pnfs: Fix a problem where we gratuitously start doing I/O through the MDS
If the client has to stop in pnfs_update_layout() to wait for another
layoutget to complete, it currently exits and defaults to I/O through
the MDS if the layoutget was successful.
Fixes: d03360aaf5cc ("pNFS: Ensure we return the error if someone kills...") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Cc: stable@vger.kernel.org # v4.20+
The NFSv4.1 protocol explicitly forbids us from using the zero stateid
together with layoutget, so when we see that nfs4_select_rw_stateid()
is unable to return a valid delegation, lock or open stateid, then
we should initiate recovery and retry.
SUNRPC: Replace division by multiplication in calculation of queue length
When checking whether or not a particular xprt queue length is shorter
than the average queue length for all xprts, prefer to use multiplication
rather than division for performance reasons.
NFSv4: Validate the stateid before applying it to state recovery
If the stateid is the zero or invalid stateid, then it is pointless
to attempt to use it for recovery. In that case, try to fall back
to using the open state stateid, or just doing a general recovery
of all state on a given inode.
Donald Buczek [Sun, 7 Jul 2019 19:26:10 +0000 (21:26 +0200)]
nfs4.0: Refetch lease_time after clientid update
RFC 7530 requires us to refetch the lease time attribute once a new
clientID is established. This is already implemented for the
nfs4.1(+) clients by nfs41_init_clientid, which calls
nfs41_finish_session_reset, which calls nfs4_setup_state_renewal.
To make nfs4_setup_state_renewal available for nfs4.0, move it
further to the top of the source file to include it regardles of
CONFIG_NFS_V4_1 and to save a forward declaration.
Call nfs4_setup_state_renewal from nfs4_init_clientid.
Signed-off-by: Donald Buczek <buczek@molgen.mpg.de> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Donald Buczek [Sun, 7 Jul 2019 19:26:08 +0000 (21:26 +0200)]
nfs4: Make nfs4_proc_get_lease_time available for nfs4.0
Compile nfs4_proc_get_lease_time, enc_get_lease_time and
dec_get_lease_time for nfs4.0. Use nfs4_sequence_done instead of
nfs41_sequence_done in nfs4_proc_get_lease_time,
Signed-off-by: Donald Buczek <buczek@molgen.mpg.de> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
That commit caused a severe memory leak in nfs_readdir_make_qstr().
When listing a directory with more than 100 files (this is how many
struct nfs_cache_array_entry elements fit in one 4kB page), all
allocated file name strings past those 100 leak.
The root of the leakage is that those string pointers are managed in
pages which are never linked into the page cache.
fs/nfs/dir.c puts pages into the page cache by calling
read_cache_page(); the callback function nfs_readdir_filler() will
then fill the given page struct which was passed to it, which is
already linked in the page cache (by do_read_cache_page() calling
add_to_page_cache_lru()).
Commit be4c2d4723a4 added another (local) array of allocated pages, to
be filled with more data, instead of discarding excess items received
from the NFS server. Those additional pages can be used by the next
nfs_readdir_filler() call (from within the same nfs_readdir() call).
The leak happens when some of those additional pages are never used
(copied to the page cache using copy_highpage()). The pages will be
freed by nfs_readdir_free_pages(), but their contents will not. The
commit did not invoke nfs_readdir_clear_array() (and doing so would
have been dangerous, because it did not track which of those pages
were already copied to the page cache, risking double free bugs).
How to reproduce the leak:
- Use a kernel with CONFIG_SLUB_DEBUG_ON.
- Create a directory on a NFS mount with more than 100 files with
names long enough to use the "kmalloc-32" slab (so we can easily
look up the allocation counts):
for i in `seq 110`; do touch ${i}_0123456789abcdef; done
110 allocations are gone, but 10 have leaked and will never be freed.
Unhelpfully, those allocations are explicitly excluded from KMEMLEAK,
that's why my initial attempts with KMEMLEAK were not successful:
/*
* Avoid a kmemleak false positive. The pointer to the name is stored
* in a page cache page which kmemleak does not scan.
*/
kmemleak_not_leak(string->name);
It would be possible to solve this bug without reverting the whole
commit:
- keep track of which pages were not used, and call
nfs_readdir_clear_array() on them, or
- manually link those pages into the page cache
But for now I have decided to just revert the commit, because the real
fix would require complex considerations, risking more dangerous
(crash) bugs, which may seem unsuitable for the stable branches.
SUNRPC: Fix transport accounting when caller specifies an rpc_xprt
Ensure that we do the required accounting for the round robin queue
when the caller to rpc_init_task() has passed in a transport to be
used.
