Vladimir Oltean [Tue, 16 Aug 2022 13:53:52 +0000 (16:53 +0300)]
net: mscc: ocelot: report ndo_get_stats64 from the wraparound-resistant ocelot->stats
Rather than reading the stats64 counters directly from the 32-bit
hardware, it's better to rely on the output produced by the periodic
ocelot_port_update_stats().
It would be even better to call ocelot_port_update_stats() right from
ocelot_get_stats64() to make sure we report the current values rather
than the ones from 2 seconds ago. But we need to export
ocelot_port_update_stats() from the switch lib towards the switchdev
driver for that, and future work will largely undo that.
There are more ocelot-based drivers waiting to be introduced, an example
of which is the SPI-controlled VSC7512. In that driver's case, it will
be impossible to call ocelot_port_update_stats() from ndo_get_stats64
context, since the latter is atomic, and reading the stats over SPI is
sleepable. So the compromise taken here, which will also hold going
forward, is to report 64-bit counters to stats64, which are not 100% up
to date.
Fixes: a556c76adc05 ("net: mscc: Add initial Ocelot switch support") Signed-off-by: Vladimir Oltean <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
Vladimir Oltean [Tue, 16 Aug 2022 13:53:51 +0000 (16:53 +0300)]
net: mscc: ocelot: keep ocelot_stat_layout by reg address, not offset
With so many counter addresses recently discovered as being wrong, it is
desirable to at least have a central database of information, rather
than two: one through the SYS_COUNT_* registers (used for
ndo_get_stats64), and the other through the offset field of struct
ocelot_stat_layout elements (used for ethtool -S).
The strategy will be to keep the SYS_COUNT_* definitions as the single
source of truth, but for that we need to expand our current definitions
to cover all registers. Then we need to convert the ocelot region
creation logic, and stats worker, to the read semantics imposed by going
through SYS_COUNT_* absolute register addresses, rather than offsets
of 32-bit words relative to SYS_COUNT_RX_OCTETS (which should have been
SYS_CNT, by the way).
Vladimir Oltean [Tue, 16 Aug 2022 13:53:50 +0000 (16:53 +0300)]
net: mscc: ocelot: make struct ocelot_stat_layout array indexable
The ocelot counters are 32-bit and require periodic reading, every 2
seconds, by ocelot_port_update_stats(), so that wraparounds are
detected.
Currently, the counters reported by ocelot_get_stats64() come from the
32-bit hardware counters directly, rather than from the 64-bit
accumulated ocelot->stats, and this is a problem for their integrity.
The strategy is to make ocelot_get_stats64() able to cherry-pick
individual stats from ocelot->stats the way in which it currently reads
them out from SYS_COUNT_* registers. But currently it can't, because
ocelot->stats is an opaque u64 array that's used only to feed data into
ethtool -S.
To solve that problem, we need to make ocelot->stats indexable, and
associate each element with an element of struct ocelot_stat_layout used
by ethtool -S.
This makes ocelot_stat_layout a fat (and possibly sparse) array, so we
need to change the way in which we access it. We no longer need
OCELOT_STAT_END as a sentinel, because we know the array's size
(OCELOT_NUM_STATS). We just need to skip the array elements that were
left unpopulated for the switch revision (ocelot, felix, seville).
Vladimir Oltean [Tue, 16 Aug 2022 13:53:49 +0000 (16:53 +0300)]
net: mscc: ocelot: fix race between ndo_get_stats64 and ocelot_check_stats_work
The 2 methods can run concurrently, and one will change the window of
counters (SYS_STAT_CFG_STAT_VIEW) that the other sees. The fix is
similar to what commit 7fbf6795d127 ("net: mscc: ocelot: fix mutex lock
error during ethtool stats read") has done for ethtool -S.
Fixes: a556c76adc05 ("net: mscc: Add initial Ocelot switch support") Signed-off-by: Vladimir Oltean <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
Vladimir Oltean [Tue, 16 Aug 2022 13:53:48 +0000 (16:53 +0300)]
net: mscc: ocelot: turn stats_lock into a spinlock
ocelot_get_stats64() currently runs unlocked and therefore may collide
with ocelot_port_update_stats() which indirectly accesses the same
counters. However, ocelot_get_stats64() runs in atomic context, and we
cannot simply take the sleepable ocelot->stats_lock mutex. We need to
convert it to an atomic spinlock first. Do that as a preparatory change.
Vladimir Oltean [Tue, 16 Aug 2022 13:53:47 +0000 (16:53 +0300)]
net: mscc: ocelot: fix address of SYS_COUNT_TX_AGING counter
This register, used as part of stats->tx_dropped in
ocelot_get_stats64(), has a wrong address. At the address currently
given, there is actually the c_tx_green_prio_6 counter.
Fixes: a556c76adc05 ("net: mscc: Add initial Ocelot switch support") Signed-off-by: Vladimir Oltean <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
Reading stats using the SYS_COUNT_* register definitions is only used by
ocelot_get_stats64() from the ocelot switchdev driver, however,
currently the bucket definitions are incorrect.
Separately, on both RX and TX, we have the following problems:
- a 256-1023 bucket which actually tracks the 256-511 packets
- the 1024-1526 bucket actually tracks the 512-1023 packets
- the 1527-max bucket actually tracks the 1024-1526 packets
=> nobody tracks the packets from the real 1527-max bucket
Additionally, the RX_PAUSE, RX_CONTROL, RX_LONGS and RX_CLASSIFIED_DROPS
all track the wrong thing. However this doesn't seem to have any
consequence, since ocelot_get_stats64() doesn't use these.
