The patch d0be8347c623: "Bluetooth: L2CAP: Fix use-after-free caused
by l2cap_chan_put" from Jul 21, 2022, leads to the following Smatch
static checker warning:
net/bluetooth/l2cap_core.c:1977 l2cap_global_chan_by_psm()
error: we previously assumed 'c' could be null (see line 1996)
Fixes: d0be8347c623 ("Bluetooth: L2CAP: Fix use-after-free caused by l2cap_chan_put") Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Potentially, someone may create as many pbuf bucket as there are indexes
in an xarray without any other restrictions bounding our memory usage,
put memory needed for the buckets under memory accounting.
There is issue as follows when test f2fs atomic write:
F2FS-fs (loop0): Can't find valid F2FS filesystem in 2th superblock
F2FS-fs (loop0): invalid crc_offset: 0
F2FS-fs (loop0): f2fs_check_nid_range: out-of-range nid=1, run fsck to fix.
F2FS-fs (loop0): f2fs_check_nid_range: out-of-range nid=2, run fsck to fix.
==================================================================
BUG: KASAN: null-ptr-deref in f2fs_get_dnode_of_data+0xac/0x16d0
Read of size 8 at addr 0000000000000028 by task rep/1990
As 3db1de0e582c commit changed atomic write way which new a cow_inode for
atomic write file, and also mark cow_inode as FI_ATOMIC_FILE.
When f2fs_do_write_data_page write cow_inode will use cow_inode's cow_inode
which is NULL. Then will trigger null-ptr-deref.
To solve above issue, introduce FI_COW_FILE flag for COW inode.
Fiexes: 3db1de0e582c("f2fs: change the current atomic write way") Signed-off-by: Ye Bin <yebin10@huawei.com> Reviewed-by: Chao Yu <chao@kernel.org> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
sparse reports
drivers/gpu/drm/vc4/vc4_drv.c:270:27: warning: symbol 'vc4_dma_range_matches' was not declared. Should it be static?
vc4_dma_range_matches is only used in vc4_drv.c, so it's storage class specifier
should be static.
Fixes: da8e393e23ef ("drm/vc4: drv: Adopt the dma configuration from the HVS or V3D component") Signed-off-by: Tom Rix <trix@redhat.com> Signed-off-by: Maxime Ripard <maxime@cerno.tech> Link: https://patchwork.freedesktop.org/patch/msgid/20220629200101.498138-1-trix@redhat.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
F2FS_IOC_ABORT_VOLATILE_WRITE was used to abort a atomic write before.
However it was removed accidentally. So revive it by changing the name,
since volatile write had gone.
Simon reports that if two LAN9514 USB adapters are directly connected
without an intermediate switch, the link fails to come up and link LEDs
remain dark. The issue was introduced by commit 1ce8b37241ed ("usbnet:
smsc95xx: Forward PHY interrupts to PHY driver to avoid polling").
The PHY suffers from a known erratum wherein link detection becomes
unreliable if Energy Detect Power-Down is used. In poll mode, the
driver works around the erratum by briefly disabling EDPD for 640 msec
to detect a neighbor, then re-enabling it to save power.
In interrupt mode, no interrupt is signaled if EDPD is used by both link
partners, so it must not be enabled at all.
We'll recoup the power savings by enabling SUSPEND1 mode on affected
LAN95xx chips in a forthcoming commit.
Fixes: 1ce8b37241ed ("usbnet: smsc95xx: Forward PHY interrupts to PHY driver to avoid polling") Reported-by: Simon Han <z.han@kunbus.com> Signed-off-by: Lukas Wunner <lukas@wunner.de> Link: https://lore.kernel.org/r/439a3f3168c2f9d44b5fd9bb8d2b551711316be6.1655714438.git.lukas@wunner.de Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Per toshiba,tc358767.yaml DT binding document, port@2 the output (e)DP
port is optional. In case this port is not described in DT, the bridge
driver operates in DPI-to-DP mode. The drm_of_find_panel_or_bridge()
call in tc_probe_edp_bridge_endpoint() returns -ENODEV in case port@2
is not present in DT and this specific return value is incorrectly
propagated outside of tc_probe_edp_bridge_endpoint() function. All
other error values must be propagated and are propagated correctly.
Return 0 in case the port@2 is missing instead, that reinstates the
original behavior before the commit this patch fixes.
Fixes: 8478095a8c4b ("drm/bridge: tc358767: Move (e)DP bridge endpoint parsing into dedicated function") Signed-off-by: Marek Vasut <marex@denx.de> Cc: Jonas Karlman <jonas@kwiboo.se> Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com> Cc: Lucas Stach <l.stach@pengutronix.de> Cc: Marek Vasut <marex@denx.de> Cc: Maxime Ripard <maxime@cerno.tech> Cc: Neil Armstrong <narmstrong@baylibre.com> Cc: Robert Foss <robert.foss@linaro.org> Cc: Sam Ravnborg <sam@ravnborg.org> Reviewed-by: Lucas Stach <l.stach@pengutronix.de> Link: https://patchwork.freedesktop.org/patch/msgid/20220428213132.447890-1-marex@denx.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
clang 14 won't build because ret is uninitialised and can be returned if
both prop and fdtprop are NULL. Drop the ret variable and return an
error in that failure case.
Fixes: b1fc44eaa9ba ("pseries/iommu/ddw: Fix kdump to work in absence of ibm,dma-window") Suggested-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Russell Currey <ruscur@russell.cc> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20220810054331.373761-1-ruscur@russell.cc Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
min_gate_len[0] and min_gate_len[1] should be 200000, while
min_gate_len[2-7] should be 0.
However what happens is that min_gate_len[0] is 200000, but
min_gate_len[1] ends up being 0 (despite gate_len[1] being 200000 at the
point where the logic detects the gate close event for TC 1).
The problem is that the code considers a "gate close" event whenever it
sees that there is a 0 for that TC (essentially it's level rather than
edge triggered). By doing that, any time a gate is seen as closed
without having been open prior, gate_len, which is 0, will be written
into min_gate_len. Once min_gate_len becomes 0, it's impossible for it
to track anything higher than that (the length of actually open
intervals).
To fix this, we make the writing to min_gate_len[tc] be edge-triggered,
which avoids writes for gates that are closed in consecutive intervals.
However what this does is it makes us need to special-case the
permanently closed gates at the end.
Fixes: 55a515b1f5a9 ("net: dsa: felix: drop oversized frames with tc-taprio instead of hanging the port") Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://lore.kernel.org/r/20220804202817.1677572-1-vladimir.oltean@nxp.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If an instance of tracing enables the same trace event as another
instance, or the top level instance, or even perf, then the va_list passed
into some tracepoints can be used more than once.
As va_list can only be traversed once, this can cause issues:
Similar to commit 7c80b038d23e ("net: fix sk_wmem_schedule() and
sk_rmem_schedule() errors"), let the MPTCP receive path schedule
exactly the required amount of memory.
Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This reverts commit a52ed4866d2b90dd5e4ae9dabd453f3ed8fa3cbc as it
causes build problems in linux-next. It needs to be reintroduced in a
way that can allow the api to evolve and not require a "flag day" to
catch all users.
Link: https://lore.kernel.org/r/20220623160723.7a44b573@canb.auug.org.au Cc: Duoming Zhou <duoming@zju.edu.cn> Cc: Brian Norris <briannorris@chromium.org> Cc: Johannes Berg <johannes@sipsolutions.net> Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
I accidentally broke IPv4 traceroute, by swapping iph->saddr
and iph->daddr.
Probably because raw_icmp_error() and raw_v4_input()
use different order for iph->saddr and iph->daddr.
Fixes: ba44f8182ec2 ("raw: use more conventional iterators") Reported-by: John Sperbeck <jsperbeck@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Link: https://lore.kernel.org/r/20220623193540.2851799-1-edumazet@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fixes: ba44f8182ec2 ("raw: use more conventional iterators") Reported-by: kernel test robot <lkp@intel.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Acked-by: Jonathan Lemon <jonathan.lemon@gmail.com> Link: https://lore.kernel.org/r/20220622032303.159394-1-edumazet@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Using 3 blocks here doesn't give us much more than using 2, and it
causes a stack frame size warning on certain compiler/config/arch
combinations:
lib/crypto/blake2s-selftest.c: In function 'blake2s_selftest':
>> lib/crypto/blake2s-selftest.c:632:1: warning: the frame size of 1088 bytes is larger than 1024 bytes [-Wframe-larger-than=]
632 | }
| ^
So this patch just reduces the block from 3 to 2, which makes the
warning go away.
