Now, extended attribute value maximum length is 64K. The memory
requested here does not need continuous physical addresses, so it is
appropriate to use kvmalloc to request memory. At the same time, it
can also cope with the situation that the extended attribute will
become longer in the future.
Signed-off-by: Ye Bin <yebin10@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20221208023233.1231330-3-yebin@huaweicloud.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When expanding inode space in ext4_expand_extra_isize_ea() we may need
to allocate external xattr block. If quota is not initialized for the
inode, the block allocation will not be accounted into quota usage. Make
sure the quota is initialized before we try to expand inode space.
Make sure we initialize quotas before possibly expanding inode space
(and thus maybe needing to allocate external xattr block) in
ext4_ioctl_setproject(). This prevents not accounting the necessary
block allocation.
This occurs in 'ext4_xattr_inode_create()'. If 'ext4_mark_inode_dirty()'
fails, dropping i_nlink of the inode is needed. Or will lead to inode leak.
Signed-off-by: Ye Bin <yebin10@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20221208023233.1231330-5-yebin@huaweicloud.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Above issue may happens as follows:
ext4_da_write_begin
ext4_da_write_inline_data_begin
ext4_da_convert_inline_data_to_extent
ext4_clear_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA);
ext4_da_write_end
ext4_run_li_request
ext4_mb_prefetch
ext4_read_block_bitmap_nowait
ext4_validate_block_bitmap
ext4_mark_group_bitmap_corrupted(sb, block_group, EXT4_GROUP_INFO_BBITMAP_CORRUPT)
percpu_counter_sub(&sbi->s_freeclusters_counter,grp->bb_free);
-> sbi->s_freeclusters_counter become zero
ext4_da_write_begin
if (ext4_nonda_switch(inode->i_sb)) -> As freeclusters_counter is zero will return true
*fsdata = (void *)FALL_BACK_TO_NONDELALLOC;
ext4_write_begin
ext4_da_write_end
if (write_mode == FALL_BACK_TO_NONDELALLOC)
ext4_write_end
if (inline_data)
ext4_write_inline_data_end
ext4_write_inline_data
BUG_ON(pos + len > EXT4_I(inode)->i_inline_size);
-> As inode is already convert to extent, so 'pos + len' > inline_size
-> then trigger BUG.
To solve this issue, instead of checking ext4_has_inline_data() which
is only cleared after data has been written back, check the
EXT4_STATE_MAY_INLINE_DATA flag in ext4_write_end().
Fixes: f19d5870cbf7 ("ext4: add normal write support for inline data") Reported-by: syzbot+4faa160fa96bfba639f8@syzkaller.appspotmail.com Reported-by: Jun Nie <jun.nie@linaro.org> Signed-off-by: Ye Bin <yebin10@huawei.com> Link: https://lore.kernel.org/r/20221206144134.1919987-1-yebin@huaweicloud.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit fb0a387dcdcd ("ext4: limit block allocations for indirect-block
files to < 2^32") added code to try to allocate xattr block with 32-bit
block number for indirect block based files on the grounds that these
files cannot use larger block numbers. It also added BUG_ON when
allocated block could not fit into 32 bits. This is however bogus
reasoning because xattr block is stored in inode->i_file_acl and
inode->i_file_acl_hi and as such even indirect block based files can
happily use full 48 bits for xattr block number. The proper handling
seems to be there basically since 64-bit block number support was added.
So remove the bogus limitation and BUG_ON.
Cc: Eric Sandeen <sandeen@redhat.com> Fixes: fb0a387dcdcd ("ext4: limit block allocations for indirect-block files to < 2^32") Signed-off-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20221121130929.32031-1-jack@suse.cz Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If a block is out of range in ext4_get_branch(), -ENOMEM will be returned
to user-space. Obviously, this error code isn't really useful. This
patch fixes it by making sure the right error code (-EFSCORRUPTED) is
propagated to user-space. EUCLEAN is more informative than ENOMEM.
When a backup superblock is updated in update_backups(), the primary
superblock's offset in the group (that is, sbi->s_sbh->b_blocknr) is used
as the backup superblock's offset in its group. However, when the block
size is 1K and bigalloc is enabled, the two offsets are not equal. This
causes the backup group descriptors to be overwritten by the superblock
in update_backups(). Moreover, if meta_bg is enabled, the file system will
be corrupted because this feature uses backup group descriptors.
To solve this issue, we use a more accurate ext4_group_first_block_no() as
the offset of the backup superblock in its group.
Fixes: d77147ff443b ("ext4: add support for online resizing with bigalloc") Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Cc: stable@kernel.org Link: https://lore.kernel.org/r/20221117040341.1380702-4-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When converting files with inline data to extents, delayed allocations
made on a file system created with both the bigalloc and inline options
can result in invalid extent status cache content, incorrect reserved
cluster counts, kernel memory leaks, and potential kernel panics.
With bigalloc, the code that determines whether a block must be
delayed allocated searches the extent tree to see if that block maps
to a previously allocated cluster. If not, the block is delayed
allocated, and otherwise, it isn't. However, if the inline option is
also used, and if the file containing the block is marked as able to
store data inline, there isn't a valid extent tree associated with
the file. The current code in ext4_clu_mapped() calls
ext4_find_extent() to search the non-existent tree for a previously
allocated cluster anyway, which typically finds nothing, as desired.
However, a side effect of the search can be to cache invalid content
from the non-existent tree (garbage) in the extent status tree,
including bogus entries in the pending reservation tree.
To fix this, avoid searching the extent tree when allocating blocks
for bigalloc + inline files that are being converted from inline to
extent mapped.
As 'ext4_rename' will modify 'old.inode' ctime and mark inode dirty,
which may trigger expand 'extra_isize' and allocate block. If inode
didn't init quota will lead to warning. To solve above issue, init
'old.inode' firstly in 'ext4_rename'.
Reported-by: syzbot+98346927678ac3059c77@syzkaller.appspotmail.com Signed-off-by: Ye Bin <yebin10@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20221107015335.2524319-1-yebin@huaweicloud.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
CPU: 1 PID: 4625 Comm: syz-executor.2 Not tainted 6.1.0-rc4-syzkaller-62821-gcb231e2f67ec #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/26/2022
=====================================================
Now, 'ext4_alloc_inode()' didn't init 'ei->i_flags'. If new inode failed
before set 'ei->i_flags' in '__ext4_new_inode()', then do 'iput()'. As after 6bc0d63dad7f commit will access 'ei->i_flags' in 'ext4_evict_inode()' which
will lead to access uninit-value.
To solve above issue just init 'ei->i_flags' in 'ext4_alloc_inode()'.