Reported-by: Olga Kornievskaia <aglo@umich.edu> Reported-by: Neil Brown <neilb@suse.com> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Merge tag 'nfs-rdma-for-5.3-1' of git://git.linux-nfs.org/projects/anna/linux-nfs
NFSoRDMA client updates for 5.3
New features:
- Add a way to place MRs back on the free list
- Reduce context switching
- Add new trace events
Bugfixes and cleanups:
- Fix a BUG when tracing is enabled with NFSv4.1
- Fix a use-after-free in rpcrdma_post_recvs
- Replace use of xdr_stream_pos in rpcrdma_marshal_req
- Fix occasional transport deadlock
- Fix show_nfs_errors macros, other tracing improvements
- Remove RPCRDMA_REQ_F_PENDING and fr_state
- Various simplifications and refactors
Chuck Lever [Wed, 19 Jun 2019 14:33:53 +0000 (10:33 -0400)]
NFS: Fix show_nfs_errors macros again
I noticed that NFS status values stopped working again.
trace_print_symbols_seq() takes an unsigned long. Passing a negative
errno or negative NFSERR value just confuses it, and since we're
using C macros here and not static inline functions, all bets are
off due to implicit type conversion.
Straight-line the calling conventions so that error codes are stored
in the trace record as positive values in an unsigned long field,
mapped to symbolic as an unsigned long, and displayed as a negative
value, to continue to enable grepping on "error=-".
It's often the case that an error value that is positive is a byte
count but when it's negative, it's an error (e.g. nfs4_write). Fix
those cases so that the value that is eventually stored in the
error field is a positive NFS status or errno, or zero.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever [Wed, 19 Jun 2019 14:33:42 +0000 (10:33 -0400)]
xprtrdma: Modernize ops->connect
Adapt and apply changes that were made to the TCP socket connect
code. See the following commits for details on the purpose of
these changes:
Commit 7196dbb02ea0 ("SUNRPC: Allow changing of the TCP timeout parameters on the fly")
Commit 3851f1cdb2b8 ("SUNRPC: Limit the reconnect backoff timer to the max RPC message timeout")
Commit 02910177aede ("SUNRPC: Fix reconnection timeouts")
Some common transport code is moved to xprt.c to satisfy the code
duplication police.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever [Wed, 19 Jun 2019 14:33:36 +0000 (10:33 -0400)]
xprtrdma: Remove rpcrdma_req::rl_buffer
Clean up.
There is only one remaining function, rpcrdma_buffer_put(), that
uses this field. Its caller can supply a pointer to the correct
rpcrdma_buffer, enabling the removal of an 8-byte pointer field
from a frequently-allocated shared data structure.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever [Wed, 19 Jun 2019 14:33:31 +0000 (10:33 -0400)]
xprtrdma: Refactor chunk encoding
Clean up.
Move the "not present" case into the individual chunk encoders. This
improves code organization and readability.
The reason for the original organization was to optimize for the
case where there there are no chunks. The optimization turned out to
be inconsequential, so let's err on the side of code readability.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever [Wed, 19 Jun 2019 14:33:26 +0000 (10:33 -0400)]
xprtrdma: Streamline rpcrdma_post_recvs
rb_lock is contended between rpcrdma_buffer_create,
rpcrdma_buffer_put, and rpcrdma_post_recvs.
Commit e340c2d6ef2a ("xprtrdma: Reduce the doorbell rate (Receive)")
causes rpcrdma_post_recvs to take the rb_lock repeatedly when it
determines more Receives are needed. Streamline this code path so
it takes the lock just once in most cases to build the Receive
chain that is about to be posted.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever [Wed, 19 Jun 2019 14:33:20 +0000 (10:33 -0400)]
xprtrdma: Simplify rpcrdma_rep_create
Clean up.
Commit 7c8d9e7c8863 ("xprtrdma: Move Receive posting to Receive
handler") reduced the number of rpcrdma_rep_create call sites to
one. After that commit, the backchannel code no longer invokes it.
Therefore the free list logic added by commit d698c4a02ee0
("xprtrdma: Fix backchannel allocation of extra rpcrdma_reps") is
no longer necessary, and in fact adds some extra overhead that we
can do without.
Simply post any newly created reps. They will get added back to
the rb_recv_bufs list when they subsequently complete.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever [Wed, 19 Jun 2019 14:33:10 +0000 (10:33 -0400)]
xprtrdma: Reduce context switching due to Local Invalidation
Since commit ba69cd122ece ("xprtrdma: Remove support for FMR memory
registration"), FRWR is the only supported memory registration mode.