Even though this problem only manifests itself for the switchdev driver,
we cannot split the fix for ocelot and for DSA, since it requires fixing
the bucket definitions from enum ocelot_reg, which makes us necessarily
adapt the structures from felix and seville as well.
Fixes: 84705fc16552 ("net: dsa: felix: introduce support for Seville VSC9953 switch") Fixes: 56051948773e ("net: dsa: ocelot: add driver for Felix switch family") Fixes: a556c76adc05 ("net: mscc: Add initial Ocelot switch support") Signed-off-by: Vladimir Oltean <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
Vladimir Oltean [Tue, 16 Aug 2022 20:14:45 +0000 (23:14 +0300)]
net: dsa: don't warn in dsa_port_set_state_now() when driver doesn't support it
ds->ops->port_stp_state_set() is, like most DSA methods, optional, and
if absent, the port is supposed to remain in the forwarding state (as
standalone). Such is the case with the mv88e6060 driver, which does not
offload the bridge layer. DSA warns that the STP state can't be changed
to FORWARDING as part of dsa_port_enable_rt(), when in fact it should not.
The error message is also not up to modern standards, so take the
opportunity to make it more descriptive.
====================
netfilter: conntrack and nf_tables bug fixes
The following patchset contains netfilter fixes for net.
Broken since 5.19:
A few ancient connection tracking helpers assume TCP packets cannot
exceed 64kb in size, but this isn't the case anymore with 5.19 when
BIG TCP got merged, from myself.
Regressions since 5.19:
1. 'conntrack -E expect' won't display anything because nfnetlink failed
to enable events for expectations, only for normal conntrack events.
2. partially revert change that added resched calls to a function that can
be in atomic context. Both broken and fixed up by myself.
Broken for several releases (up to original merge of nf_tables):
Several fixes for nf_tables control plane, from Pablo.
This fixes up resource leaks in error paths and adds more sanity
checks for mutually exclusive attributes/flags.
Kconfig:
NF_CONNTRACK_PROCFS is very old and doesn't provide all info provided
via ctnetlink, so it should not default to y. From Geert Uytterhoeven.
Selftests:
rework nft_flowtable.sh: it frequently indicated failure; the way it
tried to detect an offload failure did not work reliably.
* git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
testing: selftests: nft_flowtable.sh: rework test to detect offload failure
testing: selftests: nft_flowtable.sh: use random netns names
netfilter: conntrack: NF_CONNTRACK_PROCFS should no longer default to y
netfilter: nf_tables: check NFT_SET_CONCAT flag if field_count is specified
netfilter: nf_tables: disallow NFT_SET_ELEM_CATCHALL and NFT_SET_ELEM_INTERVAL_END
netfilter: nf_tables: NFTA_SET_ELEM_KEY_END requires concat and interval flags
netfilter: nf_tables: validate NFTA_SET_ELEM_OBJREF based on NFT_SET_OBJECT flag
netfilter: nf_tables: really skip inactive sets when allocating name
netfilter: nfnetlink: re-enable conntrack expectation events
netfilter: nf_tables: fix scheduling-while-atomic splat
netfilter: nf_ct_irc: cap packet search space to 4k
netfilter: nf_ct_ftp: prefer skb_linearize
netfilter: nf_ct_h323: cap packet size at 64k
netfilter: nf_ct_sane: remove pseudo skb linearization
netfilter: nf_tables: possible module reference underflow in error path
netfilter: nf_tables: disallow NFTA_SET_ELEM_KEY_END with NFT_SET_ELEM_INTERVAL_END flag
netfilter: nf_tables: use READ_ONCE and WRITE_ONCE for shared generation id access
====================
David Howells [Tue, 16 Aug 2022 09:34:40 +0000 (10:34 +0100)]
net: Fix suspicious RCU usage in bpf_sk_reuseport_detach()
bpf_sk_reuseport_detach() calls __rcu_dereference_sk_user_data_with_flags()
to obtain the value of sk->sk_user_data, but that function is only usable
if the RCU read lock is held, and neither that function nor any of its
callers hold it.
Fix this by adding a new helper, __locked_read_sk_user_data_with_flags()
that checks to see if sk->sk_callback_lock() is held and use that here
instead.
Alternatively, making __rcu_dereference_sk_user_data_with_flags() use
rcu_dereference_checked() might suffice.