Reported-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/linux-crypto/202206200851.gE3MHCgd-lkp@intel.com Fixes: 2d16803c562e ("crypto: blake2s - remove shash module") Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
sk_forced_mem_schedule() has a bug similar to ones fixed
in commit 7c80b038d23e ("net: fix sk_wmem_schedule() and
sk_rmem_schedule() errors")
While this bug has little chance to trigger in old kernels,
we need to fix it before the following patch.
Fixes: d83769a580f1 ("tcp: fix possible deadlock in tcp_send_fin()") Signed-off-by: Eric Dumazet <edumazet@google.com> Acked-by: Soheil Hassas Yeganeh <soheil@google.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Wei Wang <weiwan@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When a route filter is replaced and the old filter has a 0 handle, the old
one won't be removed from the hashtable, while it will still be freed.
The test was there since before commit 1109c00547fc ("net: sched: RCU
cls_route"), when a new filter was not allocated when there was an old one.
The old filter was reused and the reinserting would only be necessary if an
old filter was replaced. That was still wrong for the same case where the
old handle was 0.
Remove the old filter from the list independently from its handle value.
This fixes CVE-2022-2588, also reported as ZDI-CAN-17440.
Reported-by: Zhenpeng Lin <zplin@u.northwestern.edu> Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com> Reviewed-by: Kamal Mostafa <kamal@canonical.com> Cc: <stable@vger.kernel.org> Acked-by: Jamal Hadi Salim <jhs@mojatatu.com> Link: https://lore.kernel.org/r/20220809170518.164662-1-cascardo@canonical.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If count_max_extents() uses BTRFS_MAX_EXTENT_SIZE to calculate the number
of extents needed, btrfs release the metadata reservation too much on its
way to write out the data.
Now that BTRFS_MAX_EXTENT_SIZE is replaced with fs_info->max_extent_size,
convert count_max_extents() to use it instead, and fix the calculation of
the metadata reservation.
CC: stable@vger.kernel.org # 5.12+ Fixes: d8e3fb106f39 ("btrfs: zoned: use ZONE_APPEND write for zoned mode") Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When logging a new name, in case of a rename, we pin the log before
changing it. We then either delete a directory entry from the log or
insert a key range item to mark the old name for deletion on log replay.
However when doing one of those log changes we may have another task that
started writing out the log (at btrfs_sync_log()) and it started before
we pinned the log root. So we may end up changing a log tree while its
writeback is being started by another task syncing the log. This can lead
to inconsistencies in a log tree and other unexpected results during log
replay, because we can get some committed node pointing to a node/leaf
that ends up not getting written to disk before the next log commit.
The problem, conceptually, started to happen in commit 88d2beec7e53fc
("btrfs: avoid logging all directory changes during renames"), because
there we started to update the log without joining its current transaction
first.
However the problem only became visible with commit 259c4b96d78dda
("btrfs: stop doing unnecessary log updates during a rename"), and that is
because we used to pin the log at btrfs_rename() and then before entering
btrfs_log_new_name(), when unlinking the old dentry, we ended up at
btrfs_del_inode_ref_in_log() and btrfs_del_dir_entries_in_log(). Both
of them join the current log transaction, effectively waiting for any log
transaction writeout (due to acquiring the root's log_mutex). This made it
safe even after leaving the current log transaction, because we remained
with the log pinned when we called btrfs_log_new_name().
Then in commit 259c4b96d78dda ("btrfs: stop doing unnecessary log updates
during a rename"), we removed the log pinning from btrfs_rename() and
stopped calling btrfs_del_inode_ref_in_log() and
btrfs_del_dir_entries_in_log() during the rename, and started to do all
the needed work at btrfs_log_new_name(), but without joining the current
log transaction, only pinning the log, which is racy because another task
may have started writeout of the log tree right before we pinned the log.
Both commits landed in kernel 5.18, so it doesn't make any practical
difference which should be blamed, but I'm blaming the second commit only
because with the first one, by chance, the problem did not happen due to
the fact we joined the log transaction after pinning the log and unpinned
it only after calling btrfs_log_new_name().
So make btrfs_log_new_name() join the current log transaction instead of
pinning it, so that we never do log updates if it's writeout is starting.
Fixes: 259c4b96d78dda ("btrfs: stop doing unnecessary log updates during a rename") CC: stable@vger.kernel.org # 5.18+ Reported-by: Zygo Blaxell <ce3g8jdj@umail.furryterror.org> Tested-by: Zygo Blaxell <ce3g8jdj@umail.furryterror.org> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 9401f4e46cf6 ("powerpc: Use lwarx/ldarx directly instead of
PPC_LWARX/LDARX macros") properly handled the eh field of lwarx
in asm/bitops.h but failed to clear it for PPC32 in
asm/simple_spinlock.h
So, do as in arch_atomic_try_cmpxchg_lock(), set it to 1 if PPC64
but set it to 0 if PPC32. For that use IS_ENABLED(CONFIG_PPC64) which
returns 1 when CONFIG_PPC64 is set and 0 otherwise.
Fixes: 9401f4e46cf6 ("powerpc: Use lwarx/ldarx directly instead of PPC_LWARX/LDARX macros") Cc: stable@vger.kernel.org # v5.15+ Reported-by: Pali Rohár <pali@kernel.org> Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Tested-by: Pali Rohár <pali@kernel.org> Reviewed-by: Segher Boessenkool <segher@kernel.crashing.org>
[mpe: Use symbolic names, use 'n' constraint per Segher] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/a1176e19e627dd6a1b8d24c6c457a8ab874b7d12.1659430931.git.christophe.leroy@csgroup.eu Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In some use cases[1], the backend is created while the frontend doesn't
support the persistent grants feature, but later the frontend can be
changed to support the feature and reconnect. In the past, 'blkback'
enabled the persistent grants feature since it unconditionally checked
if frontend supports the persistent grants feature for every connect
('connect_ring()') and decided whether it should use persistent grans or
not.
However, commit aac8a70db24b ("xen-blkback: add a parameter for
disabling of persistent grants") has mistakenly changed the behavior.
It made the frontend feature support check to not be repeated once it
shown the 'feature_persistent' as 'false', or the frontend doesn't
support persistent grants.
Similar behavioral change has made on 'blkfront' by commit 74a852479c68
("xen-blkfront: add a parameter for disabling of persistent grants").
This commit changes the behavior of the parameter to make effect for
every connect, so that the previous behavior of 'blkfront' can be
restored.
In some use cases[1], the backend is created while the frontend doesn't
support the persistent grants feature, but later the frontend can be
changed to support the feature and reconnect. In the past, 'blkback'
enabled the persistent grants feature since it unconditionally checked
if frontend supports the persistent grants feature for every connect
('connect_ring()') and decided whether it should use persistent grans or
not.
However, commit aac8a70db24b ("xen-blkback: add a parameter for
disabling of persistent grants") has mistakenly changed the behavior.
It made the frontend feature support check to not be repeated once it
shown the 'feature_persistent' as 'false', or the frontend doesn't
support persistent grants.
This commit changes the behavior of the parameter to make effect for
every connect, so that the previous workflow can work again as expected.
Persistent grants feature can be used only when both backend and the
frontend supports the feature. The feature was always supported by
'blkback', but commit aac8a70db24b ("xen-blkback: add a parameter for
disabling of persistent grants") has introduced a parameter for
disabling it runtime.
To avoid the parameter be updated while being used by 'blkback', the
commit caches the parameter into 'vbd->feature_gnt_persistent' in
'xen_vbd_create()', and then check if the guest also supports the
feature and finally updates the field in 'connect_ring()'.
However, 'connect_ring()' could be called before 'xen_vbd_create()', so
later execution of 'xen_vbd_create()' can wrongly overwrite 'true' to
'vbd->feature_gnt_persistent'. As a result, 'blkback' could try to use
'persistent grants' feature even if the guest doesn't support the
feature.