We can take advantage of the asynchronous nature of FRWR's LOCAL_INV
Work Requests to get rid of the completion wait by having the
LOCAL_INV completion handler take care of DMA unmapping MRs and
waking the upper layer RPC waiter.
This eliminates two context switches when local invalidation is
necessary. As a side benefit, we will no longer need the per-xprt
deferred completion work queue.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever [Wed, 19 Jun 2019 14:33:04 +0000 (10:33 -0400)]
xprtrdma: Add mechanism to place MRs back on the free list
When a marshal operation fails, any MRs that were already set up for
that request are recycled. Recycling releases MRs and creates new
ones, which is expensive.
Since commit f2877623082b ("xprtrdma: Chain Send to FastReg WRs")
was merged, recycling FRWRs is unnecessary. This is because before
that commit, frwr_map had already posted FAST_REG Work Requests,
so ownership of the MRs had already been passed to the NIC and thus
dealing with them had to be delayed until they completed.
Since that commit, however, FAST_REG WRs are posted at the same time
as the Send WR. This means that if marshaling fails, we are certain
the MRs are safe to simply unmap and place back on the free list
because neither the Send nor the FAST_REG WRs have been posted yet.
The kernel still has ownership of the MRs at this point.
This reduces the total number of MRs that the xprt has to create
under heavy workloads and makes the marshaling logic less brittle.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever [Wed, 19 Jun 2019 14:32:59 +0000 (10:32 -0400)]
xprtrdma: Remove fr_state
Now that both the Send and Receive completions are handled in
process context, it is safe to DMA unmap and return MRs to the
free or recycle lists directly in the completion handlers.
Doing this means rpcrdma_frwr no longer needs to track the state of
each MR, meaning that a VALID or FLUSHED MR can no longer appear on
an xprt's MR free list. Thus there is no longer a need to track the
MR's registration state in rpcrdma_frwr.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever [Wed, 19 Jun 2019 14:32:54 +0000 (10:32 -0400)]
xprtrdma: Remove the RPCRDMA_REQ_F_PENDING flag
Commit 9590d083c1bb ("xprtrdma: Use xprt_pin_rqst in
rpcrdma_reply_handler") pins incoming RPC/RDMA replies so they
can be left in the pending requests queue while they are being
processed without introducing a race between ->buf_free and the
transport's reply handler. Therefore RPCRDMA_REQ_F_PENDING is no
longer necessary.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever [Wed, 19 Jun 2019 14:32:48 +0000 (10:32 -0400)]
xprtrdma: Fix occasional transport deadlock
Under high I/O workloads, I've noticed that an RPC/RDMA transport
occasionally deadlocks (IOPS goes to zero, and doesn't recover).
Diagnosis shows that the sendctx queue is empty, but when sendctxs
are returned to the queue, the xprt_write_space wake-up never
occurs. The wake-up logic in rpcrdma_sendctx_put_locked is racy.
I noticed that both EMPTY_SCQ and XPRT_WRITE_SPACE are implemented
via an atomic bit. Just one of those is sufficient. Removing
EMPTY_SCQ in favor of the generic bit mechanism makes the deadlock
un-reproducible.
Without EMPTY_SCQ, rpcrdma_buffer::rb_flags is no longer used and
is therefore removed.
Unfortunately this patch does not apply cleanly to stable. If
needed, someone will have to port it and test it.
Fixes: 2fad659209d5 ("xprtrdma: Wait on empty sendctx queue") Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Chuck Lever [Wed, 19 Jun 2019 14:32:43 +0000 (10:32 -0400)]
xprtrdma: Replace use of xdr_stream_pos in rpcrdma_marshal_req
This is a latent bug. xdr_stream_pos works by subtracting
xdr_stream::nwords from xdr_buf::len. But xdr_stream::nwords is not
initialized by xdr_init_encode().
It works today only because all fields in rpcrdma_req::rl_stream
are initialized to zero by rpcrdma_req_create, making the
subtraction in xdr_stream_pos always a no-op.
I found this issue via code inspection. It was introduced by commit 39f4cd9e9982 ("xprtrdma: Harden chunk list encoding against send
buffer overflow"), but the code has changed enough since then that
this fix can't be automatically applied to stable.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Dave Wysochanski [Wed, 26 Jun 2019 20:30:24 +0000 (16:30 -0400)]
SUNRPC: Fix possible autodisconnect during connect due to old last_used
Ensure last_used is updated before calling mod_timer inside
xprt_schedule_autodisconnect. This avoids a possible xprt_autoclose
firing immediately after a successful connect when xprt_unlock_connect
calls xprt_schedule_autodisconnect with an old value of last_used.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Darrick J. Wong [Fri, 31 May 2019 21:06:05 +0000 (14:06 -0700)]
nfs: disable client side deduplication
The NFS protocol doesn't support deduplication, so turn it off again.