Without this, the following warning can be occasionally observed:
Dave Airlie [Wed, 17 Aug 2022 22:51:04 +0000 (08:51 +1000)]
Merge tag 'drm-misc-fixes-2022-08-16' of git://anongit.freedesktop.org/drm/drm-misc into drm-fixes
One patch for imx/dcss to get rid of a warning message, one off-by-one
fix and GA103 support for nouveau, a refcounting fix for meson, a NULL
pointer dereference fix for ttm, a error check fix for lvds-codec, a
dt-binding schema fix and an underflow fix for sun4i
* tag 'ntfs3_for_6.0' of https://github.com/Paragon-Software-Group/linux-ntfs3: (39 commits)
fs/ntfs3: uninitialized variable in ntfs_set_acl_ex()
fs/ntfs3: Remove unused function wnd_bits
fs/ntfs3: Make ni_ins_new_attr return error
fs/ntfs3: Create MFT zone only if length is large enough
fs/ntfs3: Refactoring attr_insert_range to restore after errors
fs/ntfs3: Refactoring attr_punch_hole to restore after errors
fs/ntfs3: Refactoring attr_set_size to restore after errors
fs/ntfs3: New function ntfs_bad_inode
fs/ntfs3: Make MFT zone less fragmented
fs/ntfs3: Check possible errors in run_pack in advance
fs/ntfs3: Added comments to frecord functions
fs/ntfs3: Fill duplicate info in ni_add_name
fs/ntfs3: Make static function attr_load_runs
fs/ntfs3: Add new argument is_mft to ntfs_mark_rec_free
fs/ntfs3: Remove unused mi_mark_free
fs/ntfs3: Fix very fragmented case in attr_punch_hole
fs/ntfs3: Fix work with fragmented xattr
fs/ntfs3: Make ntfs_fallocate return -ENOSPC instead of -EFBIG
fs/ntfs3: extend ni_insert_nonresident to return inserted ATTR_LIST_ENTRY
fs/ntfs3: Check reserved size for maximum allowed
...
Linus Torvalds [Thu, 11 Aug 2022 22:29:46 +0000 (15:29 -0700)]
dcache: move the DCACHE_OP_COMPARE case out of the __d_lookup_rcu loop
__d_lookup_rcu() is one of the hottest functions in the kernel on
certain loads, and it is complicated by filesystems that might want to
have their own name compare function.
We can improve code generation by moving the test of DCACHE_OP_COMPARE
outside the loop, which makes the loop itself much simpler, at the cost
of some code duplication. But both cases end up being simpler, and the
"native" direct case-sensitive compare particularly so.
Arun Ramadoss [Tue, 16 Aug 2022 10:55:16 +0000 (16:25 +0530)]
net: dsa: microchip: ksz9477: fix fdb_dump last invalid entry
In the ksz9477_fdb_dump function it reads the ALU control register and
exit from the timeout loop if there is valid entry or search is
complete. After exiting the loop, it reads the alu entry and report to
the user space irrespective of entry is valid. It works till the valid
entry. If the loop exited when search is complete, it reads the alu
table. The table returns all ones and it is reported to user space. So
bridge fdb show gives ff:ff:ff:ff:ff:ff as last entry for every port.
To fix it, after exiting the loop the entry is reported only if it is
valid one.
ice: Fix VF not able to send tagged traffic with no VLAN filters
VF was not able to send tagged traffic when it didn't
have any VLAN interfaces and VLAN anti-spoofing was enabled.
Fix this by allowing VFs with no VLAN filters to send tagged
traffic. After VF adds a VLAN interface it will be able to
send tagged traffic matching VLAN filters only.
Testing hints:
1. Spawn VF
2. Send tagged packet from a VF
3. The packet should be sent out and not dropped
4. Add a VLAN interface on VF
5. Send tagged packet on that VLAN interface
6. Packet should be sent out and not dropped
7. Send tagged packet with id different than VLAN interface
8. Packet should be dropped
ice: Ignore error message when setting same promiscuous mode
Commit 1273f89578f2 ("ice: Fix broken IFF_ALLMULTI handling")
introduced new checks when setting/clearing promiscuous mode. But if the
requested promiscuous mode setting already exists, an -EEXIST error
message would be printed. This is incorrect because promiscuous mode is
either on/off and shouldn't print an error when the requested
configuration is already set.
This can happen when removing a bridge with two bonded interfaces and
promiscuous most isn't fully cleared from VLAN VSI in hardware.
Fix this by ignoring cases where requested promiscuous mode exists.
Grzegorz Siwik [Fri, 12 Aug 2022 13:25:49 +0000 (15:25 +0200)]
ice: Fix clearing of promisc mode with bridge over bond
When at least two interfaces are bonded and a bridge is enabled on the
bond, an error can occur when the bridge is removed and re-added. The
reason for the error is because promiscuous mode was not fully cleared from
the VLAN VSI in the hardware. With this change, promiscuous mode is
properly removed when the bridge disconnects from bonding.
[ 1033.676359] bond1: link status definitely down for interface enp95s0f0, disabling it
[ 1033.676366] bond1: making interface enp175s0f0 the new active one
[ 1033.676369] device enp95s0f0 left promiscuous mode
[ 1033.676522] device enp175s0f0 entered promiscuous mode
[ 1033.676901] ice 0000:af:00.0 enp175s0f0: Error setting Multicast promiscuous mode on VSI 6
[ 1041.795662] ice 0000:af:00.0 enp175s0f0: Error setting Multicast promiscuous mode on VSI 6
[ 1041.944826] bond1: link status definitely down for interface enp175s0f0, disabling it
[ 1041.944874] device enp175s0f0 left promiscuous mode
[ 1041.944918] bond1: now running without any active interface!
Grzegorz Siwik [Fri, 12 Aug 2022 13:25:48 +0000 (15:25 +0200)]
ice: Ignore EEXIST when setting promisc mode
Ignore EEXIST error when setting promiscuous mode.
This fix is needed because the driver could set promiscuous mode
when it still has not cleared properly.
Promiscuous mode could be set only once, so setting it second
time will be rejected.
