| Age | Commit message (Collapse) | Author |
|---|
|
The AGF verifier does not check that the AGF length field is within
known good bounds. This has never been checked by runtime kernel
code (i.e. the lack of verification goes back to 1993) yet we assume
in many places that it is correct and verify other metdata against
it.
Add length verification to the AGF verifier. The length of the AGF
must be equal to the size of the AG specified in the superblock,
unless it is the last AG in the filesystem. In that case, it must be
less than or equal to sb->sb_agblocks and greater than
XFS_MIN_AG_BLOCKS, which is the smallest AG a growfs operation will
allow to exist.
This requires a bit of rework of the verifier function. We want to
verify metadata before we use it to verify other metadata. Hence
we need to verify the AGF sequence numbers before using them to
verify the length of the AGF. Then we can verify the AGF length
before we verify AGFL fields. Then we can verifier other fields that
are bounds limited by the AGF length.
And, finally, by calculating agf_length only once into a local
variable, we can collapse repeated "if (xfs_has_foo() &&"
conditionaly checks into single checks. This makes the code much
easier to follow as all the checks for a given feature are obviously
in the same place.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
|
|
If the journal geometry results in a sector or log stripe unit
validation problem, it indicates that we cannot set the log up to
safely write to the the journal. In these cases, we must abort the
mount because the corruption needs external intervention to resolve.
Similarly, a journal that is too large cannot be written to safely,
either, so we shouldn't allow those geometries to mount, either.
If the log is too small, we risk having transaction reservations
overruning the available log space and the system hanging waiting
for space it can never provide. This is purely a runtime hang issue,
not a corruption issue as per the first cases listed above. We abort
mounts of the log is too small for V5 filesystems, but we must allow
v4 filesystems to mount because, historically, there was no log size
validity checking and so some systems may still be out there with
undersized logs.
The problem is that on V4 filesystems, when we discover a log
geometry problem, we skip all the remaining checks and then allow
the log to continue mounting. This mean that if one of the log size
checks fails, we skip the log stripe unit check. i.e. we allow the
mount because a "non-fatal" geometry is violated, and then fail to
check the hard fail geometries that should fail the mount.
Move all these fatal checks to the superblock verifier, and add a
new check for the two log sector size geometry variables having the
same values. This will prevent any attempt to mount a log that has
invalid or inconsistent geometries long before we attempt to mount
the log.
However, for the minimum log size checks, we can only do that once
we've setup up the log and calculated all the iclog sizes and
roundoffs. Hence this needs to remain in the log mount code after
the log has been initialised. It is also the only case where we
should allow a v4 filesystem to continue running, so leave that
handling in place, too.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
|
|
If the current transaction holds a busy extent and we are trying to
allocate a new extent to fix up the free list, we can deadlock if
the AG is entirely empty except for the busy extent held by the
transaction.
This can occur at runtime processing an XEFI with multiple extents
in this path:
__schedule+0x22f at ffffffff81f75e8f
schedule+0x46 at ffffffff81f76366
xfs_extent_busy_flush+0x69 at ffffffff81477d99
xfs_alloc_ag_vextent_size+0x16a at ffffffff8141711a
xfs_alloc_ag_vextent+0x19b at ffffffff81417edb
xfs_alloc_fix_freelist+0x22f at ffffffff8141896f
xfs_free_extent_fix_freelist+0x6a at ffffffff8141939a
__xfs_free_extent+0x99 at ffffffff81419499
xfs_trans_free_extent+0x3e at ffffffff814a6fee
xfs_extent_free_finish_item+0x24 at ffffffff814a70d4
xfs_defer_finish_noroll+0x1f7 at ffffffff81441407
xfs_defer_finish+0x11 at ffffffff814417e1
xfs_itruncate_extents_flags+0x13d at ffffffff8148b7dd
xfs_inactive_truncate+0xb9 at ffffffff8148bb89
xfs_inactive+0x227 at ffffffff8148c4f7
xfs_fs_destroy_inode+0xb8 at ffffffff81496898
destroy_inode+0x3b at ffffffff8127d2ab
do_unlinkat+0x1d1 at ffffffff81270df1
do_syscall_64+0x40 at ffffffff81f6b5f0
entry_SYSCALL_64_after_hwframe+0x44 at ffffffff8200007c
This can also happen in log recovery when processing an EFI
with multiple extents through this path:
context_switch() kernel/sched/core.c:3881
__schedule() kernel/sched/core.c:5111
schedule() kernel/sched/core.c:5186
xfs_extent_busy_flush() fs/xfs/xfs_extent_busy.c:598
xfs_alloc_ag_vextent_size() fs/xfs/libxfs/xfs_alloc.c:1641
xfs_alloc_ag_vextent() fs/xfs/libxfs/xfs_alloc.c:828
xfs_alloc_fix_freelist() fs/xfs/libxfs/xfs_alloc.c:2362
xfs_free_extent_fix_freelist() fs/xfs/libxfs/xfs_alloc.c:3029