Reported-by: syzbot+57b25da729eb0b88177d@syzkaller.appspotmail.com Signed-off-by: Ye Bin <yebin10@huawei.com> Fixes: 6bc0d63dad7f ("ext4: remove EA inode entry from mbcache on inode eviction") Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Eric Biggers <ebiggers@google.com> Link: https://lore.kernel.org/r/20221117073603.2598882-1-yebin@huaweicloud.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In the above issue, ioctl invokes the swap_inode_boot_loader function to
swap inode<5> and inode<12>. However, inode<5> contain incorrect imode and
disordered extents, and i_nlink is set to 1. The extents check for inode in
the ext4_iget function can be bypassed bacause 5 is EXT4_BOOT_LOADER_INO.
While links_count is set to 1, the extents are not initialized in
swap_inode_boot_loader. After the ioctl command is executed successfully,
the extents are swapped to inode<12>, in this case, run the `cat` command
to view inode<12>. And Bug_ON is triggered due to the incorrect extents.
When the boot loader inode is not initialized, its imode can be one of the
following:
1) the imode is a bad type, which is marked as bad_inode in ext4_iget and
set to S_IFREG.
2) the imode is good type but not S_IFREG.
3) the imode is S_IFREG.
The BUG_ON may be triggered by bypassing the check in cases 1 and 2.
Therefore, when the boot loader inode is bad_inode or its imode is not
S_IFREG, initialize the inode to avoid triggering the BUG.
Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Jason Yan <yanaijie@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20221026042310.3839669-5-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In ext4_evict_inode(), if we evicting an inode in the 'no_delete' path,
it cannot be raced by another mark_inode_dirty(). If it happens,
someone else may accidentally dirty it without holding inode refcount
and probably cause use-after-free issues in the writeback procedure.
It's indiscoverable and hard to debug, so add an WARN_ON_ONCE() to
check and detect this issue in advance.
Suggested-by: Jan Kara <jack@suse.cz> Signed-off-by: Zhang Yi <yi.zhang@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20220629112647.4141034-2-yi.zhang@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When bigalloc is enabled, reserved cluster accounting for delayed
allocation is handled in extent_status.c. With a corrupted file
system, it's possible for this accounting to be incorrect,
dsicovered by Syzbot:
EXT4-fs error (device loop0): ext4_validate_block_bitmap:398: comm rep:
bg 0: block 5: invalid block bitmap
EXT4-fs (loop0): Delayed block allocation failed for inode 18 at logical
offset 0 with max blocks 32 with error 28
EXT4-fs (loop0): This should not happen!! Data will be lost
Above issue may happen as follows:
-------------------------------------
ext4_fill_super
ext4_orphan_cleanup
ext4_enable_quotas
ext4_quota_enable
ext4_iget --> get error inode <5>
ext4_ext_check_inode --> Wrong imode makes it escape inspection
make_bad_inode(inode) --> EXT4_BOOT_LOADER_INO set imode
dquot_load_quota_inode
vfs_setup_quota_inode --> check pass
dquot_load_quota_sb
v2_check_quota_file
v2_read_header
ext4_quota_read
ext4_bread
ext4_getblk
ext4_map_blocks
ext4_ext_map_blocks
ext4_find_extent
ext4_cache_extents
ext4_es_cache_extent
__es_tree_search.isra.0
ext4_es_end --> Wrong extents trigger BUG_ON
In the above issue, s_usr_quota_inum is set to 5, but inode<5> contains
incorrect imode and disordered extents. Because 5 is EXT4_BOOT_LOADER_INO,
the ext4_ext_check_inode check in the ext4_iget function can be bypassed,
finally, the extents that are not checked trigger the BUG_ON in the
__es_tree_search function. To solve this issue, check whether the inode is
bad_inode in vfs_setup_quota_inode().
Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com> Reviewed-by: Jason Yan <yanaijie@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20221026042310.3839669-2-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Before quota is enabled, a check on the preset quota inums in
ext4_super_block is added to prevent wrong quota inodes from being loaded.
In addition, when the quota fails to be enabled, the quota type and quota
inum are printed to facilitate fault locating.
Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Jason Yan <yanaijie@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20221026042310.3839669-3-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There are many places that will get unhappy (and crash) when ext4_iget()
returns a bad inode. However, if iget the boot loader inode, allows a bad
inode to be returned, because the inode may not be initialized. This
mechanism can be used to bypass some checks and cause panic. To solve this
problem, we add a special iget flag EXT4_IGET_BAD. Only with this flag
we'd be returning bad inode from ext4_iget(), otherwise we always return
the error code if the inode is bad inode.(suggested by Jan Kara)
Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Jason Yan <yanaijie@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20221026042310.3839669-4-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Shifting signed 32-bit value by 31 bits is undefined, so changing
significant bit to unsigned. The UBSAN warning calltrace like below:
UBSAN: shift-out-of-bounds in fs/ext4/ext4.h:591:2
left shift of 1 by 31 places cannot be represented in type 'int'
Call Trace:
<TASK>
dump_stack_lvl+0x7d/0xa5
dump_stack+0x15/0x1b
ubsan_epilogue+0xe/0x4e
__ubsan_handle_shift_out_of_bounds+0x1e7/0x20c
ext4_init_fs+0x5a/0x277
do_one_initcall+0x76/0x430
kernel_init_freeable+0x3b3/0x422
kernel_init+0x24/0x1e0
ret_from_fork+0x1f/0x30
</TASK>
Fixes: 9a4c80194713 ("ext4: ensure Inode flags consistency are checked at build time") Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com> Link: https://lore.kernel.org/r/20221031055833.3966222-1-cuigaosheng1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
I caught a issue as follows:
==================================================================
BUG: KASAN: use-after-free in __list_add_valid+0x28/0x1a0
Read of size 8 at addr ffff88814b13f378 by task mount/710
In this issue, bg_inode_table_hi is overwritten as an incorrect value.
As a result, `block < end_block` cannot be met in grow_dev_page.
Therefore, __ext4_get_inode_loc always returns '-ENOMEM' and do_writepages
keeps retrying. As a result, the writeback process is in the D state due
to an infinite loop.
Add a check on inode table block in the __ext4_get_inode_loc function by
referring to ext4_read_inode_bitmap to avoid this infinite loop.
When evicting an inode with default dioread_nolock, it could be raced by
the unwritten extents converting kworker after writeback some new
allocated dirty blocks. It convert unwritten extents to written, the
extents could be merged to upper level and free extent blocks, so it
could mark the inode dirty again even this inode has been marked
I_FREEING. But the inode->i_io_list check and warning in
ext4_evict_inode() missing this corner case. Fortunately,
ext4_evict_inode() will wait all extents converting finished before this
check, so it will not lead to inode use-after-free problem, every thing
is OK besides this warning. The WARN_ON_ONCE was originally designed
for finding inode use-after-free issues in advance, but if we add
current dioread_nolock case in, it will become not quite useful, so fix
this warning by just remove this check.