Grzegorz Siwik [Fri, 12 Aug 2022 13:25:47 +0000 (15:25 +0200)]
ice: Fix double VLAN error when entering promisc mode
Avoid enabling or disabling VLAN 0 when trying to set promiscuous
VLAN mode if double VLAN mode is enabled. This fix is needed
because the driver tries to add the VLAN 0 filter twice (once for
inner and once for outer) when double VLAN mode is enabled. The
filter program is rejected by the firmware when double VLAN is
enabled, because the promiscuous filter only needs to be set once.
This issue was missed in the initial implementation of double VLAN
mode.
Linus Torvalds [Wed, 17 Aug 2022 15:58:54 +0000 (08:58 -0700)]
Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Pull virtio fixes from Michael Tsirkin:
"Most notably this drops the commits that trip up google cloud (turns
out, any legacy device).
Plus a kerneldoc patch"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
virtio: kerneldocs fixes and enhancements
virtio: Revert "virtio: find_vqs() add arg sizes"
virtio_vdpa: Revert "virtio_vdpa: support the arg sizes of find_vqs()"
virtio_pci: Revert "virtio_pci: support the arg sizes of find_vqs()"
virtio-mmio: Revert "virtio_mmio: support the arg sizes of find_vqs()"
virtio: Revert "virtio: add helper virtio_find_vqs_ctx_size()"
virtio_net: Revert "virtio_net: set the default max ring size by find_vqs()"
Josef Bacik [Wed, 3 Aug 2022 18:28:47 +0000 (14:28 -0400)]
btrfs: tree-checker: check for overlapping extent items
We're seeing a weird problem in production where we have overlapping
extent items in the extent tree. It's unclear where these are coming
from, and in debugging we realized there's no check in the tree checker
for this sort of problem. Add a check to the tree-checker to make sure
that the extents do not overlap each other.
Filipe Manana [Mon, 1 Aug 2022 13:57:52 +0000 (14:57 +0100)]
btrfs: fix warning during log replay when bumping inode link count
During log replay, at add_link(), we may increment the link count of
another inode that has a reference that conflicts with a new reference
for the inode currently being processed.
During log replay, at add_link(), we may drop (unlink) a reference from
some inode in the subvolume tree if that reference conflicts with a new
reference found in the log for the inode we are currently processing.
After the unlink, If the link count has decreased from 1 to 0, then we
increment the link count to prevent the inode from being deleted if it's
evicted by an iput() call, because we may have references to add to that
inode later on (and we will fixup its link count later during log replay).
However incrementing the link count from 0 to 1 triggers a warning:
The I_LINKABLE flag is only set when creating an O_TMPFILE file, so it's
never set during log replay.
Most of the time, the warning isn't triggered even if we dropped the last
reference of the conflicting inode, and this is because:
1) The conflicting inode was previously marked for fixup, through a call
to link_to_fixup_dir(), which increments the inode's link count;
2) And the last iput() on the inode has not triggered eviction of the
inode, nor was eviction triggered after the iput(). So at add_link(),
even if we unlink the last reference of the inode, its link count ends
up being 1 and not 0.
So this means that if eviction is triggered after link_to_fixup_dir() is
called, at add_link() we will read the inode back from the subvolume tree
and have it with a correct link count, matching the number of references
it has on the subvolume tree. So if when we are at add_link() the inode
has exactly one reference only, its link count is 1, and after the unlink
its link count becomes 0.
So fix this by using set_nlink() instead of inc_nlink(), as the former
accepts a transition from 0 to 1 and it's what we use in other similar
contexts (like at link_to_fixup_dir().
Also make add_inode_ref() use set_nlink() instead of inc_nlink() to
bump the link count from 0 to 1.
The warning is actually harmless, but it may scare users. Josef also ran
into it recently.
Filipe Manana [Mon, 1 Aug 2022 13:57:51 +0000 (14:57 +0100)]
btrfs: fix lost error handling when looking up extended ref on log replay
During log replay, when processing inode references, if we get an error
when looking up for an extended reference at __add_inode_ref(), we ignore
it and proceed, returning success (0) if no other error happens after the
lookup. This is obviously wrong because in case an extended reference
exists and it encodes some name not in the log, we need to unlink it,
otherwise the filesystem state will not match the state it had after the
last fsync.
So just make __add_inode_ref() return an error it gets from the extended
reference lookup.
This isn't necessarily new, it's just tricky to hit in practice. There
are two competing things going on here. With relocation we create a
snapshot of every fs tree with a reloc tree. Any extent buffers that
get initialized here are initialized with the reloc root lockdep key.
However since it is a snapshot, any blocks that are currently in cache
that originally belonged to the fs tree will have the normal tree
lockdep key set. This creates the lock dependency of
reloc tree -> normal tree
for the extent buffer locking during the first phase of the relocation
as we walk down the reloc root to relocate blocks.
However this is problematic because the final phase of the relocation is
merging the reloc root into the original fs root. This involves
searching down to any keys that exist in the original fs root and then
swapping the relocated block and the original fs root block. We have to
search down to the fs root first, and then go search the reloc root for
the block we need to replace. This creates the dependency of
normal tree -> reloc tree
which is why lockdep complains.
Additionally even if we were to fix this particular mismatch with a
different nesting for the merge case, we're still slotting in a block
that has a owner of the reloc root objectid into a normal tree, so that
block will have its lockdep key set to the tree reloc root, and create a
lockdep splat later on when we wander into that block from the fs root.
