In reserve_compress_blocks(), we update blkaddrs of dnode in prior to
inc_valid_block_count(), it may cause inconsistent status bewteen
i_blocks and blkaddrs once inc_valid_block_count() fails.
To fix this issue, it needs to reverse their invoking order.
Fixes: c75488fb4d ("f2fs: introduce F2FS_IOC_RESERVE_COMPRESS_BLOCKS")
Reviewed-by: Daeho Jeong <daehojeong@google.com>
Signed-off-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Bug: 313549450
(cherry picked from commit 54607494875edd636aff3c21ace3ad9a7da758a9
https://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git dev)
Change-Id: I74a5d1d47246d39b1ed0131c714c6a2a2bfab6cf
Signed-off-by: Daeho Jeong <daehojeong@google.com>
Compressed cluster may not be released due to we can fail in
release_compress_blocks(), fix to handle reserved compressed
cluster correctly in reserve_compress_blocks().
Fixes: 4c8ff7095b ("f2fs: support data compression")
Signed-off-by: Sheng Yong <shengyong@oppo.com>
Signed-off-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Bug: 313549450
(cherry picked from commit eb8fbaa53374e0a2d4381190abfe708481517bbb
https://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git dev)
Change-Id: I2a4223efb548a921a8f28f45798709743e38e899
Signed-off-by: Daeho Jeong <daehojeong@google.com>
When we overwrite compressed cluster w/ normal cluster, we should
not unlock cp_rwsem during f2fs_write_raw_pages(), otherwise data
will be corrupted if partial blocks were persisted before CP & SPOR,
due to cluster metadata wasn't updated atomically.
Fixes: 4c8ff7095b ("f2fs: support data compression")
Reviewed-by: Daeho Jeong <daehojeong@google.com>
Signed-off-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Bug: 313549450
(cherry picked from commit fd244524c2cf07b5f4c3fe8abd6a99225c76544b
https://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git dev)
Change-Id: I231eef3eb7ee5a105ad929718a6d1d67f4950dd2
Signed-off-by: Daeho Jeong <daehojeong@google.com>
If data block in compressed cluster is not persisted with metadata
during checkpoint, after SPOR, the data may be corrupted, let's
guarantee to write compressed page by checkpoint.
Fixes: 4c8ff7095b ("f2fs: support data compression")
Reviewed-by: Daeho Jeong <daehojeong@google.com>
Signed-off-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Bug: 313549450
(cherry picked from commit 8a430dd49e9cb021372b0ad91e60aeef9c6ced00
https://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git dev)
Change-Id: Ie52d0bc7d89ec3111f1f6e8140555358620e875c
Signed-off-by: Daeho Jeong <daehojeong@google.com>
Don't check the "LTO" variable in build.config.gki.aarch64.fips140,
since fips140.ko no longer depends on LTO. Also, Kleaf doesn't set the
"LTO" variable anyway; it was specific to build.sh.
Bug: 188620248
Change-Id: I213faa4c8c1a23898c08da121b0a5dc602b7218a
Signed-off-by: Eric Biggers <ebiggers@google.com>
Symbols updated to QCOM abi symbol list for Toshiba EMAC:
clk_unregister_fixed_rate
get_phy_device
__hw_addr_sync_dev
netdev_features_change
pcie_capability_read_dword
phy_device_free
phy_device_register
phy_loopback
phy_lookup_setting
phy_restart_aneg
Bug: 325987248
Change-Id: I0f4a1070283f2e9c98df4de9da0660e572a7826e
Signed-off-by: Andre Ding <quic_shuangxi@quicinc.com>
CONFIG_TASK_DELAY_ACCT cannot be enabled since `struct task_struct`
is KMI frozen. Instead, use vendor hooks to allow delay accounting
to be implemented in a vendor module.
Bug: 327566572
Bug: 310129610
Bug: 314931189
Change-Id: If814d7834889fe162aba3dd97e935289127ca3ae
Signed-off-by: Dongyun Liu <dongyun.liu@transsion.com>
(cherry picked from commit bb57557246d39dba8a66df7f43983fe1ec71bff6)
(cherry picked from commit 896cff8734)
Revert commit 9b476ebe99 in preparation of making this a
Pixel-specific change. A disadvantage of building null_blk and
scsi_debug as kernel modules for all vendors is that this causes these
kernel modules to be loaded by the init process during boot and there is
no vendor-independent mechanism for preventing kernel modules to be
loaded.
Bug: 323406846
Change-Id: I6e7bcdd9fd06194bbc134ee52251b6ae9a3c1477
Signed-off-by: Bart Van Assche <bvanassche@google.com>
This reverts commit ee42bfc791 which is
commit af42269c35 upstream.
It is reported to cause crashes, so revert it for now.