A problem about modprobe ingenic-drm failed is triggered with the following
log given:
[ 303.561088] Error: Driver 'ingenic-ipu' is already registered, aborting...
modprobe: ERROR: could not insert 'ingenic_drm': Device or resource busy
The reason is that ingenic_drm_init() returns platform_driver_register()
directly without checking its return value, if platform_driver_register()
failed, it returns without unregistering ingenic_ipu_driver_ptr, resulting
the ingenic-drm can never be installed later.
A simple call graph is shown as below:
ingenic_drm_init()
platform_driver_register() # ingenic_ipu_driver_ptr are registered
platform_driver_register()
driver_register()
bus_add_driver()
priv = kzalloc(...) # OOM happened
# return without unregister ingenic_ipu_driver_ptr
Fixing this problem by checking the return value of
platform_driver_register() and do platform_unregister_drivers() if
error happened.
intel_dsi->ports contains bitmask of enabled ports and correspondingly
logic for selecting port for VBT packet sending must use port specific
bitmask when deciding appropriate port.
Invalid userspace dma surface copies could potentially overflow
the memcpy from the surface to the snooped image leading to crashes.
To fix it the dimensions of the copybox have to be validated
against the expected size of the snooped cursor.
A typical DP-MST unplug removes a KMS connector. However care must
be taken to properly synchronize with user-space. The expected
sequence of events is the following:
1. The kernel notices that the DP-MST port is gone.
2. The kernel marks the connector as disconnected, then sends a
uevent to make user-space re-scan the connector list.
3. User-space notices the connector goes from connected to disconnected,
disables it.
4. Kernel handles the IOCTL disabling the connector. On success,
the very last reference to the struct drm_connector is dropped and
drm_connector_cleanup() is called.
5. The connector is removed from the list, and a uevent is sent to tell
user-space that the connector disappeared.
The very last step was missing. As a result, user-space thought the
connector still existed and could try to disable it again. Since the
kernel no longer knows about the connector, that would end up with
EINVAL and confused user-space.
Fix this by sending a hotplug uevent from drm_connector_cleanup().
Signed-off-by: Simon Ser <contact@emersion.fr> Cc: stable@vger.kernel.org Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Lyude Paul <lyude@redhat.com> Cc: Jonas Ådahl <jadahl@redhat.com> Tested-by: Jonas Ådahl <jadahl@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20221017153150.60675-2-contact@emersion.fr Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When add the 'a *:* rwm' entry to devcgroup A's whitelist, at first A's
exceptions will be cleaned and A's behavior is changed to
DEVCG_DEFAULT_ALLOW. Then parent's exceptions will be copyed to A's
whitelist. If copy failure occurs, just return leaving A to grant
permissions to all devices. And A may grant more permissions than
parent.
Backup A's whitelist and recover original exceptions after copy
failure.
Cc: stable@vger.kernel.org Fixes: 4cef7299b478 ("device_cgroup: add proper checking when changing default behavior") Signed-off-by: Wang Weiyang <wangweiyang2@huawei.com> Reviewed-by: Aristeu Rozanski <aris@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The second (UID) strcmp in acpi_dev_hid_uid_match considers
"0" and "00" different, which can prevent device registration.
Have the AMD IOMMU driver's ivrs_acpihid parsing code remove
any leading zeroes to make the UID strcmp succeed. Now users
can safely specify "AMDxxxxx:00" or "AMDxxxxx:0" and expect
the same behaviour.
When a driver registers with a bus, it will attempt to match with every
device on the bus through the __driver_attach() function. Currently, if
the bus_type.match() function encounters an error that is not
-EPROBE_DEFER, __driver_attach() will return a negative error code, which
causes the driver registration logic to stop trying to match with the
remaining devices on the bus.
This behavior is not correct; a failure while matching a driver to a
device does not mean that the driver won't be able to match and bind
with other devices on the bus. Update the logic in __driver_attach()
to reflect this.
Fixes: 656b8035b0ee ("ARM: 8524/1: driver cohandle -EPROBE_DEFER from bus_type.match()") Cc: stable@vger.kernel.org Cc: Saravana Kannan <saravanak@google.com> Signed-off-by: Isaac J. Manjarres <isaacmanjarres@google.com> Link: https://lore.kernel.org/r/20220921001414.4046492-1-isaacmanjarres@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Current implementation of update_mmu_cache function performs local TLB
flush. It does not take into account ASID information. Besides, it does
not take into account other harts currently running the same mm context
or possible migration of the running context to other harts. Meanwhile
TLB flush is not performed for every context switch if ASID support
is enabled.
Patch [1] proposed to add ASID support to update_mmu_cache to avoid
flushing local TLB entirely. This patch takes into account other
harts currently running the same mm context as well as possible
migration of this context to other harts.
For this purpose the approach from flush_icache_mm is reused. Remote
harts currently running the same mm context are informed via SBI calls
that they need to flush their local TLBs. All the other harts are marked
as needing a deferred TLB flush when this mm context runs on them.
The 'retp' is a pointer to the return address on the stack, so we
must pass the current return address pointer as the 'retp'
argument to ftrace_push_return_trace(). Not parent function's
return address on the stack.
When pci_create_attr() fails, pci_remove_resource_files() is called which
will iterate over the res_attr[_wc] arrays and frees every non NULL entry.
To avoid a double free here set the array entry only after it's clear we
successfully initialized it.
pci_device_is_present() previously didn't work for VFs because it reads the
Vendor and Device ID, which are 0xffff for VFs, which looks like they
aren't present. Check the PF instead.
Wei Gong reported that if virtio I/O is in progress when the driver is
unbound or "0" is written to /sys/.../sriov_numvfs, the virtio I/O
operation hangs, which may result in output like this:
This happened because pci_device_is_present(VF) returned "false" in
virtio_pci_remove(), so it called virtio_break_device(). The broken vq
meant that vring_interrupt() skipped the vq.callback() that would have
completed the virtio I/O operation via virtblk_done().
[bhelgaas: commit log, simplify to always use pci_physfn(), add stable tag] Link: https://lore.kernel.org/r/20221026060912.173250-1-mst@redhat.com Reported-by: Wei Gong <gongwei833x@gmail.com> Tested-by: Wei Gong <gongwei833x@gmail.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Erase can be zeroed in spi_nor_parse_4bait() or
spi_nor_init_non_uniform_erase_map(). In practice it happened with
mt25qu256a, which supports 4K, 32K, 64K erases with 3b address commands,
but only 4K and 64K erase with 4b address commands.
Fixes: dc92843159a7 ("mtd: spi-nor: fix erase_type array to indicate current map conf") Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com> Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20211119081412.29732-1-alexander.sverdlin@nokia.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When fixing the problem mentioned in PATCH1, we also found
the following problem:
If the IPMI is disconnected and in the sending process, the
uninstallation driver will be stuck for a long time.