Fixes: ce96e888fe48e ("Fix nfs4.2 return -EINVAL when do dedupe operation") Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Dave Wysochanski [Fri, 17 May 2019 21:06:41 +0000 (17:06 -0400)]
NFSv4: Add lease_time and lease_expired to 'nfs4:' line of mountstats
On the NFS client there is no low-impact way to determine the nfs4
lease time or whether the lease is expired, so add these to mountstats
with times displayed in seconds.
If the lease is not expired, display lease_expired=0. Otherwise,
display lease_expired=seconds_since_expired, similar to 'age:' line
in mountstats.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Trond Myklebust [Tue, 11 Jun 2019 20:49:52 +0000 (16:49 -0400)]
NFS: Clean up writeback code
Now that the VM promises never to recurse back into the filesystem
layer on writeback, remove all the GFP_NOFS references etc from
the generic writeback code.
NeilBrown [Thu, 30 May 2019 00:41:28 +0000 (10:41 +1000)]
SUNRPC: add links for all client xprts to debugfs
Now that a client can have multiple xprts, we need to add
them all to debugs.
The first one is still "xprt"
Subsequent xprts are "xprt1", "xprt2", etc.
Dave Wysochanski [Thu, 23 May 2019 20:13:50 +0000 (16:13 -0400)]
SUNRPC: Count ops completing with tk_status < 0
We often see various error conditions with NFS4.x that show up with
a very high operation count all completing with tk_status < 0 in a
short period of time. Add a count to rpc_iostats to record on a
per-op basis the ops that complete in this manner, which will
enable lower overhead diagnostics.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Dave Wysochanski [Thu, 23 May 2019 20:13:48 +0000 (16:13 -0400)]
SUNRPC: Move call to rpc_count_iostats before rpc_call_done
For diagnostic purposes, it would be useful to have an rpc_iostats
metric of RPCs completing with tk_status < 0. Unfortunately,
tk_status is reset inside the rpc_call_done functions for each
operation, and the call to tally the per-op metrics comes after
rpc_call_done. Refactor the call to rpc_count_iostat earlier in
rpc_exit_task so we can count these RPCs completing in error.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
NeilBrown [Thu, 30 May 2019 00:41:28 +0000 (10:41 +1000)]
NFS: send state management on a single connection.
With NFSv4.1, different network connections need to be explicitly
bound to a session. During session startup, this is not possible
so only a single connection must be used for session startup.
So add a task flag to disable the default round-robin choice of
connections (when nconnect > 1) and force the use of a single
connection.
Then use that flag on all requests for session management - for
consistence, include NFSv4.0 management (SETCLIENTID) and session
destruction
If the user specifies -onconnect=<number> mount option, and the transport
protocol is TCP, then set up <number> connections to the pNFS data server
as well. The connections will all go to the same IP address.
NFSv4: Allow multiple connections to NFSv4.x (x>0) servers
If the user specifies the -onconn=<number> mount option, and the transport
protocol is TCP, then set up <number> connections to the server. The
connections will all go to the same IP address.
NFS: Add a mount option to specify number of TCP connections to use
Allow the user to specify that the client should use multiple connections
to the server. For the moment, this functionality will be limited to
TCP and to NFSv4.x (x>0).
Trond Myklebust [Wed, 30 Jan 2019 02:40:10 +0000 (21:40 -0500)]
NFS: Add sysfs support for per-container identifier
In order to identify containers to the NFS client, we add a per-net
sysfs attribute that udev can fill with the appropriate identifier.
The identifier could be a unique hostname, but in most cases it
will probably be a persisted uuid.
Trond Myklebust [Wed, 22 May 2019 12:38:57 +0000 (08:38 -0400)]
NFS: Add deferred cache invalidation for close-to-open consistency violations
If the client detects that close-to-open cache consistency has been
violated, and that the file or directory has been changed on the
server, then do a cache invalidation when we're done working with
the file.
The reason we don't do an immediate cache invalidation is that we
want to avoid performance problems due to false positives. Also,
note that we cannot guarantee cache consistency in this situation
even if we do invalidate the cache.