Bug: 326555421
Change-Id: I2fb3626c306e0444f4e0eb42a95488e688942ba9
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Vendors might want to change tasks affinity settings when they are
moving from one cpuset into the other. Add vendor hook to give control
to vendor to implement what they need.
This reverts commit a42f6e7d0aa0("Revert "ANDROID: sched/cpuset: Add
vendor hook to change tasks affinity"") to effectively bring back the
original change.
Bug: 323765868
Change-Id: I47c1ee0dabda491732932c653bd80e95ee550791
Signed-off-by: Satya Durga Srinivasu Prabhala <satyap@codeaurora.org>
Signed-off-by: Sai Harshini Nimmala <quic_snimmala@quicinc.com>
When cpu loading is high, the task maybe preempted after restoring the
sched priority in trace_android_vh_binder_free_buf(). This means that
node->has_async_transaction can't be cleared immediately and the work
isn't added to the proc->todo queue as soon as possible.
To fix this we add a new hook trace_android_vh_binder_buffer_release()
to restore the priority after node->has_async_transaction has been
updated and the node->work has been added to the proc->todo queue.
Note: the old trace_android_vh_binder_free_buf() hook is kept to avoid
breaking KMI but is not extrictly needed.
Bug: 327307900
Fixes: 0eb66ec39ca8 ("ANDROID: vendor_hooks: Add hooks for binder")
Change-Id: I8126c79c9c68faa3ce0cd87ce83e2591bd61d5dd
Signed-off-by: Fuchun Liao <lfc@oppo.com>
[cmllamas: fix-up commit log and variable naming]
Signed-off-by: Carlos Llamas <cmllamas@google.com>
This reverts commit 0db211ec0f which is
commit 1361917030 upstream.
It is reported to cause crashes, so revert it for now.
Bug: 326555421
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: I3dd7744a9b706a959cbed3a793be668147dfcb9a
Add hook for trace_android_vh_copy_process, which gives the vendor a chance to monitor the total thread count of the system and the thread count under a particular process
Bug: 325765508
Change-Id: Ibeb8aa571d44997ac10623321cd00d1686bde033
Signed-off-by: chenweitao <chenweitao@oppo.com>
Existing runtime PM ftrace events (`rpm_suspend`, `rpm_resume`,
`rpm_return_int`) offer limited visibility into the exact timing of device
runtime power state transitions, particularly when asynchronous operations
are involved. When the `rpm_suspend` or `rpm_resume` functions are invoked
with the `RPM_ASYNC` flag, a return value of 0 i.e., success merely
indicates that the device power state request has been queued, not that
the device has yet transitioned.
A new ftrace event, `rpm_status`, is introduced. This event directly logs
the `power.runtime_status` value of a device whenever it changes providing
granular tracking of runtime power state transitions regardless of
synchronous or asynchronous `rpm_suspend` / `rpm_resume` usage.
Signed-off-by: Vilas Bhat <vilasbhat@google.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Bug: 325508361
(cherry picked from commit 015abee404760249a5c968b9ce29216b94b8ced1
https://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git
linux-next)
Change-Id: Iad7cae74c41b23b430331379c180b5e59bc32c40
Signed-off-by: Vilas Bhat <vilasbhat@google.com>
commit 60c0c230c6f046da536d3df8b39a20b9a9fd6af0 upstream.
rbtree lazy gc on insert might collect an end interval element that has
been just added in this transactions, skip end interval elements that
are not yet active.
Bug: 325477234
Fixes: f718863aca ("netfilter: nft_set_rbtree: fix overlap expiration walk")
Cc: stable@vger.kernel.org
Reported-by: lonial con <kongln9170@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit 1296c110c5)
Signed-off-by: Lee Jones <joneslee@google.com>
Change-Id: I42f7bca418d47948292b15ace9f371b81ccd7fe8
If command timeout happens and cq complete IRQ is raised at the same time,
ufshcd_mcq_abort clears lprb->cmd and a NULL pointer deref happens in the
ISR. Error log:
ufshcd_abort: Device abort task at tag 18
Unable to handle kernel NULL pointer dereference at virtual address
0000000000000108
pc : [0xffffffe27ef867ac] scsi_dma_unmap+0xc/0x44
lr : [0xffffffe27f1b898c] ufshcd_release_scsi_cmd+0x24/0x114
Fixes: f1304d4420 ("scsi: ufs: mcq: Added ufshcd_mcq_abort()")
Cc: stable@vger.kernel.org
Change-Id: I20d67f4bda2a83a7dc5c63e63b32232ff6401597
(cherry picked from commit 27900d7119c464b43cd9eac69c85884d17bae240)
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Link: https://lore.kernel.org/r/20231106075117.8995-1-peter.wang@mediatek.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
If a dev command times out, clk_scaling.active_reqs is not decreased which
causes abnormal clock scaling.