The main problem is that uninstalling the driver waits for curr_msg to
be sent or HOSED. After stopping tasklet, the only place to trigger the
timeout mechanism is the circular poll in shutdown_smi.
The poll function delays 10us and calls smi_event_handler(smi_info,10).
Smi_event_handler deducts 10us from kcs->ibf_timeout.
But the poll func is followed by schedule_timeout_uninterruptible(1).
The time consumed here is not counted in kcs->ibf_timeout.
So when 10us is deducted from kcs->ibf_timeout, at least 1 jiffies has
actually passed. The waiting time has increased by more than a
hundredfold.
Now instead of calling poll(). call smi_event_handler() directly and
calculate the elapsed time.
For verification, you can directly use ebpf to check the kcs->
ibf_timeout for each call to kcs_event() when IPMI is disconnected.
Decrement at normal rate before unloading. The decrement rate becomes
very slow after unloading.
Calling v4l2_subdev_get_try_crop() and v4l2_subdev_get_try_compose()
with a subdev state of NULL leads to a NULL pointer dereference. This
can currently happen in imgu_subdev_set_selection() when the state
passed in is NULL, as this method first gets pointers to both the "try"
and "active" states and only then decides which to use.
The same issue has been addressed for imgu_subdev_get_selection() with
commit 30d03a0de650 ("ipu3-imgu: Fix NULL pointer dereference in active
selection access"). However the issue still persists in
imgu_subdev_set_selection().
Therefore, apply a similar fix as done in the aforementioned commit to
imgu_subdev_set_selection(). To keep things a bit cleaner, introduce
helper functions for "crop" and "compose" access and use them in both
imgu_subdev_set_selection() and imgu_subdev_get_selection().
Fixes: 0d346d2a6f54 ("media: v4l2-subdev: add subdev-wide state struct") Cc: stable@vger.kernel.org # for v5.14 and later Signed-off-by: Maximilian Luz <luzmaximilian@gmail.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
On the JZ4740, there is a single bit that flushes (empties) both
the transmit and receive FIFO. Later SoCs have independent flush
bits for each FIFO.
Independent FIFOs can be flushed before the snd_soc_dai_active()
check because it won't disturb other active streams. This ensures
that the FIFO we're about to use is always flushed before starting
up. With shared FIFOs we can't do that because if another substream
is active, flushing its FIFO would cause underrun errors.
This also fixes a bug: since we were only setting the JZ4740's
flush bit, which corresponds to the TX FIFO flush bit on other
SoCs, other SoCs were not having their RX FIFO flushed at all.
Fixes: 967beb2e8777 ("ASoC: jz4740: Add jz4780 support") Reviewed-by: Paul Cercueil <paul@crapouillou.net> Cc: stable@vger.kernel.org Signed-off-by: Aidan MacDonald <aidanmacdonald.0x0@gmail.com> Link: https://lore.kernel.org/r/20221023143328.160866-2-aidanmacdonald.0x0@gmail.com Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
- limit bitmap chunk size internal u64 variable to values not overflowing
the u32 bitmap superblock structure variable stored on persistent media
- assign bitmap chunk size internal u64 variable from unsigned values to
avoid possible sign extension artifacts when assigning from a s32 value
The bug has been there since at least kernel 4.0.
Steps to reproduce it:
1: mdadm -C /dev/mdx -l 1 --bitmap=internal --bitmap-chunk=256M -e 1.2
-n2 /dev/rnbd1 /dev/rnbd2
2 resize member device rnbd1 and rnbd2 to 8 TB
3 mdadm --grow /dev/mdx --size=max
The bitmap_chunksize will overflow without patch.
Cc: stable@vger.kernel.org Signed-off-by: Florian-Ewald Mueller <florian-ewald.mueller@ionos.com> Signed-off-by: Jack Wang <jinpu.wang@ionos.com> Signed-off-by: Song Liu <song@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mq-deadline ensures an in order dispatching of write requests to zoned
block devices using a per zone lock (a bit). This implies that for any
purely sequential write workload, the drive is exercised most of the
time at a maximum queue depth of one.
However, when such sequential write workload crosses a zone boundary
(when sequentially writing multiple contiguous zones), zone write
locking may prevent the last write to one zone to be issued (as the
previous write is still being executed) but allow the first write to the
following zone to be issued (as that zone is not yet being writen and
not locked). This result in an out of order delivery of the sequential
write commands to the device every time a zone boundary is crossed.
While such behavior does not break the sequential write constraint of
zoned block devices (and does not generate any write error), some zoned
hard-disks react badly to seeing these out of order writes, resulting in
lower write throughput.
This problem can be addressed by always dispatching the first request
of a stream of sequential write requests, regardless of the zones
targeted by these sequential writes. To do so, the function
deadline_skip_seq_writes() is introduced and used in
deadline_next_request() to select the next write command to issue if the
target device is an HDD (blk_queue_nonrot() being false).
deadline_fifo_request() is modified using the new
deadline_earlier_request() and deadline_is_seq_write() helpers to ignore
requests in the fifo list that have a preceding request in lba order
that is sequential.
With this fix, a sequential write workload executed with the following
fio command:
Three mount options: "tcpnodelay" and "noautotune" and "noblocksend"
were not displayed when passed in on cifs/smb3 mounts (e.g. displayed
in /proc/mounts e.g.). No change to defaults so these are not
displayed if not specified on mount.
Cc: stable@vger.kernel.org Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since rc was initialised to -ENOMEM in cifs_get_smb_ses(), when an
existing smb session was found, free_xid() would be called and then
print
CIFS: fs/cifs/connect.c: Existing tcp session with server found
CIFS: fs/cifs/connect.c: VFS: in cifs_get_smb_ses as Xid: 44 with uid: 0
CIFS: fs/cifs/connect.c: Existing smb sess found (status=1)
CIFS: fs/cifs/connect.c: VFS: leaving cifs_get_smb_ses (xid = 44) rc = -12
Fix this by initialising rc to 0 and then let free_xid() print this
instead
CIFS: fs/cifs/connect.c: Existing tcp session with server found
CIFS: fs/cifs/connect.c: VFS: in cifs_get_smb_ses as Xid: 14 with uid: 0
CIFS: fs/cifs/connect.c: Existing smb sess found (status=1)
CIFS: fs/cifs/connect.c: VFS: leaving cifs_get_smb_ses (xid = 14) rc = 0
Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Cc: stable@vger.kernel.org Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The dvb-core tries to sync the releases of opened files at
dvb_dmxdev_release() with two refcounts: dvbdev->users and
dvr_dvbdev->users. A problem is present in those two syncs: when yet
another dvb_demux_open() is called during those sync waits,
dvb_demux_open() continues to process even if the device is being
closed. This includes the increment of the former refcount, resulting
in the leftover refcount after the sync of the latter refcount at
dvb_dmxdev_release(). It ends up with use-after-free, since the
function believes that all usages were gone and releases the
resources.