Unfortunately the only solution here is to make sure we do not set the
lockdep key to the reloc tree lockdep key normally, and then reset any
blocks we wander into from the reloc root when we're doing the merged.
This solves the problem of having mixed tree reloc keys intermixed with
normal tree keys, and then allows us to make sure in the merge case we
maintain the lock order of
normal tree -> reloc tree
We handle this by setting a bit on the reloc root when we do the search
for the block we want to relocate, and any block we search into or COW
at that point gets set to the reloc tree key. This works correctly
because we only ever COW down to the parent node, so we aren't resetting
the key for the block we're linking into the fs root.
With this patch we no longer have the lockdep splat in btrfs/187.
which may fail for example (or other errors could happen). When the
failure occurs, btrfs_relocate_block_group() detects the error and frees
rc and doesn't set fs_info->reloc_ctl to NULL. After that, in
btrfs_init_reloc_root(), rc is retrieved from fs_info->reloc_ctl and
then used, which may cause a use-after-free bug.
This possible bug can be triggered by calling btrfs_ioctl_balance()
before calling btrfs_ioctl_defrag().
To fix this possible bug, in prepare_to_relocate(), check if
btrfs_commit_transaction() fails. If the failure occurs,
unset_reloc_control() is called to set fs_info->reloc_ctl to NULL.
The error log in our fault-injection testing is shown as follows:
ALSA: info: Fix llseek return value when using callback
When using callback there was a flow of
ret = -EINVAL
if (callback) {
offset = callback();
goto out;
}
...
offset = some other value in case of no callback;
ret = offset;
out:
return ret;
which causes the snd_info_entry_llseek() to return -EINVAL when there is
callback handler. Fix this by setting "ret" directly to callback return
value before jumping to "out".
Florian Westphal [Tue, 16 Aug 2022 12:15:22 +0000 (14:15 +0200)]
testing: selftests: nft_flowtable.sh: rework test to detect offload failure
This test fails on current kernel releases because the flotwable path
now calls dst_check from packet path and will then remove the offload.
Test script has two purposes:
1. check that file (random content) can be sent to other netns (and vv)
2. check that the flow is offloaded (rather than handled by classic
forwarding path).
Since dst_check is in place, 2) fails because the nftables ruleset in
router namespace 1 intentionally blocks traffic under the assumption
that packets are not passed via classic path at all.
Rework this: Instead of blocking traffic, create two named counters, one
for original and one for reverse direction.
The first three test cases are handled by classic forwarding path
(path mtu discovery is disabled and packets exceed MTU).
But all other tests enable PMTUD, so the originator and responder are
expected to lower packet size and flowtable is expected to do the packet
forwarding.
For those tests, check that the packet counters (which are only
incremented for packets that are passed up to classic forward path)
are significantly lower than the file size transferred.
I've tested that the counter-checks fail as expected when the 'flow add'
statement is removed from the ruleset.
Oliver Upton [Tue, 16 Aug 2022 19:25:54 +0000 (19:25 +0000)]
KVM: arm64: Reject 32bit user PSTATE on asymmetric systems
KVM does not support AArch32 EL0 on asymmetric systems. To that end,
prevent userspace from configuring a vCPU in such a state through
setting PSTATE.
It is already ABI that KVM rejects such a write on a system where
AArch32 EL0 is unsupported. Though the kernel's definition of a 32bit
system changed in commit 2122a833316f ("arm64: Allow mismatched
32-bit EL0 support"), KVM's did not.
Jakub Kicinski [Tue, 16 Aug 2022 00:23:58 +0000 (17:23 -0700)]
tls: rx: react to strparser initialization errors
Even though the normal strparser's init function has a return
value we got away with ignoring errors until now, as it only
validates the parameters and we were passing correct parameters.
tls_strp can fail to init on memory allocation errors, which
syzbot duly induced and reported.
Florian Westphal [Tue, 16 Aug 2022 12:15:21 +0000 (14:15 +0200)]
testing: selftests: nft_flowtable.sh: use random netns names
"ns1" is a too generic name, use a random suffix to avoid
errors when such a netns exists. Also allows to run multiple
instances of the script in parallel.
Andrew Jones [Tue, 16 Aug 2022 16:30:58 +0000 (18:30 +0200)]
riscv: Ensure isa-ext static keys are writable
riscv_isa_ext_keys[] is an array of static keys used in the unified
ISA extension framework. The keys added to this array may be used
anywhere, including in modules. Ensure the keys remain writable by
placing them in the data section.
The need to change riscv_isa_ext_keys[]'s section was found when the
kvm module started failing to load. Commit 8eb060e10185 ("arch/riscv:
add Zihintpause support") adds a static branch check for a newly
added isa-ext key to cpu_relax(), which kvm uses.
Maíra Canal [Mon, 15 Aug 2022 11:39:31 +0000 (08:39 -0300)]
drm/amdgpu: Fix use-after-free on amdgpu_bo_list mutex
If amdgpu_cs_vm_handling returns r != 0, then it will unlock the
bo_list_mutex inside the function amdgpu_cs_vm_handling and again on
amdgpu_cs_parser_fini. This problem results in the following
use-after-free problem:
Mukul Joshi [Fri, 12 Aug 2022 19:23:51 +0000 (15:23 -0400)]
drm/amdgpu: Fix interrupt handling on ih_soft ring
There are no backing hardware registers for ih_soft ring.