It is complicated to handle different dev command timeout cases in both
legacy mode and MCQ mode. Besides, dev cmds are rarely used and the busy
time is short.
Remove clock scaling busy window for dev cmds like we do for UIC or TM cmds
which don't update busy window either.
Change-Id: I2f25dc5032126778ac99b6a79bd78f9578dd730c
(cherry picked from commit e66413faa5b55800ad42a570aad674aed634c5e6)
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Link: https://lore.kernel.org/r/20231004062454.29165-1-peter.wang@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add the underlying symbols for trace_android_vh_try_fixup_sea.
Bug: 320358381
Change-Id: I0fa34513d1cd503159abfef0eb6fe50284f21316
Signed-off-by: Chris Goldsworthy <quic_cgoldswo@quicinc.com>
path_lookupat() is capable of safely reading unampped VAs. If an
unmapped VA is read whilst the function is being called, the resulting
page fault will get re-directed to __do_page_fault(), which will call
fixup_exception() to handle the aforementioned unmapped VA read.
Now, for an OS running in a VM, let's say that memory was still mapped
at S1 but lent to another VM (i.e. unmapped at S2 for the given VM).
The reading of an unmapped VA in path_lookupat() still needs to be
handled. For hypervisors that inject an abort leading to a do_sea()
call, call fixup_exception() from do_sea() if
trace_android_vh_try_fixup_sea() indicates that we can do so.
Bug: 320358381
Change-Id: I0aedcd954f08e3011b27524f9a7b038debbb246d
Signed-off-by: Chris Goldsworthy <quic_cgoldswo@quicinc.com>
In dw_pcie_host_init() regardless of whether the link has been
started or not, the code waits for the link to come up. Even in
cases where start_link() is not defined the code ends up spinning
in a loop for 1 second. Since in some systems dw_pcie_host_init()
gets called during probe, this one second loop for each pcie
interface instance ends up extending the boot time.
Wait for the link up in only if the start_link() is defined.
The patch submitted to the upstream kernel (see link below) was not
accepted due to no upstream user.
The change here is a simplified version of that patch, which will wait
for a link only if start_link ops has been defined.
Bug: 315052790
Link: https://lore.kernel.org/all/20240112093006.2832105-1-ajayagarwal@google.com/
Change-Id: I4e8d00f6195062728417e41ddd51072880676920
Signed-off-by: Sajid Dalvi <sdalvi@google.com>
Currently the core UFS driver does not have a vops to notify when the
device is operational. This commit introduces a hook, which serves to
notify device completing initialization and is ready to accept I/O.
This is required by the FIPS140-2 [1] self integrity test of inline
encryption engine, which must run whenever the host controller is reset.
The code requires sleeping while waiting for I/O to complete and allocating
some memory dynamically, which requires the vendor hook to be restricted.
[1] https://csrc.nist.gov/publications/detail/fips/140/2/final
Bug: 185809932
Signed-off-by: Konstantin Vyshetsky <vkon@google.com>
(cherry picked from commit e774e4eca6)
(cherry picked from https://android-review.googlesource.com/q/commit:c0f24579002c3fb0e404f223f8574c7f4fdac200)
Merged-In: I6f476f9c2e2b50574d2898c3f1ef6b648d92df24
Change-Id: I6f476f9c2e2b50574d2898c3f1ef6b648d92df24
Adding the following symbols:
- devfreq_remove_governor
Test: Build
Bug: 325161541
Change-Id: I76dd0eed62685b71fd95f758fd9b401da87e4df2
Signed-off-by: Will Song <jinpengsong@google.com>
If a virtio_pci_device supports native PCI power management and has the
No_Soft_Reset bit set, then skip resetting and reinitializing the device
when suspending and restoring the device. This allows system-wide low
power states like s2idle to be used in systems with stateful virtio
devices that can't simply be re-initialized (e.g. virtio-fs).
Signed-off-by: David Stevens <stevensd@chromium.org>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Bug: 318540621
Change-Id: Ifc60f72a6a6bd256fa14e291e615c7514dcb4d7d
(cherry picked from commit dff4fa0e57856045359440d05af9e9b7f7048f52)
Signed-off-by: Philip Chen <philipchen@google.com>
(cherry picked from commit a7dcf250aeb8e2a572774a0e51fc0432875b6661)
commit b1db244ffd041a49ecc9618e8feb6b5c1afcdaa7 upstream.
When deactivating the catch-all set element, check the state in the next
generation that represents this transaction.
This bug uncovered after the recent removal of the element busy mark
a2dd0233cb ("netfilter: nf_tables: remove busy mark and gc batch API").