This commit fixes the issue by moving the parameter value caching to
'xen_blkif_alloc()', which allocates the 'blkif'. Because the struct
embeds 'vbd' object, which will be used by 'connect_ring()' later, this
should be called before 'connect_ring()' and therefore this should be
the right and safe place to do the caching.
Fixes: aac8a70db24b ("xen-blkback: add a parameter for disabling of persistent grants") Cc: <stable@vger.kernel.org> # 5.10.x Signed-off-by: Maximilian Heyne <mheyne@amazon.de> Signed-off-by: SeongJae Park <sj@kernel.org> Reviewed-by: Maximilian Heyne <mheyne@amazon.de> Reviewed-by: Juergen Gross <jgross@suse.com> Link: https://lore.kernel.org/r/20220715225108.193398-2-sj@kernel.org Signed-off-by: Juergen Gross <jgross@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In commit 0aa698787aa2 ("tpm: Add Upgrade/Reduced mode support for
TPM2 modules") it was said that:
"If the TPM is in Failure mode, it will successfully respond to both
tpm2_do_selftest() and tpm2_startup() calls. Although, will fail to
answer to tpm2_get_cc_attrs_tbl(). Use this fact to conclude that TPM
is in Failure mode."
But a check was never added in the commit when calling
tpm2_get_cc_attrs_tbl() to conclude that the TPM is in Failure mode.
This commit corrects this by adding a check.
Fixes: 0aa698787aa2 ("tpm: Add Upgrade/Reduced mode support for TPM2 modules") Cc: stable@vger.kernel.org # v5.17+ Signed-off-by: Mårten Lindahl <marten.lindahl@axis.com> Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org> Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
If DEBUG_SECTION_MISMATCH enabled, __calc_tpm2_event_size() will not be
inlined, this cause section mismatch like this:
WARNING: modpost: vmlinux.o(.text.unlikely+0xe30c): Section mismatch in reference from the variable L0 to the function .init.text:early_ioremap()
The function L0() references
the function __init early_memremap().
This is often because L0 lacks a __init
annotation or the annotation of early_ioremap is wrong.
Fix it by using __always_inline instead of inline for the called-once
function __calc_tpm2_event_size().
The signature verification of SM2 needs to add the Za value and
recalculate sig->digest, which requires the detection of the pkey_algo
in public_key_verify_signature(). As Eric Biggers said, the pkey_algo
field in sig is attacker-controlled and should be use pkey->pkey_algo
instead of sig->pkey_algo, and secondly, if sig->pkey_algo is NULL, it
will also cause signature verification failure.
The software_key_determine_akcipher() already forces the algorithms
are matched, so the SM3 algorithm is enforced in the SM2 signature,
although this has been checked, we still avoid using any algorithm
information in the signature as input.
Fixes: 215525639631 ("X.509: support OSCCA SM2-with-SM3 certificate verification") Reported-by: Eric Biggers <ebiggers@google.com> Cc: stable@vger.kernel.org # v5.10+ Signed-off-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com> Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org> Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Later the code in ext4_xattr_block_set() finds out the block got freed
and cancels reusal of the block but the revoke stays canceled and so in
case of block reuse and journal replay the filesystem can get corrupted.
If the race works out slightly differently, we can also hit assertions
in the jbd2 code.
Fix the problem by making sure that once matching mbcache entry is
found, code dropping the last xattr block reference (or trying to modify
xattr block in place) waits until the mbcache entry reference is
dropped. This way code trying to reuse xattr block is protected from
someone trying to drop the last reference to xattr block.
Reported-and-tested-by: Ritesh Harjani <ritesh.list@gmail.com> CC: stable@vger.kernel.org Fixes: 82939d7999df ("ext4: convert to mbcache2") Signed-off-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20220712105436.32204-5-jack@suse.cz Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Sasha Levin <sashal@kernel.org>
Remove unnecessary else (and thus indentation level) from a code block
in ext4_xattr_block_set(). It will also make following code changes
easier. No functional changes.
CC: stable@vger.kernel.org Fixes: 82939d7999df ("ext4: convert to mbcache2") Signed-off-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20220712105436.32204-4-jack@suse.cz Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Sasha Levin <sashal@kernel.org>
Currently we remove EA inode from mbcache as soon as its xattr refcount
drops to zero. However there can be pending attempts to reuse the inode
and thus refcount handling code has to handle the situation when
refcount increases from zero anyway. So save some work and just keep EA
inode in mbcache until it is getting evicted. At that moment we are sure
following iget() of EA inode will fail anyway (or wait for eviction to
finish and load things from the disk again) and so removing mbcache
entry at that moment is fine and simplifies the code a bit.
CC: stable@vger.kernel.org Fixes: 82939d7999df ("ext4: convert to mbcache2") Signed-off-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20220712105436.32204-3-jack@suse.cz Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Sasha Levin <sashal@kernel.org>
ext4_append() must always allocate a new block, otherwise we run the
risk of overwriting existing directory block corrupting the directory
tree in the process resulting in all manner of problems later on.
Add a sanity check to see if the logical block is already allocated and
error out if it is.
Currently ext4 directory handling code implicitly assumes that the
directory blocks are always within the i_size. In fact ext4_append()
will attempt to allocate next directory block based solely on i_size and
the i_size is then appropriately increased after a successful
allocation.
However, for this to work it requires i_size to be correct. If, for any
reason, the directory inode i_size is corrupted in a way that the
directory tree refers to a valid directory block past i_size, we could
end up corrupting parts of the directory tree structure by overwriting
already used directory blocks when modifying the directory.
Fix it by catching the corruption early in __ext4_read_dirblock().
If the ext4 inode does not have xattr space, 0 is returned in the
get_max_inline_xattr_value_size function. Otherwise, the function returns
a negative value when the inode does not contain EXT4_STATE_XATTR.
Cc: stable@kernel.org Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20220616021358.2504451-4-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Sasha Levin <sashal@kernel.org>
Hulk Robot reported a issue:
==================================================================
BUG: KASAN: use-after-free in ext4_xattr_set_entry+0x18ab/0x3500
Write of size 4105 at addr ffff8881675ef5f4 by task syz-executor.0/7092
Above issue may happen as follows:
-------------------------------------
ext4_xattr_set
ext4_xattr_set_handle
ext4_xattr_ibody_find
>> s->end < s->base
>> no EXT4_STATE_XATTR
>> xattr_check_inode is not executed
ext4_xattr_ibody_set
ext4_xattr_set_entry
>> size_t min_offs = s->end - s->base
>> UAF in memcpy
we can easily reproduce this problem with the following commands:
mkfs.ext4 -F /dev/sda
mount -o debug_want_extra_isize=128 /dev/sda /mnt
touch /mnt/file
setfattr -n user.cat -v `seq -s z 4096|tr -d '[:digit:]'` /mnt/file
In ext4_xattr_ibody_find, we have the following assignment logic:
header = IHDR(inode, raw_inode)
= raw_inode + EXT4_GOOD_OLD_INODE_SIZE + i_extra_isize
is->s.base = IFIRST(header)
= header + sizeof(struct ext4_xattr_ibody_header)
is->s.end = raw_inode + s_inode_size
In the calculation formula, all values except s_inode_size and
i_extra_size are fixed values. When i_extra_size is the maximum value
s_inode_size - EXT4_GOOD_OLD_INODE_SIZE, min_offs is -4 and free is -8.
The value overflows. As a result, the preceding issue is triggered when
memcpy is executed.
Therefore, when finding xattr or setting xattr, check whether
there is space for storing xattr in the inode to resolve this issue.
Cc: stable@kernel.org Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20220616021358.2504451-3-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Sasha Levin <sashal@kernel.org>
When adding an xattr to an inode, we must ensure that the inode_size is
not less than EXT4_GOOD_OLD_INODE_SIZE + extra_isize + pad. Otherwise,
the end position may be greater than the start position, resulting in UAF.
Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com> Link: https://lore.kernel.org/r/20220616021358.2504451-2-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Sasha Levin <sashal@kernel.org>
A race can occur in the unlikely event ext4 is unable to allocate a
physical cluster for a delayed allocation in a bigalloc file system
during writeback. Failure to allocate a cluster forces error recovery
that includes a call to mpage_release_unused_pages(). That function
removes any corresponding delayed allocated blocks from the extent
status tree. If a new delayed write is in progress on the same cluster
simultaneously, resulting in the addition of an new extent containing
one or more blocks in that cluster to the extent status tree, delayed
block accounting can be thrown off if that delayed write then encounters
a similar cluster allocation failure during future writeback.
Write lock the i_data_sem in mpage_release_unused_pages() to fix this
problem. Ext4's block/cluster accounting code for bigalloc relies on
i_data_sem for mutual exclusion, as is found in the delayed write path,
and the locking in mpage_release_unused_pages() is missing.
Cc: stable@kernel.org Reported-by: Ye Bin <yebin10@huawei.com> Signed-off-by: Eric Whitney <enwlinux@gmail.com> Link: https://lore.kernel.org/r/20220615160530.1928801-1-enwlinux@gmail.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Sasha Levin <sashal@kernel.org>
When doing an online resize, the on-disk superblock on-disk wasn't
updated. This means that when the file system is unmounted and
remounted, and the on-disk overhead value is non-zero, this would
result in the results of statfs(2) to be incorrect.
This was partially fixed by Commits 10b01ee92df5 ("ext4: fix overhead
calculation to account for the reserved gdt blocks"), 85d825dbf489
("ext4: force overhead calculation if the s_overhead_cluster makes no
sense"), and eb7054212eac ("ext4: update the cached overhead value in
the superblock").
However, since it was too expensive to forcibly recalculate the
overhead for bigalloc file systems at every mount, this didn't fix the
problem for bigalloc file systems. This commit should address the
problem when resizing file systems with the bigalloc feature enabled.
OOB read memory can be written to a file,
if DataOffset is 0 and Length is too large
in SMB2_WRITE request of compound request.
To prevent this, when checking the length of
the data area of SMB2_WRITE in smb2_get_data_area_len(),
let the minimum of DataOffset be the size of
SMB2 header + the size of SMB2_WRITE header.
smb-direct max read/write size can be different with smb2 max read/write
size. So smb2_read() can return error by wrong max read/write size check.
This patch use smb_direct_max_read_write_size for this check in
smb-direct read/write().
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org> Reviewed-by: Hyunchul Lee <hyc.lee@gmail.com> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
SMB2_READ/SMB2_WRITE request has to be granted the number
of rw credits, the pages the request wants to transfer
/ the maximum pages which can be registered with one
MR to read and write a file.
And allocate enough RDMA resources for the maximum
number of rw credits allowed by ksmbd.
Signed-off-by: Hyunchul Lee <hyc.lee@gmail.com> Acked-by: Namjae Jeon <linkinjeon@kernel.org> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The SMB2 Write packet contains data that is to be written
to a file or to a pipe. Depending on the client, there may
be padding between the header and the data field.
Currently, the length is validated only in the case padding
is present.
Since the DataOffset field always points to the beginning
of the data, there is no need to have a special case for
padding. By removing this, the length is validated in both
cases.
Signed-off-by: Marios Makassikis <mmakassikis@freebox.fr> Acked-by: Namjae Jeon <linkinjeon@kernel.org> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
alignof() gives an alignment of types as they would be as standalone
variables. But alignment in structures might be different, and when
building the fields of events, the alignment must be the actual
alignment otherwise the field offsets may not match what they actually
are.
This caused trace-cmd to crash, as libtraceevent did not check if the
field offset was bigger than the event. The write_msr and read_msr
events on 32 bit had their fields incorrect, because it had a u64 field
between two ints. alignof(u64) would give 8, but the u64 field was at a
4 byte alignment.
Define a macro as:
ALIGN_STRUCTFIELD(type) ((int)(offsetof(struct {char a; type b;}, b)))
which gives the actual alignment of types in a structure.
Link: https://lkml.kernel.org/r/20220731015928.7ab3a154@rorschach.local.home Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: stable@vger.kernel.org Fixes: 04ae87a52074e ("ftrace: Rework event_create_dir()") Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Instead of open coding a __dynamic_array() with a fixed length (which
defeats the purpose of the dynamic array in the first place). Use the new
__vstring() helper that will use a va_list and only write enough of the
string into the ring buffer that is needed.
Link: https://lkml.kernel.org/r/20220724191650.236b1355@rorschach.local.home Cc: Marek Lindner <mareklindner@neomailbox.ch> Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Simon Wunderlich <sw@simonwunderlich.de> Cc: Antonio Quartulli <a@unstable.cc> Cc: "David S. Miller" <davem@davemloft.net> Cc: Eric Dumazet <edumazet@google.com> Cc: Jakub Kicinski <kuba@kernel.org> Cc: Paolo Abeni <pabeni@redhat.com> Cc: b.a.t.m.a.n@lists.open-mesh.org Cc: netdev@vger.kernel.org Acked-by: Sven Eckelmann <sven@narfation.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
We forget to set cft->private for numa stat file. As a result, numa stat
of hstates[0] is always showed for all hstates. Encode the hstates index
into cft->private to fix this issue.
Link: https://lkml.kernel.org/r/20220723073804.53035-1-linmiaohe@huawei.com Fixes: f47761999052 ("hugetlb: add hugetlb.*.numa_stat file") Signed-off-by: Miaohe Lin <linmiaohe@huawei.com> Acked-by: Muchun Song <songmuchun@bytedance.com> Cc: Kees Cook <keescook@chromium.org> Cc: Mike Kravetz <mike.kravetz@oracle.com> Cc: Mina Almasry <almasrymina@google.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
damon_reclaim_init() allocates a memory chunk for ctx with
damon_new_ctx(). When damon_select_ops() fails, ctx is not released,
which will lead to a memory leak.
We should release the ctx with damon_destroy_ctx() when damon_select_ops()
fails to fix the memory leak.
Link: https://lkml.kernel.org/r/20220714063746.2343549-1-niejianglei2021@163.com Fixes: 4d69c3457821 ("mm/damon/reclaim: use damon_select_ops() instead of damon_{v,p}a_set_operations()") Signed-off-by: Jianglei Nie <niejianglei2021@163.com> Reviewed-by: SeongJae Park <sj@kernel.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
There is a KASAN warning in raid_resume when running the lvm test
lvconvert-raid.sh. The reason for the warning is that mddev->raid_disks
is greater than rs->raid_disks, so the loop touches one entry beyond
the allocated length.
There is this warning when using a kernel with the address sanitizer
and running this testsuite:
https://gitlab.com/cki-project/kernel-tests/-/tree/main/storage/swraid/scsi_raid
==================================================================
BUG: KASAN: slab-out-of-bounds in raid_status+0x1747/0x2820 [dm_raid]
Read of size 4 at addr ffff888079d2c7e8 by task lvcreate/13319
CPU: 0 PID: 13319 Comm: lvcreate Not tainted 5.18.0-0.rc3.<snip> #1
Hardware name: Red Hat KVM, BIOS 0.5.1 01/01/2011
Call Trace:
<TASK>
dump_stack_lvl+0x6a/0x9c
print_address_description.constprop.0+0x1f/0x1e0
print_report.cold+0x55/0x244
kasan_report+0xc9/0x100
raid_status+0x1747/0x2820 [dm_raid]
dm_ima_measure_on_table_load+0x4b8/0xca0 [dm_mod]
table_load+0x35c/0x630 [dm_mod]
ctl_ioctl+0x411/0x630 [dm_mod]
dm_ctl_ioctl+0xa/0x10 [dm_mod]
__x64_sys_ioctl+0x12a/0x1a0
do_syscall_64+0x5b/0x80
The warning is caused by reading conf->max_nr_stripes in raid_status. The
code in raid_status reads mddev->private, casts it to struct r5conf and
reads the entry max_nr_stripes.
However, if we have different raid type than 4/5/6, mddev->private
doesn't point to struct r5conf; it may point to struct r0conf, struct
r1conf, struct r10conf or struct mpconf. If we cast a pointer to one
of these structs to struct r5conf, we will be reading invalid memory
and KASAN warns about it.
Fix this bug by reading struct r5conf only if raid type is 4, 5 or 6.