As a result, don't try to access hardware registers for read
and write pointers when processing interrupts on the IH soft
ring.
Alvin Lee [Thu, 4 Aug 2022 18:42:47 +0000 (14:42 -0400)]
drm/amd/display: Include scaling factor for SubVP command
[Description]
For SubVP scaling cases, we must include the scaling
info as part of the cmd. This is required when converting
OTG line to HUBP line for the MALL_START_LINE programming.
Khalid Masum [Mon, 15 Aug 2022 18:34:25 +0000 (00:34 +0600)]
drm/amdgpu/vcn: Return void from the stop_dbg_mode
There is no point in returning an int here. It only returns 0 which
the caller never uses. Therefore return void and remove the unnecessary
assignment.
drm/amd/display: Update clock table policy for DCN314
[Why & How]
Depending on how the clock table is constructed from PMFW we can run
into issues where we don't think we have enough bandwidth available
due to FCLK too low - eg. when the FCLK table contains invalid entries
or a single entry.
We should always pick up the maximum clocks for each state as a final
state in this case to prevent validation from failing if the table is
malformed.
We should also contain sensible defaults in the case where values
are invalid.
Redfine the clock table structures by adding a 314 prefix to make
debugging these issues easier by avoiding symbol name clashes.
Overall this policy more closely aligns to how we did things for 315,
but because of how the voltage rail is setup we should favor keeping
DCFCLK low rather than DISPCLK or DPPCLK - so use the max for those
in every entry.
[Why]
Recent backport from opensource broke the Nightly tool build
that tests DC and DML for bugs and regressions. This was
because the backport had a header inclusion that was not
consistent with the AMD style of including headers was allowed
to be merged back in DML code that caused tool compilation
failures.
[How]
Modify the way in which the header file in included so that it
is consistent with AMD style of including headers. This then
automatically fixes the tool compilation process and also
helps maintain the code quality and consistency.
Tom Chung [Wed, 3 Aug 2022 07:15:17 +0000 (15:15 +0800)]
drm/amd/display: Fix plug/unplug external monitor will hang while playback MPO video
[Why]
Pipes for MPO primary and overlay will be power down and power up during
plug/unplug external monitor while MPO video playback.
But the pipes were the same after plug/unplug and should not need to be
power down and power up or it will make page flip interrupt disabled and
cause hang issue.
[How]
Add pipe split change condition that not only check the top pipe pointer
but also check the index of top pipe if both top pipes are available.
[Why]
During multi-vf executing benchmark (Luxmark) observed kiq error timeout.
It happenes because all of VFs do the tlb invalidation at the same time.
Although each VF has the invalidate register set, from hardware side
the invalidate requests are queue to execute.
[How]
In case of 12 VF increase timeout on 12*100ms
[Why]
Some pixel clock values could cause HDMI TMDS SSCPs to be misaligned
between different HDMI lanes when using YCbCr420 10-bit pixel format.
BIOS functions for transmitter/encoder control take pixel clock in kHz
increments, whereas the function for setting the pixel clock is in 100Hz
increments. Setting pixel clock to a value that is not on a kHz boundary
will cause the issue.
[How]
Round pixel clock down to nearest kHz in 10/12-bpc cases.
Aric Cyr [Tue, 2 Aug 2022 00:17:54 +0000 (20:17 -0400)]
drm/amd/display: 3.2.198
This version brings along following fixes:
-Fix edp panel missing event
-Set ARGB16161616 pixel format to 26
-Fix dcn32 interger issue
-Clear optc underflow bit after ODM clock off
-Fix issue with stereo3D
-Fix DML2 lightup issue
-Correct DTBCLK for dcn314
-Revert for a regression
-Fix clocks and bugs in DML2
-Enable SubVP by defalut on DCN32 & DCN321
-Corret boundary condition for engin ID on DCN303
-Fix FRL encoder override registry key
-Fix VPG for dcn314 HPO
-Fix Linux compile-time warning
-Add new prefetch modes in DML for DCN32
drm/amd/display: reverted limiting vscsdp_for_colorimetry and ARGB16161616 pixel format addition
[WHY]
Limiting vscsdp_for_colorimetry for YCbCr420/BT2020 resulted in red/green
point failures in HDR10 DTN tests. The re-implementation of ARGB16161616
was to fix this however it did not actually fix this issue but a side effect of the
issue.
Dan Carpenter [Fri, 12 Aug 2022 06:20:58 +0000 (09:20 +0300)]
drm/amdkfd: potential crash in kfd_create_indirect_link_prop()
This code has two bugs. If kfd_topology_device_by_proximity_domain()
failed on the first iteration through the loop then "cpu_link" is
uninitialized and should not be dereferenced.
The second bug is that we cannot dereference a list iterator when it
points to the list head. In other words, if we exit the
list_for_each_entry() loop exits without hitting a break then "cpu_link"
is not a valid pointer and should not be dereferenced.
Fix both of these problems by setting "cpu_link" to NULL when it is invalid
and non-NULL when it is valid. That makes it easier to test for
valid vs invalid.
Fixes: 0f28cca87e9a ("drm/amdkfd: Extend KFD device topology to surface peer-to-peer links") Signed-off-by: Dan Carpenter <[email protected]> Signed-off-by: Felix Kuehling <[email protected]> Reviewed-by: Felix Kuehling <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
Felix Kuehling [Mon, 8 Aug 2022 21:00:38 +0000 (17:00 -0400)]
drm/amdkfd: Fix mm reference in SVM eviction worker
Use the mm reference from the fence. This allows removing the
svm_bo->svms pointer, which was problematic because we cannot assume
that the struct kfd_process containing the svms is still allocated
without holding a refcount on the process.