__xfs_free_extent() fs/xfs/libxfs/xfs_alloc.c:3067
xfs_trans_free_extent() fs/xfs/xfs_extfree_item.c:370
xfs_efi_recover() fs/xfs/xfs_extfree_item.c:626
xlog_recover_process_efi() fs/xfs/xfs_log_recover.c:4605
xlog_recover_process_intents() fs/xfs/xfs_log_recover.c:4893
xlog_recover_finish() fs/xfs/xfs_log_recover.c:5824
xfs_log_mount_finish() fs/xfs/xfs_log.c:764
xfs_mountfs() fs/xfs/xfs_mount.c:978
xfs_fs_fill_super() fs/xfs/xfs_super.c:1908
mount_bdev() fs/super.c:1417
xfs_fs_mount() fs/xfs/xfs_super.c:1985
legacy_get_tree() fs/fs_context.c:647
vfs_get_tree() fs/super.c:1547
do_new_mount() fs/namespace.c:2843
do_mount() fs/namespace.c:3163
ksys_mount() fs/namespace.c:3372
__do_sys_mount() fs/namespace.c:3386
__se_sys_mount() fs/namespace.c:3383
__x64_sys_mount() fs/namespace.c:3383
do_syscall_64() arch/x86/entry/common.c:296
entry_SYSCALL_64() arch/x86/entry/entry_64.S:180
To avoid this deadlock, we should not block in
xfs_extent_busy_flush() if we hold a busy extent in the current
transaction.
Now that the EFI processing code can handle requeuing a partially
completed EFI, we can detect this situation in
xfs_extent_busy_flush() and return -EAGAIN rather than going to
sleep forever. The -EAGAIN get propagated back out to the
xfs_trans_free_extent() context, where the EFD is populated and the
transaction is rolled, thereby moving the busy extents into the CIL.
At this point, we can retry the extent free operation again with a
clean transaction. If we hit the same "all free extents are busy"
situation when trying to fix up the free list, we can safely call
xfs_extent_busy_flush() and wait for the busy extents to resolve
and wake us. At this point, the allocation search can make progress
again and we can fix up the free list.
This deadlock was first reported by Chandan in mid-2021, but I
couldn't make myself understood during review, and didn't have time
to fix it myself.
It was reported again in March 2023, and again I have found myself
unable to explain the complexities of the solution needed during
review.
As such, I don't have hours more time to waste trying to get the
fix written the way it needs to be written, so I'm just doing it
myself. This patchset is largely based on Wengang Wang's last patch,
but with all the unnecessary stuff removed, split up into multiple
patches and cleaned up somewhat.
Reported-by: Chandan Babu R <chandanrlinux@gmail.com>
Reported-by: Wengang Wang <wen.gang.wang@oracle.com>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
|
|
Extent freeing neeeds to be able to avoid a busy extent deadlock
when the transaction itself holds the only busy extents in the
allocation group. This may occur if we have an EFI that contains
multiple extents to be freed, and the freeing the second intent
requires the space the first extent free released to expand the
AGFL. If we block on the busy extent at this point, we deadlock.
We hold a dirty transaction that contains a entire atomic extent
free operations within it, so if we can abort the extent free
operation and commit the progress that we've made, the busy extent
can be resolved by a log force. Hence we can restart the aborted
extent free with a new transaction and continue to make
progress without risking deadlocks.
To enable this, we need the EFI processing code to be able to handle
an -EAGAIN error to tell it to commit the current transaction and
retry again. This mechanism is already built into the defer ops
processing (used bythe refcount btree modification intents), so
there's relatively little handling we need to add to the EFI code to
enable this.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandan.babu@oracle.com>
|
|
To avoid blocking in xfs_extent_busy_flush() when freeing extents
and the only busy extents are held by the current transaction, we
need to pass the XFS_ALLOC_FLAG_FREEING flag context all the way
into xfs_extent_busy_flush().
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandan.babu@oracle.com>
|
|
Btrees that aren't freespace management trees use the normal extent
allocation and freeing routines for their blocks. Hence when a btree
block is freed, a direct call to xfs_free_extent() is made and the
extent is immediately freed. This puts the entire free space
management btrees under this path, so we are stacking btrees on
btrees in the call stack. The inobt, finobt and refcount btrees
all do this.
However, the bmap btree does not do this - it calls
xfs_free_extent_later() to defer the extent free operation via an
XEFI and hence it gets processed in deferred operation processing
during the commit of the primary transaction (i.e. via intent
chaining).
We need to change xfs_free_extent() to behave in a non-blocking
manner so that we can avoid deadlocks with busy extents near ENOSPC
in transactions that free multiple extents. Inserting or removing a
record from a btree can cause a multi-level tree merge operation and
that will free multiple blocks from the btree in a single
transaction. i.e. we can call xfs_free_extent() multiple times, and
hence the btree manipulation transaction is vulnerable to this busy
extent deadlock vector.