Attempt to load PERF_GLOBAL_CTRL during nested VM-Enter/VM-Exit if and
only if the MSR exists (according to the guest vCPU model). KVM has very
misguided handling of VM_{ENTRY,EXIT}_LOAD_IA32_PERF_GLOBAL_CTRL and
attempts to force the nVMX MSR settings to match the vPMU model, i.e. to
hide/expose the control based on whether or not the MSR exists from the
guest's perspective.
KVM's modifications fail to handle the scenario where the vPMU is hidden
from the guest _after_ being exposed to the guest, e.g. by userspace
doing multiple KVM_SET_CPUID2 calls, which is allowed if done before any
KVM_RUN. nested_vmx_pmu_refresh() is called if and only if there's a
recognized vPMU, i.e. KVM will leave the bits in the allow state and then
ultimately reject the MSR load and WARN.
KVM should not force the VMX MSRs in the first place. KVM taking control
of the MSRs was a misguided attempt at mimicking what commit 5f76f6f5ff96
("KVM: nVMX: Do not expose MPX VMX controls when guest MPX disabled",
2018-10-01) did for MPX. However, the MPX commit was a workaround for
another KVM bug and not something that should be imitated (and it should
never been done in the first place).
In other words, KVM's ABI _should_ be that userspace has full control
over the MSRs, at which point triggering the WARN that loading the MSR
must not fail is trivial.
The intent of the WARN is still valid; KVM has consistency checks to
ensure that vmcs12->{guest,host}_ia32_perf_global_ctrl is valid. The
problem is that '0' must be considered a valid value at all times, and so
the simple/obvious solution is to just not actually load the MSR when it
does not exist. It is userspace's responsibility to provide a sane vCPU
model, i.e. KVM is well within its ABI and Intel's VMX architecture to
skip the loads if the MSR does not exist.
Add a helper to check of the guest PMU has PERF_GLOBAL_CTRL, which is
unintuitive _and_ diverges from Intel's architecturally defined behavior.
Even worse, KVM currently implements the check using two different (but
equivalent) checks, _and_ there has been at least one attempt to add a
_third_ flavor.
Cc: stable@vger.kernel.org Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20220722224409.1336532-4-seanjc@google.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Mark all MSR_CORE_PERF_GLOBAL_CTRL and MSR_CORE_PERF_GLOBAL_OVF_CTRL bits
as reserved if there is no guest vPMU. The nVMX VM-Entry consistency
checks do not check for a valid vPMU prior to consuming the masks via
kvm_valid_perf_global_ctrl(), i.e. may incorrectly allow a non-zero mask
to be loaded via VM-Enter or VM-Exit (well, attempted to be loaded, the
actual MSR load will be rejected by intel_is_valid_msr()).
Fixes: f5132b01386b ("KVM: Expose a version 2 architectural PMU to a guests") Cc: stable@vger.kernel.org Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20220722224409.1336532-3-seanjc@google.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The mask value of fixed counter control register should be dynamic
adjusted with the number of fixed counters. This patch introduces a
variable that includes the reserved bits of fixed counter control
registers. This is a generic code refactoring.
Co-developed-by: Luwei Kang <luwei.kang@intel.com> Signed-off-by: Luwei Kang <luwei.kang@intel.com> Signed-off-by: Like Xu <like.xu@linux.intel.com> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Message-Id: <20220411101946.20262-6-likexu@tencent.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When HAVE_FUNCTION_GRAPH_RET_ADDR_PTR is defined, the return
address to the fgraph caller is recovered by tagging it along with the
stack pointer of ftrace stack. This makes the stack unwinding more
reliable.
When the fgraph return address is modified to return_to_handler,
ftrace_graph_ret_addr tries to restore it to the original
value using tagged stack pointer.
Fix this by passing tagged sp to ftrace_graph_ret_addr.
The existing logic in KVM to support guests calling H_RANDOM only works
on Power8, because it looks for an RNG in the device tree, but on Power9
we just use darn.
In addition the existing code needs to work in real mode, so we have the
special cased powernv_get_random_real_mode() to deal with that.
Instead just have KVM call ppc_md.get_random_seed(), and do the real
mode check inside of there, that way we use whatever RNG is available,
including darn on Power9.
Fixes: e928e9cb3601 ("KVM: PPC: Book3S HV: Add fast real-mode H_RANDOM implementation.") Cc: stable@vger.kernel.org # v4.1+ Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> Tested-by: Sachin Sant <sachinp@linux.ibm.com>
[mpe: Rebase on previous commit, update change log appropriately] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20220727143219.2684192-2-mpe@ellerman.id.au Signed-off-by: Sasha Levin <sashal@kernel.org>
There is a problem with the current revision checks in
is_cppc_supported() that they essentially prevent the CPPC support
from working if a new _CPC package format revision being a proper
superset of the v3 and only causing _CPC to return a package with more
entries (while retaining the types and meaning of the entries defined by
the v3) is introduced in the future and used by the platform firmware.
In that case, as long as the number of entries in the _CPC return
package is at least CPPC_V3_NUM_ENT, it should be perfectly fine to
use the v3 support code and disregard the additional package entries
added by the new package format revision.
For this reason, drop is_cppc_supported() altogether, put the revision
checks directly into acpi_cppc_processor_probe() so they are easier to
follow and rework them to take the case mentioned above into account.
Fixes: 4773e77cdc9b ("ACPI / CPPC: Add support for CPPC v3") Cc: 4.18+ <stable@vger.kernel.org> # 4.18+ Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
As that commit describes, on early Sapphire Rapids Xeon platforms the C1 and
C1E states were mutually exclusive, so that users could only have either C1 and
C6, or C1E and C6.
However, Intel firmware engineers managed to remove this limitation and make C1
and C1E to be completely independent, just like on previous Xeon platforms.
Therefore, this patch:
* Removes commentary describing the old, and now non-existing SPR C1E
limitation.
* Marks SPR C1E as available by default.
* Removes the 'preferred_cstates' parameter handling for SPR. Both C1 and
C1E will be available regardless of 'preferred_cstates' value.
We expect that all SPR systems are shipping with new firmware, which includes
the C1/C1E improvement.
Cc: v5.18+ <stable@vger.kernel.org> # v5.18+ Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Similar to SPR, the C1 and C1E states on ADL are mutually exclusive.
Only one of them can be enabled at a time.
But contrast to SPR, which usually has a strong latency requirement
as a Xeon processor, C1E is preferred on ADL for better energy
efficiency.
Add custom C-state tables for ADL with both C1 and C1E, and
1. Enable the "C1E promotion" bit in MSR_IA32_POWER_CTL and mark C1
with the CPUIDLE_FLAG_UNUSABLE flag, so C1 is not available by
default.
2. Add support for the "preferred_cstates" module parameter, so that
users can choose to use C1 instead of C1E by booting with
"intel_idle.preferred_cstates=2".
Separate custom C-state tables are introduced for the ADL mobile and
desktop processors, because of the exit latency differences between
these two variants, especially with respect to PC10.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[ rjw: Changelog edits, code rearrangement ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When the allocated position doesn't progress, we cannot submit IOs to
finish a block group, but there should be ongoing IOs that will finish a
block group. So, in that case, we wait for a zone to be finished and retry
the allocation after that.
Introduce a new flag BTRFS_FS_NEED_ZONE_FINISH for fs_info->flags to
indicate we need a zone finish to have proceeded. The flag is set when the
allocator detected it cannot activate a new block group. And, it is cleared
once a zone is finished.
CC: stable@vger.kernel.org # 5.16+ Fixes: afba2bc036b0 ("btrfs: zoned: implement active zone tracking") Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
cow_file_range() works in an all-or-nothing way: if it fails to allocate an
extent for a part of the given region, it gives up all the region including
the successfully allocated parts. On cow_file_range(), run_delalloc_zoned()
writes data for the region only when it successfully allocate all the
region.
This all-or-nothing allocation and write-out are problematic when available
space in all the block groups are get tight with the active zone
restriction. btrfs_reserve_extent() try hard to utilize the left space in
the active block groups and gives up finally and fails with
-ENOSPC. However, if we send IOs for the successfully allocated region, we
can finish a zone and can continue on the rest of the allocation on a newly
allocated block group.