Use mmget_not_zero to ensure the mm is still valid, and drop the svm_bo
reference if it isn't.
Andrew Donnellan [Tue, 16 Aug 2022 05:17:20 +0000 (15:17 +1000)]
gcc-plugins: Undefine LATENT_ENTROPY_PLUGIN when plugin disabled for a file
Commit 36d4b36b6959 ("lib/nodemask: inline next_node_in() and
node_random()") refactored some code by moving node_random() from
lib/nodemask.c to include/linux/nodemask.h, thus requiring nodemask.h to
include random.h, which conditionally defines add_latent_entropy()
depending on whether the macro LATENT_ENTROPY_PLUGIN is defined.
This broke the build on powerpc, where nodemask.h is indirectly included
in arch/powerpc/kernel/prom_init.c, part of the early boot machinery that
is excluded from the latent entropy plugin using
DISABLE_LATENT_ENTROPY_PLUGIN. It turns out that while we add a gcc flag
to disable the actual plugin, we don't undefine LATENT_ENTROPY_PLUGIN.
This leads to the following:
CC arch/powerpc/kernel/prom_init.o
In file included from ./include/linux/nodemask.h:97,
from ./include/linux/mmzone.h:17,
from ./include/linux/gfp.h:7,
from ./include/linux/xarray.h:15,
from ./include/linux/radix-tree.h:21,
from ./include/linux/idr.h:15,
from ./include/linux/kernfs.h:12,
from ./include/linux/sysfs.h:16,
from ./include/linux/kobject.h:20,
from ./include/linux/pci.h:35,
from arch/powerpc/kernel/prom_init.c:24:
./include/linux/random.h: In function 'add_latent_entropy':
./include/linux/random.h:25:46: error: 'latent_entropy' undeclared (first use in this function); did you mean 'add_latent_entropy'?
25 | add_device_randomness((const void *)&latent_entropy, sizeof(latent_entropy));
| ^~~~~~~~~~~~~~
| add_latent_entropy
./include/linux/random.h:25:46: note: each undeclared identifier is reported only once for each function it appears in
make[2]: *** [scripts/Makefile.build:249: arch/powerpc/kernel/prom_init.o] Fehler 1
make[1]: *** [scripts/Makefile.build:465: arch/powerpc/kernel] Fehler 2
make: *** [Makefile:1855: arch/powerpc] Error 2
Change the DISABLE_LATENT_ENTROPY_PLUGIN flags to undefine
LATENT_ENTROPY_PLUGIN for files where the plugin is disabled.
exec: Replace kmap{,_atomic}() with kmap_local_page()
The use of kmap() and kmap_atomic() are being deprecated in favor of
kmap_local_page().
There are two main problems with kmap(): (1) It comes with an overhead as
mapping space is restricted and protected by a global lock for
synchronization and (2) it also requires global TLB invalidation when the
kmap’s pool wraps and it might block when the mapping space is fully
utilized until a slot becomes available.
With kmap_local_page() the mappings are per thread, CPU local, can take
page faults, and can be called from any context (including interrupts).
It is faster than kmap() in kernels with HIGHMEM enabled. Furthermore,
the tasks can be preempted and, when they are scheduled to run again, the
kernel virtual addresses are restored and are still valid.
Since the use of kmap_local_page() in exec.c is safe, it should be
preferred everywhere in exec.c.
As said, since kmap_local_page() can be also called from atomic context,
and since remove_arg_zero() doesn't (and shouldn't ever) rely on an
implicit preempt_disable(), this function can also safely replace
kmap_atomic().
Therefore, replace kmap() and kmap_atomic() with kmap_local_page() in
fs/exec.c.
Tested with xfstests on a QEMU/KVM x86_32 VM, 6GB RAM, booting a kernel
with HIGHMEM64GB enabled.
Linus Torvalds [Tue, 16 Aug 2022 18:42:11 +0000 (11:42 -0700)]
Merge tag 'nios2_fixes_v6.0' of git://git.kernel.org/pub/scm/linux/kernel/git/dinguyen/linux
Pull NIOS2 fixes from Dinh Nguyen:
- Security fixes from Al Viro
* tag 'nios2_fixes_v6.0' of git://git.kernel.org/pub/scm/linux/kernel/git/dinguyen/linux:
nios2: add force_successful_syscall_return()
nios2: restarts apply only to the first sigframe we build...
nios2: fix syscall restart checks
nios2: traced syscall does need to check the syscall number
nios2: don't leave NULLs in sys_call_table[]
nios2: page fault et.al. are *not* restartable syscalls...