Bug: 323034444
Fixes: aaa31047a6 ("netfilter: nftables: add catch-all set element support")
Cc: stable@vger.kernel.org
Reported-by: lonial con <kongln9170@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit a372f1d01b)
Signed-off-by: Lee Jones <joneslee@google.com>
Change-Id: I40ac92a33db48a2709079d27d7d529946ba297c7
kthread_create_on_cpu no longer marks the created thread as a per cpu
thread, so the affinity might get lost on suspend or other hotplug
events.
Export kthread_set_per_cpu so a module that needs a kthread to stay on a
specific cpu can accomplish that.
Bug: 274202992
Change-Id: Iaafc12f93f341f9e0586cb051b7f1c941f140866
Signed-off-by: Arve Hjønnevåg <arve@android.com>
(cherry picked from commit 4fdc7e4658)
The following list of commits, which are in the 6.1.57 release, add a
new field in struct ipv6_devconf and this breaks the abi. Fix this by
using one of the reserved slots for upstream changes and update the stg
file to preserve the build.
b691264274 ("net: release reference to inet6_dev pointer")
bad004c384 ("net: change accept_ra_min_rtr_lft to affect all RA lifetimes")
ec4162bb70 ("net: add sysctl accept_ra_min_rtr_lft")
type 'struct ipv6_devconf' changed
member 'union { struct { __s32 accept_ra_min_lft; u32 padding; }; struct { u64 android_kabi_reserved1; }; union { }; }' was added
member 'u64 android_kabi_reserved1' was removed
Bug: 320243175
Change-Id: Icddcad574f3c29f6d3e63b10f8c7e51bdf46fe7f
Signed-off-by: Carlos Llamas <cmllamas@google.com>
The android_kabi.h header file uses the __stringify() macro without
explicitly including its definition via linux/stringify.h. This can
result in build breaking when using the ANDROID_KABI_USE macro:
common/include/linux/ipv6.h:83:2: error: expected ')'
ANDROID_KABI_USE(1, struct { __s32 accept_ra_min_lft; u32 padding; });
^
[...]
common/include/linux/android_kabi.h:44:24: note: expanded from macro '__ANDROID_KABI_CHECK_SIZE_ALIGN'
__FILE__ ":" __stringify(__LINE__) ": " \
^
To fix this let android_kabi.h include stringify.h explicitly instead of
relying on includes of previous unrelated header files.
Bug: 324437514
Change-Id: I16cced44e723871b2e1a92b312e60f38e41fea70
Signed-off-by: Carlos Llamas <cmllamas@google.com>
I encountered a race issue after lengthy (~594647 sec) stress tests on
a 64k-page arm64 VM with several 4k-block EROFS images. The timing
is like below:
z_erofs_try_inplace_io z_erofs_fill_bio_vec
cmpxchg(&compressed_bvecs[].page,
NULL, ..)
[access bufvec]
compressed_bvecs[] = *bvec;
Previously, z_erofs_submit_queue() just accessed bufvec->page only, so
other fields in bufvec didn't matter. After the subpage block support
is landed, .offset and .end can be used too, but filling bufvec isn't
an atomic operation which can cause inconsistency.
Let's use a spinlock to keep the atomicity of each bufvec.
Fixes: 192351616a9d ("erofs: support I/O submission for sub-page compressed blocks")
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Reviewed-by: Sandeep Dhavale <dhavale@google.com>
Reviewed-by: Yue Hu <huyue2@coolpad.com>
Link: https://lore.kernel.org/r/20240125120039.3228103-1-hsiangkao@linux.alibaba.com
Bug: 324640522
(cherry picked from commit cc4b2dd95f0d1eba8c691b36e8f4d1795582f1ff
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/ master)
[dhavale: introduced spinlock in struct erofs_workgroup, upstream has a
change where atomic is replaced by lockref but pulling that and related
changes will cause unnecessary churn. Adding spinlock keeps the spirit
of the change in-tact by fixing the race. Also updated commit message
as we are not using lockref.]
Change-Id: Id20a0a433277ab71d46bce48c81824564d1b391d
Signed-off-by: Sandeep Dhavale <dhavale@google.com>
Instead of invalidating the tlb of the guest every page it relinquishes,
invalidate once before the host is informed about the free pages.
Bug: 315173520
Change-Id: I8450fd552a27dfc791efa15e884982aeedaa3234
Signed-off-by: Prakruthi Deepak Heragu <quic_pheragu@quicinc.com>
Move out the kvm specific page_relinquish call and introduce a generic
hypervisor_ops structure. Based on the hypervisor running underneath the
guest, populate the operations to make specific hypercalls.
Bug: 315173520
Change-Id: I93d5a92553a04d56811b0666dec6eacc1ca168e7
Signed-off-by: Prakruthi Deepak Heragu <quic_pheragu@quicinc.com>