This patch implements the partial write-out for run_delalloc_zoned(). With
this patch applied, cow_file_range() returns -EAGAIN to tell the caller to
do something to progress the further allocation, and tells the successfully
allocated region with done_offset. Furthermore, the zoned extent allocator
returns -EAGAIN to tell cow_file_range() going back to the caller side.
Actually, we still need to wait for an IO to complete to continue the
allocation. The next patch implements that part.
CC: stable@vger.kernel.org # 5.16+ Fixes: afba2bc036b0 ("btrfs: zoned: implement active zone tracking") Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
There are two places where allocating a chunk is not enough. These two
places are trying to ensure the space by allocating a chunk. To meet the
condition for active_total_bytes, we also need to activate a block group
there.
CC: stable@vger.kernel.org # 5.16+ Fixes: afba2bc036b0 ("btrfs: zoned: implement active zone tracking") Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
For metadata space on zoned filesystem, reaching ALLOC_CHUNK{,_FORCE}
means we don't have enough space left in the active_total_bytes. Before
allocating a new chunk, we can try to activate an existing block group
in this case.
Also, allocating a chunk is not enough to grant a ticket for metadata
space on zoned filesystem we need to activate the block group to
increase the active_total_bytes.
btrfs_zoned_activate_one_bg() implements the activation feature. It will
activate a block group by (maybe) finishing a block group. It will give up
activating a block group if it cannot finish any block group.
CC: stable@vger.kernel.org # 5.16+ Fixes: afba2bc036b0 ("btrfs: zoned: implement active zone tracking") Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The active_total_bytes, like the total_bytes, accounts for the total bytes
of active block groups in the space_info.
With an introduction of active_total_bytes, we can check if the reserved
bytes can be written to the block groups without activating a new block
group. The check is necessary for metadata allocation on zoned
filesystem. We cannot finish a block group, which may require waiting
for the current transaction, from the metadata allocation context.
Instead, we need to ensure the ongoing allocation (reserved bytes) fits
in active block groups.
The chunk size is stored in the btrfs_space_info structure. It is
initialized at the start and is then used.
A new API is added to update the current chunk size. This API is used
to be able to expose the chunk_size as a sysfs setting.
Signed-off-by: Stefan Roesch <shr@fb.com> Reviewed-by: David Sterba <dsterba@suse.com>
[ rename and merge helpers, switch atomic type to u64, style fixes ] Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
For non-zoned file systems it's useful to have the auto reclaim feature,
however there are different use cases for non-zoned, for example we may
not want to reclaim metadata chunks ever, only data chunks. Move this
sysfs flag to per-space_info. This won't affect current users because
this tunable only ever did anything for zoned, and that is currently
hidden behind BTRFS_CONFIG_DEBUG.
Tested-by: Pankaj Raghav <p.raghav@samsung.com> Signed-off-by: Josef Bacik <josef@toxicpanda.com>
[ jth restore global bg_reclaim_threshold ] Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The metadata overcommit makes the space reservation flexible but it is also
harmful to active zone tracking. Since we cannot finish a block group from
the metadata allocation context, we might not activate a new block group
and might not be able to actually write out the overcommit reservations.
So, disable metadata overcommit for zoned filesystems. We will ensure
the reservations are under active_total_bytes in the following patches.
CC: stable@vger.kernel.org # 5.16+ Fixes: afba2bc036b0 ("btrfs: zoned: implement active zone tracking") Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When we run out of active zones and no sufficient space is left in any
block groups, we need to finish one block group to make room to activate a
new block group.
However, we cannot do this for metadata block groups because we can cause a
deadlock by waiting for a running transaction commit. So, do that only for
a data block group.
Furthermore, the block group to be finished has two requirements. First,
the block group must not have reserved bytes left. Having reserved bytes
means we have an allocated region but did not yet send bios for it. If that
region is allocated by the thread calling btrfs_zone_finish(), it results
in a deadlock.
Second, the block group to be finished must not be a SYSTEM block
group. Finishing a SYSTEM block group easily breaks further chunk
allocation by nullifying the SYSTEM free space.
In a certain case, we cannot find any zone finish candidate or
btrfs_zone_finish() may fail. In that case, we fall back to split the
allocation bytes and fill the last spaces left in the block groups.
CC: stable@vger.kernel.org # 5.16+ Fixes: afba2bc036b0 ("btrfs: zoned: implement active zone tracking") Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
On zoned filesystem, data write out is limited by max_zone_append_size,
and a large ordered extent is split according the size of a bio. OTOH,
the number of extents to be written is calculated using
BTRFS_MAX_EXTENT_SIZE, and that estimated number is used to reserve the
metadata bytes to update and/or create the metadata items.
The metadata reservation is done at e.g, btrfs_buffered_write() and then
released according to the estimation changes. Thus, if the number of extent
increases massively, the reserved metadata can run out.
The increase of the number of extents easily occurs on zoned filesystem
if BTRFS_MAX_EXTENT_SIZE > max_zone_append_size. And, it causes the
following warning on a small RAM environment with disabling metadata
over-commit (in the following patch).
To fix the estimation, we need to introduce fs_info->max_extent_size to
replace BTRFS_MAX_EXTENT_SIZE, which allow setting the different size for
regular vs zoned filesystem.
Set fs_info->max_extent_size to BTRFS_MAX_EXTENT_SIZE by default. On zoned
filesystem, it is set to fs_info->max_zone_append_size.
CC: stable@vger.kernel.org # 5.12+ Fixes: d8e3fb106f39 ("btrfs: zoned: use ZONE_APPEND write for zoned mode") Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
This patch is basically a revert of commit 5a80d1c6a270 ("btrfs: zoned:
remove max_zone_append_size logic"), but without unnecessary ASSERT and
check. The max_zone_append_size will be used as a hint to estimate the
number of extents to cover delalloc/writeback region in the later commits.
The size of a ZONE APPEND bio is also limited by queue_max_segments(), so
this commit considers it to calculate max_zone_append_size. Technically, a
bio can be larger than queue_max_segments() * PAGE_SIZE if the pages are
contiguous. But, it is safe to consider "queue_max_segments() * PAGE_SIZE"
as an upper limit of an extent size to calculate the number of extents
needed to write data.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Add a helper to check the max supported sectors for zone append based on
the block_device instead of having to poke into the block layer internal
request_queue.
Signed-off-by: Christoph Hellwig <hch@lst.de> Acked-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com> Link: https://lore.kernel.org/r/20220415045258.199825-16-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 6f93e834fa7c seemingly inadvertently moved the code responsible
for flagging the filesystem as having BIG_METADATA to a place where
setting the flag was essentially lost. This means that
filesystems created with kernels containing this bug (starting with 5.15)
can potentially be mounted by older (pre-3.4) kernels. In reality
chances for this happening are low because there are other incompat
flags introduced in the mean time. Still the correct behavior is to set
INCOMPAT_BIG_METADATA flag and persist this in the superblock.
Fixes: 6f93e834fa7c ("btrfs: fix upper limit for max_inline for page size 64K") CC: stable@vger.kernel.org # 5.4+ Reviewed-by: Qu Wenruo <wqu@suse.com> Signed-off-by: Nikolay Borisov <nborisov@suse.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
If you try to force a chunk allocation, but you race with another chunk
allocation, you will end up waiting on the chunk allocation that just
occurred and then allocate another chunk. If you have many threads all
doing this at once you can way over-allocate chunks.
Fix this by resetting force to NO_FORCE, that way if we think we need to
allocate we can, otherwise we don't force another chunk allocation if
one is already happening.
Reviewed-by: Filipe Manana <fdmanana@suse.com> CC: stable@vger.kernel.org # 5.4+ Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When cow_file_range() fails in the middle of the allocation loop, it
unlocks the pages but leaves the ordered extents intact. Thus, we need
to call btrfs_cleanup_ordered_extents() to finish the created ordered
extents.
Also, we need to call end_extent_writepage() if locked_page is available
because btrfs_cleanup_ordered_extents() never processes the region on
the locked_page.
Furthermore, we need to set the mapping as error if locked_page is
unavailable before unlocking the pages, so that the errno is properly
propagated to the user space.
CC: stable@vger.kernel.org # 5.18+ Reviewed-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
While we debug the issue, we found running fstests generic/551 on 5GB
non-zoned null_blk device in the emulated zoned mode also had a
similar hung issue.