Trond Myklebust [Wed, 1 May 2019 14:49:27 +0000 (10:49 -0400)]
SUNRPC: Replace the queue timer with a delayed work function
The queue timer function, which walks the RPC queue in order to locate
candidates for waking up is one of the current constraints against
removing the bh-safe queue spin locks. Replace it with a delayed
work queue, so that we can do the actual rpc task wake ups from an
ordinary process context.
Trond Myklebust [Thu, 27 Jun 2019 10:41:45 +0000 (06:41 -0400)]
NFSv4: Handle the special Linux file open access mode
According to the open() manpage, Linux reserves the access mode 3
to mean "check for read and write permission on the file and return
a file descriptor that can't be used for reading or writing."
Currently, the NFSv4 code will ask the server to open the file,
and will use an incorrect share access mode of 0. Since it has
an incorrect share access mode, the client later forgets to send
a corresponding close, meaning it can leak stateids on the server.
Trond Myklebust [Thu, 27 Jun 2019 10:30:48 +0000 (06:30 -0400)]
NFSv4: Handle open for execute correctly
When mapping the NFSv4 context to an open mode and access mode,
we need to treat the FMODE_EXEC flag differently. For the open
mode, FMODE_EXEC means we need read share access. For the access
mode checking, we need to verify that the user actually has
execute access.
Merge tag 'mips_fixes_5.2_2' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux
Pull MIPS fixes from Paul Burton:
"A few more MIPS fixes:
- Fix a silly typo in virt_addr_valid which led to completely bogus
behavior (that happened to stop tripping up hardened usercopy
despite being broken).
- Fix UART parity setup on AR933x systems.
- A build fix for non-Linux build machines.
- Have the 'all' make target build DTBs, primarily to fit in with the
behavior of scripts/package/builddeb.
- Handle an execution hazard in TLB exceptions that use KScratch
registers, which could inadvertently clobber the $1 register on
some generally higher-end out-of-order CPUs.
- A MAINTAINERS update to fix the path to the NAND driver for Ingenic
systems"
* tag 'mips_fixes_5.2_2' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:
MAINTAINERS: Correct path to moved files
MIPS: Add missing EHB in mtc0 -> mfc0 sequence.
MIPS: have "plain" make calls build dtbs for selected platforms
MIPS: fix build on non-linux hosts
MIPS: ath79: fix ar933x uart parity mode
MIPS: Fix bounds check virt_addr_valid
Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull SCSI fixes from James Bottomley:
"Two iscsi fixes.
One for an oops in the client which can be triggered by the server
authentication protocol and the other in the target code which causes
data corruption"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: iscsi: set auth_protocol back to NULL if CHAP_A value is not supported
scsi: target/iblock: Fix overrun in WRITE SAME emulation
Merge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Pull vfs fixlet from Al Viro:
"Fix bogus default y in Kconfig (VALIDATE_FS_PARSER)
That thing should not be turned on by default, especially since it's
not quiet in case it finds no problems. Geert has sent the obvious fix
quite a few times, but it fell through the cracks"
* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
fs: VALIDATE_FS_PARSER should default to n
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull kvm fixes from Paolo Bonzini:
"x86 bugfix patches and one compilation fix for ARM"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
KVM: arm64/sve: Fix vq_present() macro to yield a bool
KVM: LAPIC: Fix pending interrupt in IRR blocked by software disable LAPIC
KVM: nVMX: Change KVM_STATE_NESTED_EVMCS to signal vmcs12 is copied from eVMCS
KVM: nVMX: Allow restore nested-state to enable eVMCS when vCPU in SMM
KVM: x86: degrade WARN to pr_warn_ratelimited
Merge tag 'mtd/fixes-for-5.2-final' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux
Pull mtf fixes from Miquel Raynal:
- Fix the memory organization structure of a Macronix SPI-NAND chip.
- Fix a build dependency wrongly described.
- Fix the sunxi NAND driver for A23/A33 SoCs by (a) reverting the
faulty commit introducing broken DMA support and (b) applying another
commit bringing working DMA support.