To fix this, convert all the remaining callers of xfs_free_extent()
to use xfs_free_extent_later() to queue XEFIs and hence defer
processing of the extent frees to a context that can be safely
restarted if a deadlock condition is detected.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandan.babu@oracle.com>
|
|
XFS has strict metadata ordering requirements. One of the things it
does is maintain the commit order of items from transaction commit
through the CIL and into the AIL. That is, if a transaction logs
item A before item B in a modification, then they will be inserted
into the CIL in the order {A, B}. These items are then written into
the iclog during checkpointing in the order {A, B}. When the
checkpoint commits, they are supposed to be inserted into the AIL in
the order {A, B}, and when they are pushed from the AIL, they are
pushed in the order {A, B}.
If we crash, log recovery then replays the two items from the
checkpoint in the order {A, B}, resulting in the objects the items
apply to being queued for writeback at the end of the checkpoint
in the order {A, B}. This means recovery behaves the same way as the
runtime code.
In places, we have subtle dependencies on this ordering being
maintained. One of this place is performing intent recovery from the
log. It assumes that recovering an intent will result in a
non-intent object being the first thing that is modified in the
recovery transaction, and so when the transaction commits and the
journal flushes, the first object inserted into the AIL beyond the
intent recovery range will be a non-intent item. It uses the
transistion from intent items to non-intent items to stop the
recovery pass.
A recent log recovery issue indicated that an intent was appearing
as the first item in the AIL beyond the recovery range, hence
breaking the end of recovery detection that exists.
Tracing indicated insertion of the items into the AIL was apparently
occurring in the right order (the intent was last in the commit item
list), but the intent was appearing first in the AIL. IOWs, the
order of items in the AIL was {D,C,B,A}, not {A,B,C,D}, and bulk
insertion was reversing the order of the items in the batch of items
being inserted.
Lucky for us, all the items fed to bulk insertion have the same LSN,
so the reversal of order does not affect the log head/tail tracking
that is based on the contents of the AIL. It only impacts on code
that has implicit, subtle dependencies on object order, and AFAICT
only the intent recovery loop is impacted by it.
Make sure bulk AIL insertion does not reorder items incorrectly.
Fixes: 0e57f6a36f9b ("xfs: bulk AIL insertion during transaction commit")
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandan.babu@oracle.com>
|
|
Pointers drop_leaf and save_leaf are initialized with values that are never
read, they are being re-assigned later on just before they are used. Remove
the redundant early initializations and keep the later assignments at the
point where they are used. Cleans up two clang scan build warnings:
fs/xfs/libxfs/xfs_attr_leaf.c:2288:29: warning: Value stored to 'drop_leaf'
during its initialization is never read [deadcode.DeadStores]
fs/xfs/libxfs/xfs_attr_leaf.c:2289:29: warning: Value stored to 'save_leaf'
during its initialization is never read [deadcode.DeadStores]
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
|
|
As Darrick prefers to micro-manage this without looking at my input
for code I wrote and then complain about getting burned out by that
I might as well drop myself from the maintainers file.
[djwong: and I don't like it when you all bikeshed a single-patch
submission to the point that new bugs are being introduced in the scope
creep that wasn't part of V1-V7!]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
|
|
iomap_file_buffered_write_punch_delalloc
The return value type of i_blocksize() is 'unsigned int', so the
type of blocksize has been modified from 'int' to 'unsigned int'
to ensure data type consistency.
Signed-off-by: Lu Hongfei <luhongfei@vivo.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
|
|
It is perfectly valid to not find session not found errors
when a reconnect of a session happens when requests for the
same session are happening in parallel.
We had these log messages as VFS logs. My last change dumped
these logs as FYI logs.
This change just creates a new dynamic tracepoint to capture
events of this type, just in case it is useful while
debugging issues in the future.
Signed-off-by: Shyam Prasad N <sprasad@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
|
|
We do not log the session id in crypt_setup when a matching
session is not found. Printing the session id helps debugging
here. This change does just that.
This change also changes this log to FYI, since it is normal to
see then during a reconnect. Doing the same for a similar log
in case of signed connections.
The plan is to have a tracepoint for this event, so that we will
be able to see this event if need be. That will be done as
another change.