Also, we can reproduce the same symptom with an error injected
cow_file_range() setup.
The hang occurs when cow_file_range() fails in the middle of
allocation. cow_file_range() called from do_allocation_zoned() can
split the give region ([start, end]) for allocation depending on
current block group usages. When btrfs can allocate bytes for one part
of the split regions but fails for the other region (e.g. because of
-ENOSPC), we return the error leaving the pages in the succeeded regions
locked. Technically, this occurs only when @unlock == 0. Otherwise, we
unlock the pages in an allocated region after creating an ordered
extent.
Considering the callers of cow_file_range(unlock=0) won't write out
the pages, we can unlock the pages on error exit from
cow_file_range(). So, we can ensure all the pages except @locked_page
are unlocked on error case.
In summary, cow_file_range now behaves like this:
- page_started == 1 (return value)
- All the pages are unlocked. IO is started.
- unlock == 1
- All the pages except @locked_page are unlocked in any case
- unlock == 0
- On success, all the pages are locked for writing out them
- On failure, all the pages except @locked_page are unlocked
Fixes: 42c011000963 ("btrfs: zoned: introduce dedicated data write path for zoned filesystems") CC: stable@vger.kernel.org # 5.12+ Reviewed-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Currently we will return 1 or -EAGAIN if we decide we need to commit
the transaction rather than sync the log. In practice this doesn't
really matter, we interpret any !0 and !BTRFS_NO_LOG_SYNC as needing to
commit the transaction. However this makes it hard to figure out what
the correct thing to do is.
Fix this up by defining BTRFS_LOG_FORCE_COMMIT and using this in all the
places where we want to force the transaction to be committed.
CC: stable@vger.kernel.org # 5.15+ Reviewed-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Josef Bacik <josef@toxicpanda.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In our test of iocost, we encountered some list add/del corruptions of
inner_walk list in ioc_timer_fn.
The reason can be described as follows:
cpu 0 cpu 1
ioc_qos_write ioc_qos_write
ioc = q_to_ioc(queue);
if (!ioc) {
ioc = kzalloc();
ioc = q_to_ioc(queue);
if (!ioc) {
ioc = kzalloc();
...
rq_qos_add(q, rqos);
}
...
rq_qos_add(q, rqos);
...
}
When the io.cost.qos file is written by two cpus concurrently, rq_qos may
be added to one disk twice. In that case, there will be two iocs enabled
and running on one disk. They own different iocgs on their active list. In
the ioc_timer_fn function, because of the iocgs from two iocs have the
same root iocg, the root iocg's walk_list may be overwritten by each other
and this leads to list add/del corruptions in building or destroying the
inner_walk list.
And so far, the blk-rq-qos framework works in case that one instance for
one type rq_qos per queue by default. This patch make this explicit and
also fix the crash above.
Various places like I/O schedulers or the QOS infrastructure try to
register debugfs files on demans, which can race with creating and
removing the main queue debugfs directory. Use the existing
debugfs_mutex to serialize all debugfs operations that rely on
q->debugfs_dir or the directories hanging off it.
To make the teardown code a little simpler declare all debugfs dentry
pointers and not just the main one uncoditionally in blkdev.h.
Move debugfs_mutex next to the dentries that it protects and document
what it is used for.
The csdlock_debug kernel-boot parameter is parsed by the
early_param() function csdlock_debug(). If set, csdlock_debug()
invokes static_branch_enable() to enable csd_lock_wait feature, which
triggers a panic on arm64 for kernels built with CONFIG_SPARSEMEM=y and
CONFIG_SPARSEMEM_VMEMMAP=n.
With CONFIG_SPARSEMEM_VMEMMAP=n, __nr_to_section is called in
static_key_enable() and returns NULL, resulting in a NULL dereference
because mem_section is initialized only later in sparse_init().
This is also a problem for powerpc because early_param() functions
are invoked earlier than jump_label_init(), also resulting in
static_key_enable() failures. These failures cause the warning "static
key 'xxx' used before call to jump_label_init()".
Thus, early_param is too early for csd_lock_wait to run
static_branch_enable(), so changes it to __setup to fix these.
Fixes: 8d0968cc6b8f ("locking/csd_lock: Add boot parameter for controlling CSD lock debugging") Cc: stable@vger.kernel.org Reported-by: Chen jingwen <chenjingwen6@huawei.com> Signed-off-by: Chen Zhongjin <chenzhongjin@huawei.com> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The rng's random_init() function contributes the real time to the rng at
boot time, so that events can at least start in relation to something
particular in the real world. But this clock might not yet be set that
point in boot, so nothing is contributed. In addition, the relation
between minor clock changes from, say, NTP, and the cycle counter is
potentially useful entropic data.
This commit addresses this by mixing in a time stamp on calls to
settimeofday and adjtimex. No entropy is credited in doing so, so it
doesn't make initialization faster, but it is still useful input to
have.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable@vger.kernel.org Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Ensure that the fid's iounit field is set to zero when a new fid is
created. Certain 9P operations, such as OPEN and CREATE, allow the
server to reply with an iounit size which the client code assigns to the
p9_fid struct shortly after the fid is created by p9_fid_create(). On
the other hand, an XATTRWALK operation doesn't allow for the server to
specify an iounit value. The iounit field of the newly allocated p9_fid
struct remained uninitialized in that case. Depending on allocation
patterns, the iounit value could have been something reasonable that was
carried over from previously freed fids or, in the worst case, could
have been arbitrary values from non-fid related usages of the memory
location.
The bug was detected in the Windows Subsystem for Linux 2 (WSL2) kernel
after the uninitialized iounit field resulted in the typical sequence of
two getxattr(2) syscalls, one to get the size of an xattr and another
after allocating a sufficiently sized buffer to fit the xattr value, to
hit an unexpected ERANGE error in the second call to getxattr(2). An
uninitialized iounit field would sometimes force rsize to be smaller
than the xattr value size in p9_client_read_once() and the 9P server in
WSL refused to chunk up the READ on the attr_fid and, instead, returned
ERANGE to the client. The virtfs server in QEMU seems happy to chunk up
the READ and this problem goes undetected there.
Fault inject on pool metadata device reports:
BUG: KASAN: use-after-free in dm_pool_register_metadata_threshold+0x40/0x80
Read of size 8 at addr ffff8881b9d50068 by task dmsetup/950
CPU: 7 PID: 950 Comm: dmsetup Tainted: G W 5.19.0-rc6 #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-1.fc33 04/01/2014
Call Trace:
<TASK>
dump_stack_lvl+0x34/0x44
print_address_description.constprop.0.cold+0xeb/0x3f4
kasan_report.cold+0xe6/0x147
dm_pool_register_metadata_threshold+0x40/0x80
pool_ctr+0xa0a/0x1150
dm_table_add_target+0x2c8/0x640
table_load+0x1fd/0x430
ctl_ioctl+0x2c4/0x5a0
dm_ctl_ioctl+0xa/0x10
__x64_sys_ioctl+0xb3/0xd0
do_syscall_64+0x35/0x80
entry_SYSCALL_64_after_hwframe+0x46/0xb0
This can be easily reproduced using:
echo offline > /sys/block/sda/device/state
dd if=/dev/zero of=/dev/mapper/thin bs=4k count=10
dmsetup load pool --table "0 20971520 thin-pool /dev/sda /dev/sdb 128 0 0"
If a metadata commit fails, the transaction will be aborted and the
metadata space maps will be destroyed. If a DM table reload then
happens for this failed thin-pool, a use-after-free will occur in
dm_sm_register_threshold_callback (called from
dm_pool_register_metadata_threshold).
Fix this by in dm_pool_register_metadata_threshold() by returning the
-EINVAL error if the thin-pool is in fail mode. Also fail pool_ctr()
with a new error message: "Error registering metadata threshold".
Fixes: ac8c3f3df65e4 ("dm thin: generate event when metadata threshold passed") Cc: stable@vger.kernel.org Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Luo Meng <luomeng12@huawei.com> Signed-off-by: Mike Snitzer <snitzer@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
To load a string created by variable array va_list.
The main issue with this approach is that "MSG_MAX" usage in the
__dynamic_array() portion. That actually just reserves the MSG_MAX in the
event, and even wastes space because there's dynamic meta data also saved
in the event to denote the offset and size of the dynamic array. It would
have been better to just use a static __array() field.