* tag 'mtd/fixes-for-5.2-final' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux:
mtd: rawnand: sunxi: Add A23/A33 DMA support with extra MBUS configuration
Revert "mtd: rawnand: sunxi: Add A23/A33 DMA support"
mtd: rawnand: ingenic: Fix ingenic_ecc dependency
mtd: spinand: Fix max_bad_eraseblocks_per_lun info in memorg
Merge tag 'nfsd-5.2-2' of git://linux-nfs.org/~bfields/linux
Pull nfsd fixes from Bruce Fields:
"Two more quick bugfixes for nfsd: fixing a regression causing mount
failures on high-memory machines and fixing the DRC over RDMA"
* tag 'nfsd-5.2-2' of git://linux-nfs.org/~bfields/linux:
nfsd: Fix overflow causing non-working mounts on 1 TB machines
svcrdma: Ignore source port when computing DRC hash
mtd: rawnand: sunxi: Add A23/A33 DMA support with extra MBUS configuration
Allwinner NAND controllers can make use of DMA to enhance the I/O
throughput thanks to ECC pipelining. DMA handling with A23/A33 NAND IP
is a bit different than with the older SoCs, hence the introduction of
a new compatible to handle:
* the differences between register offsets,
* the burst length change from 4 to minimum 8,
* manage SRAM accesses through MBUS with extra configuration.
Dmitry Osipenko [Sun, 23 Jun 2019 17:46:55 +0000 (20:46 +0300)]
i2c: tegra: Add Dmitry as a reviewer
I'm contributing to Tegra's upstream development in general and happened
to review the Tegra's I2C patches for awhile because I'm actively using
upstream kernel on all of my Tegra-powered devices and initially some of
the submitted patches were getting my attention since they were causing
problems. Recently Wolfram Sang asked whether I'm interested in becoming
a reviewer for the driver and I don't mind at all.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
[wsa: ack was expressed by Thierry Reding in a mail thread] Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
CONFIG_VALIDATE_FS_PARSER is a debugging tool to check that the parser
tables are vaguely sane. It was set to default to 'Y' for the moment to
catch errors in upcoming fs conversion development.
Make sure it is not enabled by default in the final release of v5.1.
Zhang Lei [Wed, 3 Jul 2019 17:42:50 +0000 (18:42 +0100)]
KVM: arm64/sve: Fix vq_present() macro to yield a bool
The original implementation of vq_present() relied on aggressive
inlining in order for the compiler to know that the code is
correct, due to some const-casting issues. This was causing sparse
and clang to complain, while GCC compiled cleanly.
Commit 0c529ff789bc addressed this problem, but since vq_present()
is no longer a function, there is now no implicit casting of the
returned value to the return type (bool).
In set_sve_vls(), this uncast bit value is compared against a bool,
and so may spuriously compare as unequal when both are nonzero. As
a result, KVM may reject valid SVE vector length configurations as
invalid, and vice versa.
Fix it by forcing the returned value to a bool.
Signed-off-by: Zhang Lei <zhang.lei@jp.fujitsu.com> Fixes: 0c529ff789bc ("KVM: arm64: Implement vq_present() as a macro") Signed-off-by: Dave Martin <Dave.Martin@arm.com> [commit message rewrite] Cc: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
One space is left unused in circular FIFO to differentiate
'full' and 'empty' cases. So take that in to account while
counting for the descriptors completed.
Fixes the issue reported here,
https://lkml.org/lkml/2019/6/18/669
Robin Gong [Fri, 21 Jun 2019 08:23:06 +0000 (16:23 +0800)]
dmaengine: imx-sdma: remove BD_INTR for channel0
It is possible for an irq triggered by channel0 to be received later
after clks are disabled once firmware loaded during sdma probe. If
that happens then clearing them by writing to SDMA_H_INTR won't work
and the kernel will hang processing infinite interrupts. Actually,
don't need interrupt triggered on channel0 since it's pollling
SDMA_H_STATSTOP to know channel0 done rather than interrupt in
current code, just clear BD_INTR to disable channel0 interrupt to
avoid the above case.
This issue was brought by commit 1d069bfa3c78 ("dmaengine: imx-sdma:
ack channel 0 IRQ in the interrupt handler") which didn't take care
the above case.
Fixes: 1d069bfa3c78 ("dmaengine: imx-sdma: ack channel 0 IRQ in the interrupt handler") Cc: stable@vger.kernel.org #5.0+ Signed-off-by: Robin Gong <yibin.gong@nxp.com> Reported-by: Sven Van Asbroeck <thesven73@gmail.com> Tested-by: Sven Van Asbroeck <thesven73@gmail.com> Reviewed-by: Michael Olbrich <m.olbrich@pengutronix.de> Signed-off-by: Vinod Koul <vkoul@kernel.org>
dmaengine: imx-sdma: fix use-after-free on probe error path
If probe() fails anywhere beyond the point where
sdma_get_firmware() is called, then a kernel oops may occur.
Problematic sequence of events:
1. probe() calls sdma_get_firmware(), which schedules the
firmware callback to run when firmware becomes available,
using the sdma instance structure as the context
2. probe() encounters an error, which deallocates the
sdma instance structure
3. firmware becomes available, firmware callback is
called with deallocated sdma instance structure
4. use after free - kernel oops !