Signed-off-by: Shyam Prasad N <sprasad@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
|
|
The kdb_kbd_cleanup_state() is called from another file through
the kdb_private.h file, but that is not included before the
definition, causing a W=1 warning:
kernel/debug/kdb/kdb_keyboard.c:198:6: error: no previous prototype for 'kdb_kbd_cleanup_state' [-Werror=missing-prototypes]
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20230517124802.929751-1-arnd@kernel.org
Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org>
|
|
Commit 7ae24fcee992 ("cdrom: remove the unused mode argument to
cdrom_release") was supposed to remove an unused argument from
cdrom_release(). but instead removed a used argument from
cdrom_open(). This results in the following build error.
drivers/cdrom/gdrom.c: In function 'gdrom_bdops_open':
drivers/cdrom/gdrom.c:484:15: error: too few arguments to function 'cdrom_open'
drivers/cdrom/gdrom.c: In function 'gdrom_bdops_release':
drivers/cdrom/gdrom.c:492:35: error: 'mode' undeclared
Fix it up.
Fixes: 7ae24fcee992 ("cdrom: remove the unused mode argument to cdrom_release")
Cc: Christoph Hellwig <hch@lst.de>
Cc: Phillip Potter <phil@philpotter.co.uk>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/20230629083047.3487172-1-linux@roeck-us.net
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
To save architectures from needing to wrap the call in #ifdefs, add a
stub no-op version of kgdb_nmicallback(), which returns 1 if it didn't
handle anything.
Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://lore.kernel.org/r/20230601143109.v9.6.Ia3aeac89bb6751b682237e76e5ba594318e4b1aa@changeid
Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org>
|
|
For now, DIE_PAGE_FAULT, DIE_BREAK, DIE_SSTEPBP, DIE_UPROBE and
DIE_UPROBE_XOL are not used by any code, remove them.
Tested-by: Jeff Xie <xiehuan09@gmail.com>
Suggested-by: Youling Tang <tangyouling@loongson.cn>
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Uprobes is the user-space counterpart to kprobes, this patch adds
uprobes support for LoongArch.
Here is a simple example with CONFIG_UPROBE_EVENTS=y:
# cat test.c
#include <stdio.h>
int add(int a, int b)
{
return a + b;
}
int main()
{
return add(2, 7);
}
# gcc test.c -o /tmp/test
# nm /tmp/test | grep add
0000000120004194 T add
# cd /sys/kernel/debug/tracing
# echo > uprobe_events
# echo "p:myuprobe /tmp/test:0x4194 %r4 %r5" > uprobe_events
# echo "r:myuretprobe /tmp/test:0x4194 %r4" >> uprobe_events
# echo 1 > events/uprobes/enable
# echo 1 > tracing_on
# /tmp/test
# cat trace
...
# TASK-PID CPU# ||||| TIMESTAMP FUNCTION
# | | | ||||| | |
test-1060 [001] DNZff 1015.770620: myuprobe: (0x120004194) arg1=0x2 arg2=0x7
test-1060 [001] DNZff 1015.770930: myuretprobe: (0x1200041f0 <- 0x120004194) arg1=0x9
Tested-by: Jeff Xie <xiehuan09@gmail.com>
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
For now, we can use larch_insn_gen_break() to define KPROBE_BP_INSN and
KPROBE_SSTEPBP_INSN. Because larch_insn_gen_break() returns instruction
word, define kprobe_opcode_t as u32, then do some small changes related
with type conversion, no functional change intended.
Tested-by: Jeff Xie <xiehuan09@gmail.com>
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
There exist various break insns such as BRK_KPROBE_BP, BRK_KPROBE_SSTEPBP,
BRK_UPROBE_BP and BRK_UPROBE_XOLBP, add larch_insn_gen_break() to generate
break insns simpler, this is preparation for later patch.
Tested-by: Jeff Xie <xiehuan09@gmail.com>
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Like llsc instructions, the atomic memory access instructions shouldn't
be supported for probing, so check for them in insns_not_supported().
Closes: https://lore.kernel.org/all/SY4P282MB351877A70A0333C790FE85A5C09C9@SY4P282MB3518.AUSP282.PROD.OUTLOOK.COM/
Tested-by: Jeff Xie <xiehuan09@gmail.com>
Reported-by: Hengqi Chen <hengqi.chen@gmail.com>
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
The three functions insns_not_supported(), insns_need_simulation() and
arch_simulate_insn() will be used for uprobes, move them from kprobes.c
to inst.c, this is preparation for later patch, no functionality change.
Tested-by: Jeff Xie <xiehuan09@gmail.com>
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
This is an adaptation of commit f3a112c0c40d ("x86,rethook,kprobes:
Replace kretprobe with rethook on x86") and commit b57c2f124098 ("riscv:
add riscv rethook implementation") to LoongArch. Mainly refer to commit
b57c2f124098 ("riscv: add riscv rethook implementation").
Replaces the kretprobe code with rethook on LoongArch. With this patch,
kretprobe on LoongArch uses the rethook instead of kretprobe specific
trampoline code.
Signed-off-by: Haoran Jiang <jianghaoran@kylinos.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Add support for jump labels based on the ARM64 version.