This patch addresses the problem by adding the check of dmxdev->exit
flag at dvb_demux_open(), just like dvb_dvr_open() already does. With
the exit flag check, the second call of dvb_demux_open() fails, hence
the further corruption can be avoided.
Also for avoiding the races of the dmxdev->exit flag reference, this
patch serializes the dmxdev->exit set up and the sync waits with the
dmxdev->mutex lock at dvb_dmxdev_release(). Without the mutex lock,
dvb_demux_open() (or dvb_dvr_open()) may run concurrently with
dvb_dmxdev_release(), which allows to skip the exit flag check and
continue the open process that is being closed.
CVE-2022-41218 is assigned to those bugs above.
Reported-by: Hyunwoo Kim <imv4bel@gmail.com> Cc: <stable@vger.kernel.org> Link: https://lore.kernel.org/20220908132754.30532-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In function dvb_register_device() -> dvb_register_media_device() ->
dvb_create_media_entity(), dvb->entity is allocated and initialized. If
the initialization fails, it frees the dvb->entity, and return an error
code. The caller takes the error code and handles the error by calling
dvb_media_device_free(), which unregisters the entity and frees the
field again if it is not NULL. As dvb->entity may not NULLed in
dvb_create_media_entity() when the allocation of dvbdev->pad fails, a
double free may occur. This may also cause an Use After free in
media_device_unregister_entity().
Fix this by storing NULL to dvb->entity when it is freed.
clang-15's ability to elide loops completely became more aggressive when
it can deduce how a variable is being updated in a loop. Counting down
one variable by an increment of another can be replaced by a modulo
operation.
For 64b variables on 32b ARM EABI targets, this can result in the
compiler generating calls to __aeabi_uldivmod, which it does for a do
while loop in float64_rem().
For the kernel, we'd generally prefer that developers not open code 64b
division via binary / operators and instead use the more explicit
helpers from div64.h. On arm-linux-gnuabi targets, failure to do so can
result in linkage failures due to undefined references to
__aeabi_uldivmod().
While developers can avoid open coding divisions on 64b variables, the
compiler doesn't know that the Linux kernel has a partial implementation
of a compiler runtime (--rtlib) to enforce this convention.
It's also undecidable for the compiler whether the code in question
would be faster to execute the loop vs elide it and do the 64b division.
While I actively avoid using the internal -mllvm command line flags, I
think we get better code than using barrier() here, which will force
reloads+spills in the loop for all toolchains.
Link: https://github.com/ClangBuiltLinux/linux/issues/1666 Reported-by: Nathan Chancellor <nathan@kernel.org> Reviewed-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Nick Desaulniers <ndesaulniers@google.com> Tested-by: Nathan Chancellor <nathan@kernel.org> Cc: stable@vger.kernel.org Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
* tegra_csi_init
* tegra_csi_channels_alloc
* for_each_child_of_node(node, channel) -- iterates over channels
* automatically gets 'channel'
* tegra_csi_channel_alloc()
* saves into chan->of_node a pointer to the channel OF node
* automatically gets and puts 'channel'
* now the node saved in chan->of_node has refcount 0, can disappear
* tegra_csi_channels_init
* iterates over channels
* tegra_csi_channel_init -- uses chan->of_node
After that, chan->of_node keeps storing the node until the device is
removed.
of_node_get() the node and of_node_put() it during teardown to avoid any
risk.
Fixes: 1ebaeb09830f ("media: tegra-video: Add support for external sensor capture") Cc: stable@vger.kernel.org Cc: Sowjanya Komatineni <skomatineni@nvidia.com> Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
chan->mipi takes the return value of tegra_mipi_request() which can be a
valid pointer or an error. However chan->mipi is checked in several places,
including error-cleanup code in tegra_csi_channels_cleanup(), as 'if
(chan->mipi)', which suggests the initial intent was that chan->mipi should
be either NULL or a valid pointer, never an error. As a consequence,
cleanup code in case of tegra_mipi_request() errors would dereference an
invalid pointer.
Fix by ensuring chan->mipi always contains either NULL or a void pointer.
When creating probe names, a check is done to make sure it matches basic C
standard variable naming standards. Basically, starts with alphabetic or
underline, and then the rest of the characters have alpha-numeric or
underline in them.
But system names do not have any true naming conventions, as they are
created by the TRACE_SYSTEM macro and nothing tests to see what they are.
The "xhci-hcd" trace events has a '-' in the system name. When trying to
attach a eprobe to one of these trace points, it fails because the system
name does not follow the variable naming convention because of the
hyphen, and the eprobe checks fail on this.
Allow hyphens in the system name so that eprobes can attach to the
"xhci-hcd" trace events.
Both CONFIG_OSNOISE_TRACER and CONFIG_HWLAT_TRACER partially enables the
CONFIG_TRACER_MAX_TRACE code, but that is complicated and has
introduced a bug; It declares tracing_max_lat_fops data structure outside
of #ifdefs, but since it is defined only when CONFIG_TRACER_MAX_TRACE=y
or CONFIG_HWLAT_TRACER=y, if only CONFIG_OSNOISE_TRACER=y, that
declaration comes to a definition(!).
To fix this issue, and do not repeat the similar problem, makes
CONFIG_OSNOISE_TRACER and CONFIG_HWLAT_TRACER enables the
CONFIG_TRACER_MAX_TRACE always. It has there benefits;
- Fix the tracing_max_lat_fops bug
- Simplify the #ifdefs
- CONFIG_TRACER_MAX_TRACE code is fully enabled, or not.
The flag that tells the event to call its triggers after reading the event
is set for eprobes after the eprobe is enabled. This leads to a race where
the eprobe may be triggered at the beginning of the event where the record
information is NULL. The eprobe then dereferences the NULL record causing
a NULL kernel pointer bug.
Test for a NULL record to keep this from happening.
Since the CONFIG_RETHUNK and CONFIG_SLS will use INT3 for stopping
speculative execution after function return, kprobe jump optimization
always fails on the functions with such INT3 inside the function body.
(It already checks the INT3 padding between functions, but not inside
the function)
To avoid this issue, as same as kprobes, check whether the INT3 comes
from kgdb or not, and if so, stop decoding and make it fail. The other
INT3 will come from CONFIG_RETHUNK/CONFIG_SLS and those can be
treated as a one-byte instruction.
Since the CONFIG_RETHUNK and CONFIG_SLS will use INT3 for stopping
speculative execution after RET instruction, kprobes always failes to
check the probed instruction boundary by decoding the function body if
the probed address is after such sequence. (Note that some conditional
code blocks will be placed after function return, if compiler decides
it is not on the hot path.)
This is because kprobes expects kgdb puts the INT3 as a software
breakpoint and it will replace the original instruction.