Solution: only attempt to load firmware when we're certain
that probe() will succeed. This guarantees that the firmware
callback's context will remain valid.
Note that the remove() path is unaffected by this issue: the
firmware loader will increment the driver module's use count,
ensuring that the module cannot be unloaded while the
firmware callback is pending or running.
Signed-off-by: Sven Van Asbroeck <TheSven73@gmail.com> Reviewed-by: Robin Gong <yibin.gong@nxp.com>
[vkoul: fixed braces for if condition] Signed-off-by: Vinod Koul <vkoul@kernel.org>
Dan Carpenter [Mon, 24 Jun 2019 13:49:40 +0000 (16:49 +0300)]
dmaengine: jz4780: Fix an endian bug in IRQ handler
The "pending" variable was a u32 but we cast it to an unsigned long
pointer when we do the for_each_set_bit() loop. The problem is that on
big endian 64bit systems that results in an out of bounds read.
Fixes: 4e4106f5e942 ("dmaengine: jz4780: Fix transfers being ACKed too soon") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Vinod Koul <vkoul@kernel.org>
Merge tag 'drm-fixes-2019-07-05-1' of git://anongit.freedesktop.org/drm/drm
Pull drm fixes from Dave Airlie:
"I skipped last week because there wasn't much worth doing, this week
got a few more fixes in.
amdgpu:
- default register value change
- runpm regression fix
- fan control fix
i915:
- fix Ironlake regression
panfrost:
- fix a double free
virtio:
- fix a locking bug
imx:
- crtc disable fixes"
* tag 'drm-fixes-2019-07-05-1' of git://anongit.freedesktop.org/drm/drm:
drm/imx: only send event on crtc disable if kept disabled
drm/imx: notify drm core before sending event during crtc disable
drm/i915/ringbuffer: EMIT_INVALIDATE *before* switch context
drm/amdgpu/gfx9: use reset default for PA_SC_FIFO_SIZE
drm/amdgpu: Don't skip display settings in hwmgr_resume()
drm/amd/powerplay: use hardware fan control if no powerplay fan table
drm/panfrost: Fix a double-free error
drm/etnaviv: add missing failure path to destroy suballoc
drm/virtio: move drm_connector_update_edid_property() call
Dave Airlie [Fri, 5 Jul 2019 02:54:48 +0000 (12:54 +1000)]
Merge tag 'imx-drm-fixes-2019-07-04' of git://git.pengutronix.de/git/pza/linux into drm-fixes
drm/imx: fix stale vblank timestamp after a modeset
This series fixes stale vblank timestamps in the first event sent after
a crtc was disabled. The core now is notified via drm_crtc_vblank_off
before sending the last pending event in atomic_disable. If the crtc is
reenabled right away during to a modeset, the event is not sent at all,
as the next vblank will take care of it.
Merge tag 'dax-fix-5.2-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm
Pull dax fix from Dan Williams:
"A single dax fix that has been soaking awaiting other fixes under
discussion to join it. As it is getting late in the cycle lets proceed
with this fix and save follow-on changes for post-v5.3-rc1.
- Fix xarray entry association for mixed mappings"
* tag 'dax-fix-5.2-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
dax: Fix xarray entry association for mixed mappings
swap_readpage(): avoid blk_wake_io_task() if !synchronous
swap_readpage() sets waiter = bio->bi_private even if synchronous = F,
this means that the caller can get the spurious wakeup after return.
This can be fatal if blk_wake_io_task() does
set_current_state(TASK_RUNNING) after the caller does
set_special_state(), in the worst case the kernel can crash in
do_task_dead().
devm_ioremap_resource() does not currently take 'const' arguments, which
results in a warning from the first driver trying to do it anyway:
drivers/gpio/gpio-amd-fch.c: In function 'amd_fch_gpio_probe':
drivers/gpio/gpio-amd-fch.c:171:49: error: passing argument 2 of 'devm_ioremap_resource' discards 'const' qualifier from pointer target type [-Werror=discarded-qualifiers]
priv->base = devm_ioremap_resource(&pdev->dev, &amd_fch_gpio_iores);
^~~~~~~~~~~~~~~~~~~
Change the prototype to allow it, as there is no real reason not to.
In production we have noticed hard lockups on large machines running
large jobs due to kswaps hoarding lru lock within isolate_lru_pages when
sc->reclaim_idx is 0 which is a small zone. The lru was couple hundred
GiBs and the condition (page_zonenum(page) > sc->reclaim_idx) in
isolate_lru_pages() was basically skipping GiBs of pages while holding
the LRU spinlock with interrupt disabled.