Instead, create __vstring() and __assign_vstr() that work like __string
and __assign_str() but instead of taking a destination string to copy,
take a format string and a va_list pointer and fill in the values.
To figure out the length to store the string. It may be slightly slower as
it needs to run the vsnprintf() twice, but it now saves space on the ring
buffer.
Link: https://lkml.kernel.org/r/20220705224749.053570613@goodmis.org Cc: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Jason Gunthorpe <jgg@ziepe.ca> Cc: Leon Romanovsky <leon@kernel.org> Cc: Kalle Valo <kvalo@kernel.org> Cc: "David S. Miller" <davem@davemloft.net> Cc: Eric Dumazet <edumazet@google.com> Cc: Jakub Kicinski <kuba@kernel.org> Cc: Paolo Abeni <pabeni@redhat.com> Cc: Arend van Spriel <aspriel@gmail.com> Cc: Franky Lin <franky.lin@broadcom.com> Cc: Hante Meuleman <hante.meuleman@broadcom.com> Cc: Gregory Greenman <gregory.greenman@intel.com> Cc: Peter Chen <peter.chen@kernel.org> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Mathias Nyman <mathias.nyman@intel.com> Cc: Chunfeng Yun <chunfeng.yun@mediatek.com> Cc: Bin Liu <b-liu@ti.com> Cc: Marek Lindner <mareklindner@neomailbox.ch> Cc: Simon Wunderlich <sw@simonwunderlich.de> Cc: Antonio Quartulli <a@unstable.cc> Cc: Sven Eckelmann <sven@narfation.org> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Jim Cromie <jim.cromie@gmail.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
commit e23a8020ce4e ("s390/kexec_file: Signature verification prototype")
adds support for KEXEC_SIG verification with keys from platform keyring
but the built-in keys and secondary keyring are not used.
Add support for the built-in keys and secondary keyring as x86 does.
Before commit 105e10e2cf1c ("kexec_file: drop weak attribute from
functions"), there was already no arch-specific implementation
of arch_kexec_kernel_verify_sig. With weak attribute dropped by that
commit, arch_kexec_kernel_verify_sig is completely useless. So clean it
up.
Note later patches are dependent on this patch so it should be backported
to the stable tree as well.
Cc: stable@vger.kernel.org Suggested-by: Eric W. Biederman <ebiederm@xmission.com> Reviewed-by: Michal Suchanek <msuchanek@suse.de> Acked-by: Baoquan He <bhe@redhat.com> Signed-off-by: Coiby Xu <coxu@redhat.com>
[zohar@linux.ibm.com: reworded patch description "Note"] Link: https://lore.kernel.org/linux-integrity/20220714134027.394370-1-coxu@redhat.com/ Signed-off-by: Mimi Zohar <zohar@linux.ibm.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
As requested
(http://lkml.kernel.org/r/87ee0q7b92.fsf@email.froward.int.ebiederm.org),
this series converts weak functions in kexec to use the #ifdef approach.
Quoting the 3e35142ef99fe ("kexec_file: drop weak attribute from
arch_kexec_apply_relocations[_add]") changelog:
: Since commit d1bcae833b32f1 ("ELF: Don't generate unused section symbols")
: [1], binutils (v2.36+) started dropping section symbols that it thought
: were unused. This isn't an issue in general, but with kexec_file.c, gcc
: is placing kexec_arch_apply_relocations[_add] into a separate
: .text.unlikely section and the section symbol ".text.unlikely" is being
: dropped. Due to this, recordmcount is unable to find a non-weak symbol in
: .text.unlikely to generate a relocation record against.
This patch (of 2);
Drop __weak attribute from functions in kexec_file.c:
- arch_kexec_kernel_image_probe()
- arch_kimage_file_post_load_cleanup()
- arch_kexec_kernel_image_load()
- arch_kexec_locate_mem_hole()
- arch_kexec_kernel_verify_sig()
arch_kexec_kernel_image_load() calls into kexec_image_load_default(), so
drop the static attribute for the latter.
arch_kexec_kernel_verify_sig() is not overridden by any architecture, so
drop the __weak attribute.
dm-writecache has the capability to limit the number of writeback jobs
in progress. However, this feature was off by default. As such there
were some out-of-memory crashes observed when lowering the low
watermark while the cache is full.
This commit enables writeback limit by default. It is set to 256MiB or
1/16 of total system memory, whichever is smaller.
Add support for some of the Brainboxes PCIe (PX) range of
serial cards, including the PX-101, PX-235/PX-246,
PX-203/PX-257, PX-260/PX-701, PX-310, PX-313,
PX-320/PX-324/PX-376/PX-387, PX-335/PX-346, PX-368, PX-420,
PX-803 and PX-846.
Oxford Semiconductor PCIe (Tornado) 950 serial port devices are driven
by a fixed 62.5MHz clock input derived from the 100MHz PCI Express clock.
We currently drive the device using its default oversampling rate of 16
and the clock prescaler disabled, consequently yielding the baud base of 3906250. This base is inadequate for some of the high-speed baud rates
such as 460800bps, for which the closest rate possible can be obtained
by dividing the baud base by 8, yielding the baud rate of 488281.25bps,
which is off by 5.9638%. This is enough for data communication to break
with the remote end talking actual 460800bps, where missed stop bits
have been observed.
We can do better however, by taking advantage of a reduced oversampling
rate, which can be set to any integer value from 4 to 16 inclusive by
programming the TCR register, and by using the clock prescaler, which
can be set to any value from 1 to 63.875 in increments of 0.125 in the
CPR/CPR2 register pair. The prescaler has to be explicitly enabled
though by setting bit 7 in the MCR or otherwise it is bypassed (in the
enhanced mode that we enable) as if the value of 1 was used.
Make use of these features then as follows:
- Set the baud base to 15625000, reflecting the minimum oversampling
rate of 4 with the clock prescaler and divisor both set to 1.
- Override the `set_mctrl' and set the MCR shadow there so as to have
MCR[7] always set and have the 8250 core propagate these settings.
- Override the `get_divisor' handler and determine a good combination of
parameters by using a lookup table with predetermined value pairs of
the oversampling rate and the clock prescaler and finding a pair that
divides the input clock such that the quotient, when rounded to the
nearest integer, deviates the least from the exact result. Calculate
the clock divisor accordingly.
Scale the resulting oversampling rate (only by powers of two) if
possible so as to maximise it, reducing the divisor accordingly, and
avoid a divisor overflow for very low baud rates by scaling the
oversampling rate and/or the prescaler even if that causes some
accuracy loss.
Also handle the historic spd_cust feature so as to allow one to set
all the three parameters manually to arbitrary values, by keeping the
low 16 bits for the divisor and then putting TCR in bits 19:16 and
CPR/CPR2 in bits 28:20, sanitising the bit pattern supplied such as
to clamp CPR/CPR2 values between 0.000 and 0.875 inclusive to 33.875.
This preserves compatibility with any existing setups, that is where
requesting a custom divisor that only has any bits set among the low
16 the oversampling rate of 16 and the clock prescaler of 33.875 will
be used as with the original 8250.
Finally abuse the `frac' argument to store the determined bit patterns
for the TCR, CPR and CPR2 registers.
- Override the `set_divisor' handler so as to set the TCR, CPR and CPR2
registers from the `frac' value supplied. Set the divisor as usual.
With the baud base set to 15625000 and the unsigned 16-bit UART_DIV_MAX
limitation imposed by `serial8250_get_baud_rate' standard baud rates
below 300bps become unavailable in the regular way, e.g. the rate of
200bps requires the baud base to be divided by 78125 and that is beyond
the unsigned 16-bit range. The historic spd_cust feature can still be
used to obtain such rates if so required.
See Documentation/tty/device_drivers/oxsemi-tornado.rst for more details.
The EndRun PTP/1588 dual serial port device is based on the Oxford
Semiconductor OXPCIe952 UART device with the PCI vendor:device ID set
for EndRun Technologies and uses the same sequence to determine the
number of ports available. Despite that we have duplicate code
specific to the EndRun device.
Remove redundant code then and factor out OxSemi Tornado device
detection.