But these INT3 are not such purpose, it doesn't need to recover the
original instruction.
To avoid this issue, kprobes checks whether the INT3 is owned by
kgdb or not, and if so, stop decoding and make it fail. The other
INT3 will come from CONFIG_RETHUNK/CONFIG_SLS and those can be
treated as a one-byte instruction.
After someone reported a bug report with a failed modification due to the
expected value not matching what was found, it came to my attention that
the ftrace_expected is no longer set when that happens. This makes for
debugging the issue a bit more difficult.
Set ftrace_expected to the expected code before calling ftrace_bug, so
that it shows what was expected and why it failed.
The retries in load_ucode_intel_ap() were in place to support systems
with mixed steppings. Mixed steppings are no longer supported and there is
only one microcode image at a time. Any retries will simply reattempt to
apply the same image over and over without making progress.
[ bp: Zap the circumstantial reasoning from the commit message. ]
Set ENABLE_USR_WAIT_PAUSE in KVM's supported VMX MSR configuration if the
feature is supported in hardware and enabled in KVM's base, non-nested
configuration, i.e. expose ENABLE_USR_WAIT_PAUSE to L1 if it's supported.
This fixes a bug where saving/restoring, i.e. migrating, a vCPU will fail
if WAITPKG (the associated CPUID feature) is enabled for the vCPU, and
obviously allows L1 to enable the feature for L2.
KVM already effectively exposes ENABLE_USR_WAIT_PAUSE to L1 by stuffing
the allowed-1 control ina vCPU's virtual MSR_IA32_VMX_PROCBASED_CTLS2 when
updating secondary controls in response to KVM_SET_CPUID(2), but (a) that
depends on flawed code (KVM shouldn't touch VMX MSRs in response to CPUID
updates) and (b) runs afoul of vmx_restore_control_msr()'s restriction
that the guest value must be a strict subset of the supported host value.
Although no past commit explicitly enabled nested support for WAITPKG,
doing so is safe and functionally correct from an architectural
perspective as no additional KVM support is needed to virtualize TPAUSE,
UMONITOR, and UMWAIT for L2 relative to L1, and KVM already forwards
VM-Exits to L1 as necessary (commit bf653b78f960, "KVM: vmx: Introduce
handle_unexpected_vmexit and handle WAITPKG vmexit").
Note, KVM always keeps the hosts MSR_IA32_UMWAIT_CONTROL resident in
hardware, i.e. always runs both L1 and L2 with the host's power management
settings for TPAUSE and UMWAIT. See commit bf09fb6cba4f ("KVM: VMX: Stop
context switching MSR_IA32_UMWAIT_CONTROL") for more details.
Fixes: e69e72faa3a0 ("KVM: x86: Add support for user wait instructions") Cc: stable@vger.kernel.org Reported-by: Aaron Lewis <aaronlewis@google.com> Reported-by: Yu Zhang <yu.c.zhang@linux.intel.com> Signed-off-by: Sean Christopherson <seanjc@google.com> Reviewed-by: Jim Mattson <jmattson@google.com>
Message-Id: <20221213062306.667649-2-seanjc@google.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Inject #GP for if VMXON is attempting with a CR0/CR4 that fails the
generic "is CRx valid" check, but passes the CR4.VMXE check, and do the
generic checks _after_ handling the post-VMXON VM-Fail.
The CR4.VMXE check, and all other #UD cases, are special pre-conditions
that are enforced prior to pivoting on the current VMX mode, i.e. occur
before interception if VMXON is attempted in VMX non-root mode.
All other CR0/CR4 checks generate #GP and effectively have lower priority
than the post-VMXON check.
Per the SDM:
IF (register operand) or (CR0.PE = 0) or (CR4.VMXE = 0) or ...
THEN #UD;
ELSIF not in VMX operation
THEN
IF (CPL > 0) or (in A20M mode) or
(the values of CR0 and CR4 are not supported in VMX operation)
THEN #GP(0);
ELSIF in VMX non-root operation
THEN VMexit;
ELSIF CPL > 0
THEN #GP(0);
ELSE VMfail("VMXON executed in VMX root operation");
FI;
which, if re-written without ELSIF, yields:
IF (register operand) or (CR0.PE = 0) or (CR4.VMXE = 0) or ...
THEN #UD
IF in VMX non-root operation
THEN VMexit;
IF CPL > 0
THEN #GP(0)
IF in VMX operation
THEN VMfail("VMXON executed in VMX root operation");
IF (in A20M mode) or
(the values of CR0 and CR4 are not supported in VMX operation)
THEN #GP(0);
Note, KVM unconditionally forwards VMXON VM-Exits that occur in L2 to L1,
i.e. there is no need to check the vCPU is not in VMX non-root mode. Add
a comment to explain why unconditionally forwarding such exits is
functionally correct.
Reported-by: Eric Li <ercli@ucdavis.edu> Fixes: c7d855c2aff2 ("KVM: nVMX: Inject #UD if VMXON is attempted with incompatible CR0/CR4") Cc: stable@vger.kernel.org Signed-off-by: Sean Christopherson <seanjc@google.com> Link: https://lore.kernel.org/r/20221006001956.329314-1-seanjc@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Resume the guest immediately when injecting a #GP on ECREATE due to an
invalid enclave size, i.e. don't attempt ECREATE in the host. The #GP is
a terminal fault, e.g. skipping the instruction if ECREATE is successful
would result in KVM injecting #GP on the instruction following ECREATE.
Fixes: 70210c044b4e ("KVM: VMX: Add SGX ENCLS[ECREATE] handler to enforce CPUID restrictions") Cc: stable@vger.kernel.org Cc: Kai Huang <kai.huang@intel.com> Signed-off-by: Sean Christopherson <seanjc@google.com> Reviewed-by: Kai Huang <kai.huang@intel.com> Link: https://lore.kernel.org/r/20220930233132.1723330-1-seanjc@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
"linux,initrd-start" and "linux,initrd-end" can be 32-bit values even on
a 64-bit platform. Ideally, the size should be based on
'#address-cells', but that has never been enforced in the kernel's FDT
boot parsing code (early_init_dt_check_for_initrd()). Bootloader
behavior is known to vary. For example, kexec always writes these as
64-bit. The result of incorrectly reading 32-bit values is most likely
the reserved memory for the original initrd will still be reserved
for the new kernel. The original arm64 equivalent of this code failed to
release the initrd reserved memory in *all* cases.
Use of_read_number() to mirror the early_init_dt_check_for_initrd()
code.
Fixes: b30be4dc733e ("of: Add a common kexec FDT setup function") Cc: stable@vger.kernel.org Reported-by: Peter Maydell <peter.maydell@linaro.org> Link: https://lore.kernel.org/r/20221128202440.1411895-1-robh@kernel.org Signed-off-by: Rob Herring <robh@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When generate a synthetic event with many params and then create a trace
action for it [1], kernel panic happened [2].