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Youling Tang <tangyouling@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
We can see that DEBUG_KMEMLEAK depends on HAVE_DEBUG_KMEMLEAK after
commit b69ec42b1b19 ("Kconfig: clean up the long arch list for the
DEBUG_KMEMLEAK config option"), just select HAVE_DEBUG_KMEMLEAK to
support kmemleak on LoongArch.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Some PMC (Power Management Controllers) need to support DTS and will use
the suspend interfaces thus this patch was to export such interfaces for
their use.
Signed-off-by: Yinbo Zhu <zhuyinbo@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Loongson-3A6000 and newer processors have hardware page table walker
(PTW) support. PTW can handle all fastpaths of TLBI/TLBL/TLBS/TLBM
exceptions by hardware, software only need to handle slowpaths (page
faults).
BTW, PTW doesn't append _PAGE_MODIFIED for page table entries, so we
change pmd_dirty() and pte_dirty() to also check _PAGE_DIRTY for the
"dirty" attribute.
Signed-off-by: Liang Gao <gaoliang@loongson.cn>
Signed-off-by: Jun Yi <yijun@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Traditionally, LoongArch uses "dbar 0" (full completion barrier) for
everything. But the full completion barrier is a performance killer, so
Loongson-3A6000 and newer processors have made finer granularity hints
available:
Bit4: ordering or completion (0: completion, 1: ordering)
Bit3: barrier for previous read (0: true, 1: false)
Bit2: barrier for previous write (0: true, 1: false)
Bit1: barrier for succeeding read (0: true, 1: false)
Bit0: barrier for succeeding write (0: true, 1: false)
Hint 0x700: barrier for "read after read" from the same address, which
is needed by LL-SC loops on old models (dbar 0x700 behaves the same as
nop if such reordering is disabled on new models).
This patch makes use of the various new hints for different kinds of
memory barriers. It brings performance improvements on Loongson-3A6000
series, while not affecting the existing models because all variants are
treated as 'dbar 0' there.
Why override queued_spin_unlock()?
After commit 01e3b958efe85a26d9b ("drivers: Remove explicit invocations
of mmiowb()") we need a completion barrier in queued_spin_unlock(), but
the generic implementation use smp_store_release() which only provide an
ordering barrier.
Signed-off-by: Jun Yi <yijun@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Loongson-3A6000 has SMT (Simultaneous Multi-Threading) support, each
physical core has two logical cores (threads). This patch add SMT probe
and scheduler support via ACPI PPTT.
If SCHED_SMT enabled, Loongson-3A6000 is treated as 4 cores, 8 threads;
If SCHED_SMT disabled, Loongson-3A6000 is treated as 8 cores, 8 threads.
Remove smp_num_siblings to support HMP (Heterogeneous Multi-Processing).
Signed-off-by: Liupu Wang <wangliupu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Add LoongArch's vector extensions support, which including 128bit LSX
(i.e., Loongson SIMD eXtension) and 256bit LASX (i.e., Loongson Advanced
SIMD eXtension).
Linux kernel doesn't use vector itself, it only handle exceptions and
context save/restore. So it only needs a subset of these instructions:
* Vector load/store: vld vst vldx vstx xvld xvst xvldx xvstx
* 8bit-elements move: vpickve2gr.b xvpickve2gr.b vinsgr2vr.b xvinsgr2vr.b
* 16bit-elements move: vpickve2gr.h xvpickve2gr.h vinsgr2vr.h xvinsgr2vr.h
* 32bit-elements move: vpickve2gr.w xvpickve2gr.w vinsgr2vr.w xvinsgr2vr.w
* 64bit-elements move: vpickve2gr.d xvpickve2gr.d vinsgr2vr.d xvinsgr2vr.d
* Elements permute: vpermi.w vpermi.d xvpermi.w xvpermi.d xvpermi.q
Introduce AS_HAS_LSX_EXTENSION and AS_HAS_LASX_EXTENSION to avoid non-
vector toolchains complains unsupported instructions.
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
We can see that "Time namespaces are not supported" on LoongArch:
(1) clone3 test
# cd tools/testing/selftests/clone3 && make && ./clone3
...
# Time namespaces are not supported
ok 18 # SKIP Skipping clone3() with CLONE_NEWTIME
# Totals: pass:17 fail:0 xfail:0 xpass:0 skip:1 error:0
(2) timens test
# cd tools/testing/selftests/timens && make && ./timens
...
1..0 # SKIP Time namespaces are not supported
On LoongArch the current kernel does not support CONFIG_TIME_NS which
depends on GENERIC_VDSO_TIME_NS, select GENERIC_VDSO_TIME_NS to enable
CONFIG_TIME_NS to build kernel/time/namespace.c.
Additionally, it needs to define some arch-dependent functions for the
timens, such as __arch_get_timens_vdso_data(), arch_get_vdso_data() and
vdso_join_timens().