On further inspection, it seems like there are two issues:
(1) If kswapd on the return from balance_pgdat() could not sleep (i.e.
node is still unbalanced), the classzone_idx is unintentionally set
to 0 and the whole reclaim cycle of kswapd will try to reclaim only
the lowest and smallest zone while traversing the whole memory.
(2) Fundamentally isolate_lru_pages() is really bad when the
allocation has woken kswapd for a smaller zone on a very large machine
running very large jobs. It can hoard the LRU spinlock while skipping
over 100s of GiBs of pages.
This patch only fixes (1). (2) needs a more fundamental solution. To
fix (1), in the kswapd context, if pgdat->kswapd_classzone_idx is
invalid use the classzone_idx of the previous kswapd loop otherwise use
the one the waker has requested.
Link: http://lkml.kernel.org/r/20190701201847.251028-1-shakeelb@google.com Fixes: e716f2eb24de ("mm, vmscan: prevent kswapd sleeping prematurely due to mismatched classzone_idx") Signed-off-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Yang Shi <yang.shi@linux.alibaba.com> Acked-by: Mel Gorman <mgorman@techsingularity.net> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Hillf Danton <hdanton@sina.com> Cc: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Eric Biggers [Thu, 4 Jul 2019 22:14:39 +0000 (15:14 -0700)]
fs/userfaultfd.c: disable irqs for fault_pending and event locks
When IOCB_CMD_POLL is used on a userfaultfd, aio_poll() disables IRQs
and takes kioctx::ctx_lock, then userfaultfd_ctx::fd_wqh.lock.
This may have to wait for userfaultfd_ctx::fd_wqh.lock to be released by
userfaultfd_ctx_read(), which in turn can be waiting for
userfaultfd_ctx::fault_pending_wqh.lock or
userfaultfd_ctx::event_wqh.lock.
But elsewhere the fault_pending_wqh and event_wqh locks are taken with
IRQs enabled. Since the IRQ handler may take kioctx::ctx_lock, lockdep
reports that a deadlock is possible.
Fix it by always disabling IRQs when taking the fault_pending_wqh and
event_wqh locks.
Commit ae62c16e105a ("userfaultfd: disable irqs when taking the
waitqueue lock") didn't fix this because it only accounted for the
fd_wqh lock, not the other locks nested inside it.
Link: http://lkml.kernel.org/r/20190627075004.21259-1-ebiggers@kernel.org Fixes: bfe4037e722e ("aio: implement IOCB_CMD_POLL") Signed-off-by: Eric Biggers <ebiggers@google.com> Reported-by: syzbot+fab6de82892b6b9c6191@syzkaller.appspotmail.com Reported-by: syzbot+53c0b767f7ca0dc0c451@syzkaller.appspotmail.com Reported-by: syzbot+a3accb352f9c22041cfa@syzkaller.appspotmail.com Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Christoph Hellwig <hch@lst.de> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: <stable@vger.kernel.org> [4.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/page_alloc.c: fix regression with deferred struct page init
Commit 0e56acae4b4d ("mm: initialize MAX_ORDER_NR_PAGES at a time
instead of doing larger sections") is causing a regression on some
systems when the kernel is booted as Xen dom0.
The system will just hang in early boot.
Reason is an endless loop in get_page_from_freelist() in case the first
zone looked at has no free memory. deferred_grow_zone() is always
returning true due to the following code snipplet:
/* If the zone is empty somebody else may have cleared out the zone */
if (!deferred_init_mem_pfn_range_in_zone(&i, zone, &spfn, &epfn,
first_deferred_pfn)) {
pgdat->first_deferred_pfn = ULONG_MAX;
pgdat_resize_unlock(pgdat, &flags);
return true;
}
This in turn results in the loop as get_page_from_freelist() is assuming
forward progress can be made by doing some more struct page
initialization.
Link: http://lkml.kernel.org/r/20190620160821.4210-1-jgross@suse.com Fixes: 0e56acae4b4d ("mm: initialize MAX_ORDER_NR_PAGES at a time instead of doing larger sections") Signed-off-by: Juergen Gross <jgross@suse.com> Suggested-by: Alexander Duyck <alexander.h.duyck@linux.intel.com> Acked-by: Alexander Duyck <alexander.h.duyck@linux.intel.com> Cc: Michal Hocko <mhocko@suse.com> Cc: Pavel Tatashin <pasha.tatashin@soleen.com> Cc: Mike Rapoport <rppt@linux.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>