It is because that in trace_action_create() 'data->n_params' is up to
SYNTH_FIELDS_MAX (current value is 64), and array 'data->var_ref_idx'
keeps indices into array 'hist_data->var_refs' for each synthetic event
param, but the length of 'data->var_ref_idx' is TRACING_MAP_VARS_MAX
(current value is 16), so out-of-bound write happened when 'data->n_params'
more than 16. In this case, 'data->match_data.event' is overwritten and
eventually cause the panic.
To solve the issue, adjust the length of 'data->var_ref_idx' to be
SYNTH_FIELDS_MAX and add sanity checks to avoid out-of-bound write.
[1]
# cd /sys/kernel/tracing/
# echo "my_synth_event int v1; int v2; int v3; int v4; int v5; int v6;\
int v7; int v8; int v9; int v10; int v11; int v12; int v13; int v14;\
int v15; int v16; int v17; int v18; int v19; int v20; int v21; int v22;\
int v23; int v24; int v25; int v26; int v27; int v28; int v29; int v30;\
int v31; int v32; int v33; int v34; int v35; int v36; int v37; int v38;\
int v39; int v40; int v41; int v42; int v43; int v44; int v45; int v46;\
int v47; int v48; int v49; int v50; int v51; int v52; int v53; int v54;\
int v55; int v56; int v57; int v58; int v59; int v60; int v61; int v62;\
int v63" >> synthetic_events
# echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="bash"' >> \
events/sched/sched_waking/trigger
# echo "hist:keys=next_pid:onmatch(sched.sched_waking).my_synth_event(\
pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\
pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\
pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\
pid,pid,pid,pid,pid,pid,pid,pid,pid)" >> events/sched/sched_switch/trigger
Otherwise the commit that will be aborted will be associated with the
metadata objects that will be torn down. Must write needs_check flag
to metadata with a reset block manager.
Found through code-inspection (and compared against dm-thin.c).
When dm_resume() and dm_destroy() are concurrent, it will
lead to UAF, as follows:
BUG: KASAN: use-after-free in __run_timers+0x173/0x710
Write of size 8 at addr ffff88816d9490f0 by task swapper/0/0
<snip>
Call Trace:
<IRQ>
dump_stack_lvl+0x73/0x9f
print_report.cold+0x132/0xaa2
_raw_spin_lock_irqsave+0xcd/0x160
__run_timers+0x173/0x710
kasan_report+0xad/0x110
__run_timers+0x173/0x710
__asan_store8+0x9c/0x140
__run_timers+0x173/0x710
call_timer_fn+0x310/0x310
pvclock_clocksource_read+0xfa/0x250
kvm_clock_read+0x2c/0x70
kvm_clock_get_cycles+0xd/0x20
ktime_get+0x5c/0x110
lapic_next_event+0x38/0x50
clockevents_program_event+0xf1/0x1e0
run_timer_softirq+0x49/0x90
__do_softirq+0x16e/0x62c
__irq_exit_rcu+0x1fa/0x270
irq_exit_rcu+0x12/0x20
sysvec_apic_timer_interrupt+0x8e/0xc0
One of the concurrency UAF can be shown as below:
use free
do_resume |
__find_device_hash_cell |
dm_get |
atomic_inc(&md->holders) |
| dm_destroy
| __dm_destroy
| if (!dm_suspended_md(md))
| atomic_read(&md->holders)
| msleep(1)
dm_resume |
__dm_resume |
dm_table_resume_targets |
pool_resume |
do_waker #add delay work |
dm_put |
atomic_dec(&md->holders) |
| dm_table_destroy
| pool_dtr
| __pool_dec
| __pool_destroy
| destroy_workqueue
| kfree(pool) # free pool
time out
__do_softirq
run_timer_softirq # pool has already been freed
This can be easily reproduced using:
1. create thin-pool
2. dmsetup suspend pool
3. dmsetup resume pool
4. dmsetup remove_all # Concurrent with 3
The root cause of this UAF bug is that dm_resume() adds timer after
dm_destroy() skips cancelling the timer because of suspend status.
After timeout, it will call run_timer_softirq(), however pool has
already been freed. The concurrency UAF bug will happen.
Therefore, cancelling timer again in __pool_destroy().
Following process may generate a broken btree mixed with fresh and
stale btree nodes, which could get dm thin trapped in an infinite loop
while looking up data block:
Transaction 1: pmd->root = A, A->B->C // One path in btree
pmd->root = X, X->Y->Z // Copy-up
Transaction 2: X,Z is updated on disk, Y write failed.
// Commit failed, dm thin becomes read-only.
process_bio_read_only
dm_thin_find_block
__find_block
dm_btree_lookup(pmd->root)
The pmd->root points to a broken btree, Y may contain stale node
pointing to any block, for example X, which gets dm thin trapped into
a dead loop while looking up Z.
Fix this by setting pmd->root in __open_metadata(), so that dm thin
will use the last transaction's pmd->root if commit failed.
Function metadata_operation_failed() is called when operations failed
on dm pool metadata, dm pool will destroy and recreate metadata. So,
shrinker will be unregistered and registered, which could down write
shrinker_rwsem under pmd_write_lock.
Fix it by allocating dm_block_manager before locking pmd->root_lock
and destroying old dm_block_manager after unlocking pmd->root_lock,
then old dm_block_manager is replaced with new dm_block_manager under
pmd->root_lock. So, shrinker register/unregister could be done without
holding pmd->root_lock.
Fetch a reproducer in [Link].
Link: https://bugzilla.kernel.org/show_bug.cgi?id=216676 Cc: stable@vger.kernel.org #v5.2+ Fixes: e49e582965b3 ("dm thin: add read only and fail io modes") Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com> Signed-off-by: Mike Snitzer <snitzer@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Same ABBA deadlock pattern fixed in commit 4b60f452ec51 ("dm thin: Fix
ABBA deadlock between shrink_slab and dm_pool_abort_metadata") to
DM-cache's metadata.
To ease the maintenance, it is often recommended to avoid having #ifdef
preprocessor conditions.
Here the section related to CONFIG_MPTCP was quite short but the next
commit needs to add more code around. It is then cleaner to move
specific MPTCP code to functions located in net/mptcp directory.
Now that mptcp_subflow_request_sock_ops structure can be static, it can
also be marked as "read only after init".
Suggested-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com> Cc: stable@vger.kernel.org Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch fixes a race if we get two times an socket data ready event
while the listen connection worker is queued. Currently it will be
served only once but we need to do it (in this case twice) until we hit
-EAGAIN which tells us there is no pending accept going on.
This patch wraps an do while loop until we receive a return value which
is different than 0 as it was done before commit d11ccd451b65 ("fs: dlm:
listen socket out of connection hash").