At the same time, modify the layout of vvar to use one page size for
generic vdso data, expand another page size for timens vdso data and
assign LOONGARCH_VDSO_DATA_SIZE (maybe exceeds a page size if expand in
the future) for loongarch vdso data, at last add the callback function
vvar_fault() and modify stack_top().
With this patch under CONFIG_TIME_NS:
(1) clone3 test
# cd tools/testing/selftests/clone3 && make && ./clone3
...
ok 18 [739] Result (0) matches expectation (0)
# Totals: pass:18 fail:0 xfail:0 xpass:0 skip:0 error:0
(2) timens test
# cd tools/testing/selftests/timens && make && ./timens
...
# Totals: pass:10 fail:0 xfail:0 xpass:0 skip:0 error:0
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
The LoongArch kernel is 64-bit and built with the soft-float ABI,
hence the loongarch64-linux-gnusf target. (The "libc" part can affect
the codegen of libcalls: other arches do not use a bare-metal target,
and currently the only fully supported libc on LoongArch is glibc
anyway.)
See: https://lore.kernel.org/loongarch/CAKwvOdnimxv8oJ4mVY74zqtt1x7KTMrWvn2_T9x22SFDbU6rHQ@mail.gmail.com/
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Confirmed working with QEMU system emulation.
Acked-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
This is a port of commit 08f6554ff90e ("mips: Include KBUILD_CPPFLAGS in
CHECKFLAGS invocation") to arch/loongarch, for fixing cross-compilation
of Linux/LoongArch with Clang, where previously the `--target` flag
would no longer be present for the CHECKFLAGS cc invocation leading to
build failure.
Reported-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Link: https://github.com/ClangBuiltLinux/linux/issues/1787#issuecomment-1608306002
Signed-off-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
This is a port of commit 76d7fff22be3e ("MIPS: VDSO: Use CLANG_FLAGS
instead of filtering out '--target='") to arch/loongarch, for fixing
cross-compilation with Clang.
Reported-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Link: https://github.com/ClangBuiltLinux/linux/issues/1787#issuecomment-1608306002
Signed-off-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Now the arch code is mostly ready for LLVM/Clang consumption, it is time
to re-organize the CFLAGS a little to actually enable the LLVM build.
Namely, all -G0 switches from CFLAGS are removed, and -mexplicit-relocs
and -mdirect-extern-access are now wrapped with cc-option (with the
related asm/percpu.h definition guarded against toolchain combos that
are known to not work).
A build with !RELOCATABLE && !MODULE is confirmed working within a QEMU
environment; support for the two features are currently blocked on
LLVM/Clang, and will come later.
Why -G0 can be removed:
In GCC, -G stands for "small data threshold", that instructs the
compiler to put data smaller than the specified threshold in a dedicated
"small data" section (called .sdata on LoongArch and several other
arches).
However, benefiting from this would require ABI cooperation, which is
not the case for LoongArch; and current GCC behave the same whether -G0
(equal to disabling this optimization) is given or not. So, remove -G0
from CFLAGS altogether for one less thing to care about. This also
benefits LLVM/Clang compatibility where the -G switch is not supported.
Why -mexplicit-relocs can now be conditionally applied without
regressions:
Originally -mexplicit-relocs is unconditionally added to CFLAGS in case
of CONFIG_AS_HAS_EXPLICIT_RELOCS, because not having it (i.e. old GCC +
new binutils) would not work: modules will have R_LARCH_ABS_* relocs
inside, but given the rarity of such toolchain combo in the wild, it may
not be worthwhile to support it, so support for such relocs in modules
were not added back when explicit relocs support was upstreamed, and
-mexplicit-relocs is unconditionally added to fail the build early.
Now that Clang compatibility is desired, given Clang is behaving like
-mexplicit-relocs from day one but without support for the CLI flag, we
must ensure the flag is not passed in case of Clang. However, explicit
compiler flavor checks can be more brittle than feature detection: in
this case what actually matters is support for __attribute__((model))
when building modules. Given neither older GCC nor current Clang support
this attribute, probing for the attribute support and #error'ing out
would allow proper UX without checking for Clang, and also automatically
work when Clang support for the attribute is to be added in the future.
Why -mdirect-extern-access is now conditionally applied:
This is actually a nice-to-have optimization that can reduce GOT
accesses, but not having it is harmless either. Because Clang does not
support the option currently, but might do so in the future, conditional
application via cc-option ensures compatibility with both current and
future Clang versions.
Suggested-by: Xi Ruoyao <xry111@xry111.site> # cc-option changes
Signed-off-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
The invtlb instruction has been supported by upstream LoongArch
toolchains from day one, so ditch the raw opcode trickery and just use
plain inline asm for it.
While at it, also make the invtlb asm statements barriers, for proper
modeling of the side effects. The functions are also marked as
__always_inline instead of just "inline", because they cannot work at
all if not inlined: the op argument will not be compile-time const in
that case, thus failing to satisfy the "i" constraint.
