings for MediaTek MDP3 components")
Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Acked-by: Rob Herring <robh@kernel.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 44ab3ee76a5a977864ba0bb6c352dcf6206804e0 ]
The Hi6421 PMIC regulator child nodes do not have unit addresses so drop
the incorrect '#address-cells' and '#size-cells' properties.
Fixes: 6219b20e1e ("arm64: dts: hisilicon: Add support for Hikey 970 PMIC")
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Signed-off-by: Wei Xu <xuwei5@hisilicon.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a833a17aeac73b33f79433d7cee68d5cafd71e4f ]
This patch fixes a bug around the verification of possibly-zero-sized
stack accesses. When the access was done through a var-offset stack
pointer, check_stack_access_within_bounds was incorrectly computing the
maximum-offset of a zero-sized read to be the same as the register's min
offset. Instead, we have to take in account the register's maximum
possible value. The patch also simplifies how the max offset is checked;
the check is now simpler than for min offset.
The bug was allowing accesses to erroneously pass the
check_stack_access_within_bounds() checks, only to later crash in
check_stack_range_initialized() when all the possibly-affected stack
slots are iterated (this time with a correct max offset).
check_stack_range_initialized() is relying on
check_stack_access_within_bounds() for its accesses to the
stack-tracking vector to be within bounds; in the case of zero-sized
accesses, we were essentially only verifying that the lowest possible
slot was within bounds. We would crash when the max-offset of the stack
pointer was >= 0 (which shouldn't pass verification, and hopefully is
not something anyone's code attempts to do in practice).
Thanks Hao for reporting!
Fixes: 01f810ace9 ("bpf: Allow variable-offset stack access")
Reported-by: Hao Sun <sunhao.th@gmail.com>
Signed-off-by: Andrei Matei <andreimatei1@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20231207041150.229139-2-andreimatei1@gmail.com
Closes: https://lore.kernel.org/bpf/CACkBjsZGEUaRCHsmaX=h-efVogsRfK1FPxmkgb0Os_frnHiNdw@mail.gmail.com/
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 92184eae1d5ad804884e2c6e289d885b9e3194d1 ]
The ieee80211_queue_work function queues work into the mac80211
local->workqueue, which is widely used for mac80211 internal
work processes. In the mt76 driver, both the mt76-sido-status and
mt76-sdio-net threads enqueue workers to the workqueue with this
function. However, in some cases, when two workers are enqueued
to the workqueue almost simultaneously, the second worker may not
be scheduled immediately and may get stuck for a while.
This can cause timing issues. To avoid these timing
conflicts caused by worker scheduling, replace the worker
with an independent thread.
Fixes: 48fab5bbef ("mt76: mt7921: introduce mt7921s support")
Signed-off-by: Wang Zhao <wang.zhao@mediatek.com>
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 24f8aba9a7c77c7e9d814a5754798e8346c7dd28 ]
The DP/DM wakeup interrupts are edge triggered and which edge to trigger
on depends on use-case and whether a Low speed or Full/High speed device
is connected.
Note that only triggering on rising edges can be used to detect resume
events but not disconnect events.
Fixes: bb9efa59c6 ("arm64: dts: qcom: sc7280: Add USB related nodes")
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Link: https://lore.kernel.org/r/20231120164331.8116-6-johan+linaro@kernel.org
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 94085049fdad7a36fe14dd55e72e712fe55d6bca ]
GPU_SMMU SID 1 is meant for Adreno LPAC (Low Priority Async Compute).
On platforms that support it (in firmware), it is necessary to
describe that link, or Adreno register access will hang the board.
The current settings are functionally identical, *but* due to what is
likely hardcoded security policies, the secure firmware rejects them,
resulting in the board hanging. To avoid that, alter the settings such
that SID 0 and 1 are described separately.
Fixes: 96c471970b ("arm64: dts: qcom: sc7280: Add gpu support")
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Link: https://lore.kernel.org/r/20230926-topic-a643-v2-2-06fa3d899c0a@linaro.org
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit ab125ed3ec1c10ccc36bc98c7a4256ad114a3dae ]
When register is spilled onto a stack as a 1/2/4-byte register, we set
slot_type[BPF_REG_SIZE - 1] (plus potentially few more below it,
depending on actual spill size). So to check if some stack slot has
spilled register we need to consult slot_type[7], not slot_type[0].
To avoid the need to remember and double-check this in the future, just
use is_spilled_reg() helper.
Fixes: 27113c59b6 ("bpf: Check the other end of slot_type for STACK_SPILL")
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20231205184248.1502704-4-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 7f770d28f2e5abfd442ad689ba1129dd66593529 ]
When do arping, the interface need to be specified. Or we will
get error: Interface "lo" is not ARPable. And the test failed.
]# ./arp_ndisc_untracked_subnets.sh
TEST: test_arp: accept_arp=0 [ OK ]
TEST: test_arp: accept_arp=1 [FAIL]
TEST: test_arp: accept_arp=2 same_subnet=0 [ OK ]
TEST: test_arp: accept_arp=2 same_subnet=1 [FAIL]
After fix:
]# ./arp_ndisc_untracked_subnets.sh
TEST: test_arp: accept_arp=0 [ OK ]
TEST: test_arp: accept_arp=1 [ OK ]
TEST: test_arp: accept_arp=2 same_subnet=0 [ OK ]
TEST: test_arp: accept_arp=2 same_subnet=1 [ OK ]
Fixes: 0ea7b0a454 ("selftests: net: arp_ndisc_untracked_subnets: test for arp_accept and accept_untracked_na")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 876673364161da50eed6b472d746ef88242b2368 ]
When updating or deleting an inner map in map array or map htab, the map
may still be accessed by non-sleepable program or sleepable program.
However bpf_map_fd_put_ptr() decreases the ref-counter of the inner map
directly through bpf_map_put(), if the ref-counter is the last one
(which is true for most cases), the inner map will be freed by
ops->map_free() in a kworker. But for now, most .map_free() callbacks
don't use synchronize_rcu() or its variants to wait for the elapse of a
RCU grace period, so after the invocation of ops->map_free completes,
the bpf program which is accessing the inner map may incur
use-after-free problem.
Fix the free of inner map by invoking bpf_map_free_deferred() after both
one RCU grace period and one tasks trace RCU grace period if the inner
map has been removed from the outer map before. The deferment is
accomplished by using call_rcu() or call_rcu_tasks_trace() when
releasing the last ref-counter of bpf map. The newly-added rcu_head
field in bpf_map shares the same storage space with work field to
reduce the size of bpf_map.
Fixes: bba1dc0b55 ("bpf: Remove redundant synchronize_rcu.")
Fixes: 638e4b825d ("bpf: Allows per-cpu maps and map-in-map in sleepable programs")
Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20231204140425.1480317-5-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