Cc: stable@vger.kernel.org Fixes: d11ccd451b65 ("fs: dlm: listen socket out of connection hash") Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch fixes a double sock_release() call when the listen() is
called for the dlm lowcomms listen socket. The caller of
dlm_listen_for_all should never care about releasing the socket if
dlm_listen_for_all() fails, it's done now only once if listen() fails.
The Dell Latiture 3340/3440/3540 laptops with Realtek ALC3204 have
dual codecs and need the ALC1220_FIXUP_GB_DUAL_CODECS to fix the
conflicts of Master controls. The existing headset mic fixup for
Dell is also required to enable the jack sense and the headset mic.
Introduce a new fixup to fix the dual codec and headset mic issues
for particular Dell laptops since other old Dell laptops with the
same codec configuration are already well handled by the fixup in
alc269_fallback_pin_fixup_tbl[].
The Dell Inspiron Plus 16, in both laptop and 2in1 form factor, has top
speakers connected on NID 0x17, which the codec reports as unconnected.
These speakers should be connected to the DAC on NID 0x03.
Signed-off-by: Philipp Jungkamp <p.jungkamp@gmx.net> Link: https://lore.kernel.org/r/20221205163713.7476-1-p.jungkamp@gmx.net Signed-off-by: Takashi Iwai <tiwai@suse.de>
Stable-dep-of: a4517c4f3423 ("ALSA: hda/realtek: Apply dual codec fixup for Dell Latitude laptops") Signed-off-by: Sasha Levin <sashal@kernel.org>
In cpufreq_policy_alloc(), it will call uninitialed completion in
cpufreq_sysfs_release() when kobject_init_and_add() fails. And
that will cause a crash such as the following page fault in complete:
The member void *data in the structure devfreq can be overwrite
by governor_userspace. For example:
1. The device driver assigned the devfreq governor to simple_ondemand
by the function devfreq_add_device() and init the devfreq member
void *data to a pointer of a static structure devfreq_simple_ondemand_data
by the function devfreq_add_device().
2. The user changed the devfreq governor to userspace by the command
"echo userspace > /sys/class/devfreq/.../governor".
3. The governor userspace alloced a dynamic memory for the struct
userspace_data and assigend the member void *data of devfreq to
this memory by the function userspace_init().
4. The user changed the devfreq governor back to simple_ondemand
by the command "echo simple_ondemand > /sys/class/devfreq/.../governor".
5. The governor userspace exited and assigned the member void *data
in the structure devfreq to NULL by the function userspace_exit().
6. The governor simple_ondemand fetched the static information of
devfreq_simple_ondemand_data in the function
devfreq_simple_ondemand_func() but the member void *data of devfreq was
assigned to NULL by the function userspace_exit().
7. The information of upthreshold and downdifferential is lost
and the governor simple_ondemand can't work correctly.
The member void *data in the structure devfreq is designed for
a static pointer used in a governor and inited by the function
devfreq_add_device(). This patch add an element named governor_data
in the devfreq structure which can be used by a governor(E.g userspace)
who want to assign a private data to do some private things.
The pin configuration (done with generic pin controller helpers and
as expressed by bindings) requires children nodes with either:
1. "pins" property and the actual configuration,
2. another set of nodes with above point.
The qup_i2c12_default pin configuration used second method - with a
"pinmux" child.
Sparse reports that calling add_device_randomness() on `uid` is a
violation of address spaces. And indeed the next usage uses readl()
properly, but that was left out when passing it toadd_device_
randomness(). So instead copy the whole thing to the stack first.
If a file consists of an inline extent followed by a regular or prealloc
extent, then a legitimate attempt to resolve a logical address in the
non-inline region will result in add_all_parents reading the invalid
offset field of the inline extent. If the inline extent item is placed
in the leaf eb s.t. it is the first item, attempting to access the
offset field will not only be meaningless, it will go past the end of
the eb and cause this panic:
The pin configuration (done with generic pin controller helpers and
as expressed by bindings) requires children nodes with either:
1. "pins" property and the actual configuration,
2. another set of nodes with above point.
The qup_spi2_default pin configuration uses alreaady the second method
with a "pinmux" child, so configure drive-strength similarly in
"pinconf". Otherwise the PIN drive strength would not be applied.
Fixes: 8d23a0040475 ("arm64: dts: qcom: db845c: add Low speed expansion i2c and spi nodes") Cc: <stable@vger.kernel.org> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Douglas Anderson <dianders@chromium.org> Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org> Signed-off-by: Bjorn Andersson <andersson@kernel.org> Link: https://lore.kernel.org/r/20221010114417.29859-2-krzysztof.kozlowski@linaro.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Current clear_attr_update procedure in pmu_set_mapping() sets attr_update
field in NULL that is not correct because intel_uncore_type pmu types can
contain several groups in attr_update field. For example, SPR platform
already has uncore_alias_group to update and then UPI topology group will
be added in next patches.
Fix current behavior and clear attr_update group related to mapping only.
Fixes: bb42b3d39781 ("perf/x86/intel/uncore: Expose an Uncore unit to IIO PMON mapping") Signed-off-by: Alexander Antonov <alexander.antonov@linux.intel.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Reviewed-by: Kan Liang <kan.liang@linux.intel.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20221117122833.3103580-4-alexander.antonov@linux.intel.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The print format error was found when using ftrace event:
<...>-1406 [000] .... 23599442.895823: jbd2_end_commit: dev 252,8 transaction -1866216965 sync 0 head -1866217368
<...>-1406 [000] .... 23599442.896299: jbd2_start_commit: dev 252,8 transaction -1866216964 sync 0
Use the correct print format for transaction, head and tid.
Fixes: 879c5e6b7cb4 ('jbd2: convert instrumentation from markers to tracepoints') Signed-off-by: Bixuan Cui <cuibixuan@linux.alibaba.com> Reviewed-by: Jason Yan <yanaijie@huawei.com> Link: https://lore.kernel.org/r/1665488024-95172-1-git-send-email-cuibixuan@linux.alibaba.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
After a full run of a make_min_config test, I noticed there were a lot of
CONFIGs still enabled that really should not be. Looking at them, I
noticed they were all defined as "default y". The issue is that the test
simple removes the config and re-runs make oldconfig, which enables it
again because it is set to default 'y'. Instead, explicitly disable the
config with writing "# CONFIG_FOO is not set" to the file to keep it from
being set again.
With this change, one of my box's minconfigs went from 768 configs set,
down to 521 configs set.
Link: https://lkml.kernel.org/r/20221202115936.016fce23@gandalf.local.home Cc: stable@vger.kernel.org Fixes: 0a05c769a9de5 ("ktest: Added config_bisect test type") Reviewed-by: John 'Warthog9' Hawley (VMware) <warthog9@eaglescrag.net> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
projects / linux.git / log