The signature of the other more specific invtlb wrappers contain unused
arguments right now, but these are not removed right away in order for
the patch to be focused. In the meantime, assertions are added to ensure
no accidental misuse happens before the refactor. (The more specific
wrappers cannot re-use the generic invtlb wrapper, because the ISA
manual says $zero shall be used in case a particular op does not take
the respective argument: re-using the generic wrapper would mean losing
control over the register usage.)
Signed-off-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
In addition to less visual clutter, this also makes Clang happy
regarding the const-ness of arguments. In the original approach, all
Clang gets to see is the incoming arguments whose const-ness cannot be
proven without first being inlined; so Clang errors out here while GCC
is fine.
While at it, tweak several printk format strings because the return type
of csr_read64 becomes effectively unsigned long, instead of unsigned
long long.
Signed-off-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
The GNU assembler (as of 2.40) mis-treats FCSR operands as GPRs, but
the LLVM IAS does not. Probe for this and refer to FCSRs as "$fcsrNN"
if support is present.
Signed-off-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
When the kernel is compiled with LLVM, the register names being handled
during exception fixup building are ABI names instead of bare $rNN
style. Add mapping for the ABI names for LLVM compatibility.
Signed-off-by: WANG Rui <wangrui@loongson.cn>
Signed-off-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Taking the address delta between symbols in different sections is not
supported by the LLVM IAS. Instead, do this in the linker script, so
the same data can be properly referenced in assembly.
Signed-off-by: WANG Rui <wangrui@loongson.cn>
Signed-off-by: WANG Xuerui <git@xen0n.name>
[chenhuacai: Fix build with !CONFIG_EFI_STUB]
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Add guard for the larch_insn_gen_xxx functions to verify whether the
immediate operand is within the acceptable range.
Signed-off-by: WANG Rui <wangrui@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Debugfs functions are not supposed to be checked for errors. This
is sort of unusual but it is described in the comments for the
debugfs_create_dir() function. Also debugfs_create_dir() can never
return NULL.
Reviewed-by: WANG Xuerui <git@xen0n.name>
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
ACPI systems set io masters by parsing ACPI MADT, FDT systems have no
MADT so we explicitly set CPU#0 as the io master. Otherwise CPU#0 will
be considered as hotpluggable.
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
|
|
Vladimir Oltean says:
====================
Fix PTP received on wrong port with bridged SJA1105 DSA
Since the changes were made to tag_8021q to support imprecise RX for
bridged ports, the tag_sja1105 driver still prefers the source port
information deduced from the VLAN headers for link-local traffic, even
though the switch can theoretically do better and report the precise
source port.
The problem is that the tagger doesn't know when to trust one source of
information over another, because the INCL_SRCPT option (to "tag" link
local frames) is sometimes enabled and sometimes it isn't.
The first patch makes the switch provide the hardware tag for link local
traffic under all circumstances, and the second patch makes the tagger
always use that hardware tag as primary source of information for link
local packets.
====================
Link: https://lore.kernel.org/r/20230627094207.3385231-1-vladimir.oltean@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Currently the sja1105 tagging protocol prefers using the source port
information from the VLAN header if that is available, falling back to
the INCL_SRCPT option if it isn't. The VLAN header is available for all
frames except for META frames initiated by the switch (containing RX
timestamps), and thus, the "if (is_link_local)" branch is practically
dead.
The tag_8021q source port identification has become more loose
("imprecise") and will report a plausible rather than exact bridge port,
when under a bridge (be it VLAN-aware or VLAN-unaware). But link-local
traffic always needs to know the precise source port. With incorrect
source port reporting, for example PTP traffic over 2 bridged ports will
all be seen on sockets opened on the first such port, which is incorrect.
Now that the tagging protocol has been changed to make link-local frames
always contain source port information, we can reverse the order of the
checks so that we always give precedence to that information (which is
always precise) in lieu of the tag_8021q VID which is only precise for a
standalone port.
Fixes: d7f9787a763f ("net: dsa: tag_8021q: add support for imprecise RX based on the VBID")
Fixes: 91495f21fcec ("net: dsa: tag_8021q: replace the SVL bridging with VLAN-unaware IVL bridging")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Link-local traffic on bridged SJA1105 ports is sometimes tagged by the
hardware with source port information (when the port is under a VLAN
aware bridge).
The tag_8021q source port identification has become more loose
("imprecise") and will report a plausible rather than exact bridge port,
when under a bridge (be it VLAN-aware or VLAN-unaware). But link-local
traffic always needs to know the precise source port.
Modify the driver logic (and therefore: the tagging protocol itself) to
always include the source port information with link-local packets,
regardless of whether the port is standalone, under a VLAN-aware or
VLAN-unaware bridge. This makes it possible for the tagging driver to
give priority to that information over the tag_8021q VLAN header.
