]> git.apps.os.sepia.ceph.com Git - ceph-client.git/log
ceph-client.git
3 years ago[DO NOT MERGE] libceph: pr_err when we get a socket write error that's not -EAGAIN wip-jlayton-55823
Jeff Layton [Fri, 3 Jun 2022 16:59:56 +0000 (12:59 -0400)]
[DO NOT MERGE] libceph: pr_err when we get a socket write error that's not -EAGAIN

The complex error_msg stuff makes it difficult to track what error was
returned. For now, when we get a socket write error, throw a pr_err
immediately as well.

URL: https://tracker.ceph.com/issues/55823
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agoceph: make ceph_msdc_build_path use ref-walk
Jeff Layton [Wed, 5 Aug 2020 18:50:48 +0000 (14:50 -0400)]
ceph: make ceph_msdc_build_path use ref-walk

Encryption potentially requires allocation, at which point we'll need to
be in a non-atomic context. Convert ceph_msdc_build_path to take dentry
spinlocks and references instead of using rcu_read_lock to walk the
path.

This is slightly less efficient, and we may want to eventually allow
using RCU when the leaf dentry isn't encrypted.

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agoceph: preallocate inode for ops that may create one
Jeff Layton [Wed, 26 Aug 2020 17:11:00 +0000 (13:11 -0400)]
ceph: preallocate inode for ops that may create one

When creating a new inode, we need to determine the crypto context
before we can transmit the RPC. The fscrypt API has a routine for getting
a crypto context before a create occurs, but it requires an inode.

Change the ceph code to preallocate an inode in advance of a create of
any sort (open(), mknod(), symlink(), etc). Move the existing code that
generates the ACL and SELinux blobs into this routine since that's
mostly common across all the different codepaths.

In most cases, we just want to allow ceph_fill_trace to use that inode
after the reply comes in, so add a new field to the MDS request for it
(r_new_inode).

The async create codepath is a bit different though. In that case, we
want to hash the inode in advance of the RPC so that it can be used
before the reply comes in. If the call subsequently fails with
-EJUKEBOX, then just put the references and clean up the as_ctx. Note
that with this change, we now need to regenerate the as_ctx when this
occurs, but it's quite rare for it to happen.

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agoceph: add new mount option to enable sparse reads
Jeff Layton [Sat, 26 Feb 2022 11:33:03 +0000 (06:33 -0500)]
ceph: add new mount option to enable sparse reads

Add a new mount option that has the client issue sparse reads instead of
normal ones. The callers now preallocate an sparse extent buffer that
the libceph receive code can populate and hand back after the operation
completes.

After a successful sparse read, we can't use the req->r_result value to
determine the amount of data "read", so instead we set the received
length to be from the end of the last extent in the buffer. Any
interstitial holes will have been filled by the receive code.

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agolibceph: add sparse read support to msgr1
Jeff Layton [Thu, 24 Mar 2022 17:33:06 +0000 (13:33 -0400)]
libceph: add sparse read support to msgr1

Add 2 new fields to ceph_connection_v1_info to track the necessary info
in sparse reads. Skip initializing the cursor for a sparse read.

Break out read_partial_message_section into a wrapper around a new
read_partial_message_chunk function that doesn't zero out the crc first.

Add new helper functions to drive receiving into the destinations
provided by the sparse_read state machine.

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agolibceph: support sparse reads on msgr2 secure codepath
Jeff Layton [Wed, 23 Mar 2022 16:17:15 +0000 (12:17 -0400)]
libceph: support sparse reads on msgr2 secure codepath

Add a new init_sgs_pages helper that populates the scatterlist from
an arbitrary point in an array of pages.

Change setup_message_sgs to take an optional pointer to an array of
pages. If that's set, then the scatterlist will be set using that
array instead of the cursor.

When given a sparse read on a secure connection, decrypt the data
in-place rather than into the final destination, by passing it the
in_enc_pages array.

After decrypting, run the sparse_read state machine in a loop, copying
data from the decrypted pages until it's complete.

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agolibceph: add sparse read support to OSD client
Jeff Layton [Fri, 11 Feb 2022 16:38:02 +0000 (11:38 -0500)]
libceph: add sparse read support to OSD client

Have get_reply check for the presence of sparse read ops in the
request and set the sparse_read boolean in the msg. That will queue the
messenger layer to use the sparse read codepath instead of the normal
data receive.

Add a new sparse_read operation for the OSD client, driven by its own
state machine. The messenger will repeatedly call the sparse_read
operation, and it will pass back the necessary info to set up to read
the next extent of data, while zero-filling the sparse regions.

The state machine will stop at the end of the last extent, and will
attach the extent map buffer to the ceph_osd_req_op so that the caller
can use it.

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agolibceph: add sparse read support to msgr2 crc state machine
Jeff Layton [Tue, 25 Jan 2022 13:26:31 +0000 (08:26 -0500)]
libceph: add sparse read support to msgr2 crc state machine

Add support for a new sparse_read ceph_connection operation. The idea is
that the client driver can define this operation use it to do special
handling for incoming reads.