Linus Torvalds [Tue, 16 Aug 2022 18:40:15 +0000 (11:40 -0700)]
Merge tag 'spi-fix-v6.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Pull spi fixes from Mark Brown:
"A few fixes that came in since my pull request, the Meson fix is a
little large since it's fixing all possible cases of the problem that
was observed with the driver and clock API trying to share
configuration by integrating the device clocking fully with the clock
API rather than spot fixing the one instance that was observed"
* tag 'spi-fix-v6.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
spi: dt-bindings: Drop Pratyush Yadav
spi: meson-spicc: add local pow2 clock ops to preserve rate between messages
MAINTAINERS: rectify entry for ARM/HPE GXP ARCHITECTURE
spi: spi.c: Add missing __percpu annotations in users of spi_statistics
Linus Torvalds [Tue, 16 Aug 2022 18:36:38 +0000 (11:36 -0700)]
Merge tag 'regulator-fix-v6.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator
Pull regulator fixes from Mark Brown:
"A couple of small fixes that came in since my pull request, nothing
major here"
* tag 'regulator-fix-v6.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
regulator: core: Fix missing error return from regulator_bulk_get()
regulator: pca9450: Remove restrictions for regulator-name
The exception for the "unaligned access at the end of the page, next
page not mapped" never happens, but the fixup code ends up causing
trouble for compilers to optimize well.
clang in particular ends up seeing it being in the middle of a loop, and
tries desperately to optimize the exception fixup code that is never
really reached.
The simple solution is to just move all the fixups into the exception
handler itself, which moves it all out of the hot case code, and means
that the compiler never sees it or needs to worry about it.
Hector Martin [Tue, 16 Aug 2022 07:03:11 +0000 (16:03 +0900)]
locking/atomic: Make test_and_*_bit() ordered on failure
These operations are documented as always ordered in
include/asm-generic/bitops/instrumented-atomic.h, and producer-consumer
type use cases where one side needs to ensure a flag is left pending
after some shared data was updated rely on this ordering, even in the
failure case.
This is the case with the workqueue code, which currently suffers from a
reproducible ordering violation on Apple M1 platforms (which are
notoriously out-of-order) that ends up causing the TTY layer to fail to
deliver data to userspace properly under the right conditions. This
change fixes that bug.
Change the documentation to restrict the "no order on failure" story to
the _lock() variant (for which it makes sense), and remove the
early-exit from the generic implementation, which is what causes the
missing barrier semantics in that case. Without this, the remaining
atomic op is fully ordered (including on ARM64 LSE, as of recent
versions of the architecture spec).
Suggested-by: Linus Torvalds <[email protected]> Cc: [email protected] Fixes: e986a0d6cb36 ("locking/atomics, asm-generic/bitops/atomic.h: Rewrite using atomic_*() APIs") Fixes: 61e02392d3c7 ("locking/atomic/bitops: Document and clarify ordering semantics for failed test_and_{}_bit()") Signed-off-by: Hector Martin <[email protected]> Acked-by: Will Deacon <[email protected]> Reviewed-by: Arnd Bergmann <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
Alan Brady [Tue, 2 Aug 2022 08:19:17 +0000 (10:19 +0200)]
i40e: Fix to stop tx_timeout recovery if GLOBR fails
When a tx_timeout fires, the PF attempts to recover by incrementally
resetting. First we try a PFR, then CORER and finally a GLOBR. If the
GLOBR fails, then we keep hitting the tx_timeout and incrementing the
recovery level and issuing dmesgs, which is both annoying to the user
and accomplishes nothing.
If the GLOBR fails, then we're pretty much totally hosed, and there's
not much else we can do to recover, so this makes it such that we just
kill the VSI and stop hitting the tx_timeout in such a case.
i40e: Fix tunnel checksum offload with fragmented traffic
Fix checksum offload on VXLAN tunnels.
In case, when mpls protocol is not used, set l4 header to transport
header of skb. This fixes case, when user tries to offload checksums
of VXLAN tunneled traffic.
Steps for reproduction (requires link partner with tunnels):
ip l s enp130s0f0 up
ip a f enp130s0f0
ip a a 10.10.110.2/24 dev enp130s0f0
ip l s enp130s0f0 mtu 1600
ip link add vxlan12_sut type vxlan id 12 group 238.168.100.100 dev \
enp130s0f0 dstport 4789
ip l s vxlan12_sut up
ip a a 20.10.110.2/24 dev vxlan12_sut
iperf3 -c 20.10.110.1 #should connect
Without this patch, TX descriptor was using wrong data, due to
l4 header pointing wrong address. NIC would then drop those packets
internally, due to incorrect TX descriptor data, which increased
GLV_TEPC register.
Cheng Xu [Wed, 10 Aug 2022 01:43:19 +0000 (09:43 +0800)]
RDMA/erdma: Correct the max_qp and max_cq capacities of the device
QP0 in HW is used for CMDQ, and the rest is for RDMA QPs. So the actual
max_qp capacity reported to core should be max_qp (reported by HW) - 1.
So does max_cq.
RDMA/cxgb4: fix accept failure due to increased cpl_t5_pass_accept_rpl size
Commit 'c2ed5611afd7' has increased the cpl_t5_pass_accept_rpl{} structure
size by 8B to avoid roundup. cpl_t5_pass_accept_rpl{} is a HW specific
structure and increasing its size will lead to unwanted adapter errors.
Current commit reverts the cpl_t5_pass_accept_rpl{} back to its original
and allocates zeroed skb buffer there by avoiding the memset for iss field.
Reorder code to minimize chip type checks.
Mark Bloch [Mon, 8 Aug 2022 07:48:06 +0000 (10:48 +0300)]
RDMA/mlx5: Use the proper number of ports
The cited commit allowed the driver to operate over HCAs that have
4 physical ports. Use the number of ports of the RDMA device in the for
loop instead of using the struct size.
projects / linux.git / log