The big drawback with INCL_SRCPT is that it makes it impossible to
distinguish between an original MAC DA of 01:80:C2:XX:YY:ZZ and
01:80:C2:AA:BB:ZZ, because the tagger just patches MAC DA bytes 3 and 4
with zeroes. Only if PTP RX timestamping is enabled, the switch will
generate a META follow-up frame containing the RX timestamp and the
original bytes 3 and 4 of the MAC DA. Those will be used to patch up the
original packet. Nonetheless, in the absence of PTP RX timestamping, we
have to live with this limitation, since it is more important to have
the more precise source port information for link-local traffic.
Fixes: d7f9787a763f ("net: dsa: tag_8021q: add support for imprecise RX based on the VBID")
Fixes: 91495f21fcec ("net: dsa: tag_8021q: replace the SVL bridging with VLAN-unaware IVL bridging")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Vladimir Oltean says:
====================
Fix PTP packet drops with ocelot-8021q DSA tag protocol
Changes in v2:
- Distinguish between L2 and L4 PTP packets
v1 at:
https://lore.kernel.org/netdev/20230626154003.3153076-1-vladimir.oltean@nxp.com/
Patch 3/3 fixes an issue with the ocelot/felix driver, where it would
drop PTP traffic on RX unless hardware timestamping for that packet type
was enabled.
Fixing that requires the driver to know whether it had previously
configured the hardware to timestamp PTP packets on that port. But it
cannot correctly determine that today using the existing code structure,
so patches 1/3 and 2/3 fix the control path of the code such that
ocelot->ports[port]->trap_proto faithfully reflects whether that
configuration took place.
====================
Link: https://lore.kernel.org/r/20230627163114.3561597-1-vladimir.oltean@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
is disabled
The driver implements a workaround for the fact that it doesn't have an
IRQ source to tell it whether PTP frames are available through the
extraction registers, for those frames to be processed and passed
towards the network stack. That workaround is to configure the switch,
through felix_hwtstamp_set() -> felix_update_trapping_destinations(),
to create two copies of PTP packets: one sent over Ethernet to the DSA
master, and one to be consumed through the aforementioned CPU extraction
queue registers.
The reason why we want PTP packets to be consumed through the CPU
extraction registers in the first place is because we want to see their
hardware RX timestamp. With tag_8021q, that is only visible that way,
and it isn't visible with the copy of the packet that's transmitted over
Ethernet.
The problem with the workaround implementation is that it drops the
packet received over Ethernet, in expectation of its copy being present
in the CPU extraction registers. However, if felix_hwtstamp_set() hasn't
run (aka PTP RX timestamping is disabled), the driver will drop the
original PTP frame and there will be no copy of it in the CPU extraction
registers. So, the network stack will simply not see any PTP frame.
Look at the port's trapping configuration to see whether the driver has
previously enabled the CPU extraction registers. If it hasn't, just
don't RX timestamp the frame and let it be passed up the stack by DSA,
which is perfectly fine.
Fixes: 0a6f17c6ae21 ("net: dsa: tag_ocelot_8021q: add support for PTP timestamping")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
In a future change, the driver will need to determine whether PTP RX
timestamping is enabled on a port (including whether traps were set up
on that port in particular) and that is currently not possible.
The driver supports different RX filters (L2, L4) and kinds of TX
timestamping (one-step, two-step) on its ports, but it saves all
configuration in a single struct hwtstamp_config that is global to the
switch. So, the latest timestamping configuration on one port
(including a request to disable timestamping) affects what gets reported
for all ports, even though the configuration itself is still individual
to each port.
The port timestamping configurations are only coupled because of the
common structure, so replace the hwtstamp_config with a mask of trapped
protocols saved per port. We also have the ptp_cmd to distinguish
between one-step and two-step PTP timestamping, so with those 2 bits of
information we can fully reconstruct a descriptive struct
hwtstamp_config for each port, during the SIOCGHWTSTAMP ioctl.
Fixes: 4e3b0468e6d7 ("net: mscc: PTP Hardware Clock (PHC) support")
Fixes: 96ca08c05838 ("net: mscc: ocelot: set up traps for PTP packets")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
PTP RX timestamping should be enabled when the user requests it, not by
default. If it is enabled by default, it can be problematic when the
ocelot driver is a DSA master, and it sidesteps what DSA tries to avoid
through __dsa_master_hwtstamp_validate().
Additionally, after the change which made ocelot trap PTP packets only
to the CPU at ocelot_hwtstamp_set() time, it is no longer even true that
RX timestamping is enabled by default, because until ocelot_hwtstamp_set()
is called, the PTP traps are actually not set up. So the rx_filter field
of ocelot->hwtstamp_config reflects an incorrect reality.
Fixes: 96ca08c05838 ("net: mscc: ocelot: set up traps for PTP packets")
Fixes: 4e3b0468e6d7 ("net: mscc: PTP Hardware Clock (PHC) support")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