The alloc_msg routine will look at the request and determine whether the
reply is expected to be sparse. If it is, then we'll dispatch to a
different set of state machine states that will repeatedly call the
driver's sparse_read op to get length and placement info for reading the
extent map, and the extents themselves.

This necessitates adding some new field to some other structs:

- The msg gets a new bool to track whether it's a sparse_read request.

- A new field is added to the cursor to track the amount remaining in the
current extent. This is used to cap the read from the socket into the
msg_data

- Handing a revoke with all of this is particularly difficult, so I've
added a new data_len_remain field to the v2 connection info, and then
use that to skip that much on a revoke. We may want to expand the use of
that to the normal read path as well, just for consistency's sake.

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agolibceph: define struct ceph_sparse_extent and add some helpers
Jeff Layton [Wed, 16 Mar 2022 19:23:00 +0000 (15:23 -0400)]
libceph: define struct ceph_sparse_extent and add some helpers

When the OSD sends back a sparse read reply, it contains an array of
these structures. Define the structure and add a couple of helpers for
dealing with them.

Also add a place in struct ceph_osd_req_op to store the extent buffer,
and code to free it if it's populated when the req is torn down.

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agolibceph: add spinlock around osd->o_requests
Jeff Layton [Mon, 14 Mar 2022 19:47:18 +0000 (15:47 -0400)]
libceph: add spinlock around osd->o_requests

In a later patch, we're going to need to search for a request in
the rbtree, but taking the o_mutex is inconvenient as we already
hold the con mutex at the point where we need it.

Add a new spinlock that we take when inserting and erasing entries from
the o_requests tree. Search of the rbtree can be done with either the
mutex or the spinlock, but insertion and removal requires both.

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agoceph: choose auth MDS for getxattr with the Xs caps
Xiubo Li [Fri, 27 May 2022 04:39:17 +0000 (12:39 +0800)]
ceph: choose auth MDS for getxattr with the Xs caps

And for the 'Xs' caps for getxattr we will also choose the auth MDS,
because the MDS side code is buggy due to setxattr won't notify the
replica MDSes when the values changed and the replica MDS will return
the old values. Though we will fix it in MDS code, but this still
makes sense for old ceph.

URL: https://tracker.ceph.com/issues/55331
Signed-off-by: Xiubo Li <xiubli@redhat.com>
3 years agoceph: add session already open notify support
Xiubo Li [Thu, 26 May 2022 05:21:31 +0000 (13:21 +0800)]
ceph: add session already open notify support

If the connection was accidently closed due to the socket issue or
something else the clients will try to open the opened sessions, the
MDSes will send the session open reply one more time if the clients
support the notify feature.

When the clients retry to open the sessions the s_seq will be 0 as
default, we need to update it anyway.

URL: https://tracker.ceph.com/issues/53911
Signed-off-by: Xiubo Li <xiubli@redhat.com>
3 years agolibceph: drop last_piece flag from ceph_msg_data_cursor
Jeff Layton [Wed, 25 May 2022 10:11:00 +0000 (06:11 -0400)]
libceph: drop last_piece flag from ceph_msg_data_cursor

ceph_msg_data_next is always passed a NULL pointer for this field. Some
of the "next" operations look at it in order to determine the length,
but we can just take the min of the data on the page or cursor->resid.

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agoceph: wait the first reply of inflight async unlink
Xiubo Li [Tue, 10 May 2022 01:47:01 +0000 (09:47 +0800)]
ceph: wait the first reply of inflight async unlink

In async unlink case the kclient won't wait for the first reply
from MDS and just drop all the links and unhash the dentry and then
succeeds immediately.

For any new create/link/rename,etc requests followed by using the
same file names we must wait for the first reply of the inflight
unlink request, or the MDS possibly will fail these following
requests with -EEXIST if the inflight async unlink request was
delayed for some reasons.

And the worst case is that for the none async openc request it will
successfully open the file if the CDentry hasn't been unlinked yet,
but later the previous delayed async unlink request will remove the
CDenty. That means the just created file is possiblly deleted later
by accident.

We need to wait for the inflight async unlink requests to finish
when creating new files/directories by using the same file names.

[ xiubli: fixed the build warning reported by kernel test rotot]

URL: https://tracker.ceph.com/issues/55332
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Xiubo Li <xiubli@redhat.com>
3 years agofs/dcache: export d_same_name() helper
Xiubo Li [Mon, 16 May 2022 03:23:19 +0000 (11:23 +0800)]
fs/dcache: export d_same_name() helper

Compare dentry name with case-exact name, return true if names
are same, or false.

Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Xiubo Li <xiubli@redhat.com>
3 years agoceph: remove useless CEPHFS_FEATURES_CLIENT_REQUIRED
Xiubo Li [Thu, 26 May 2022 06:07:21 +0000 (14:07 +0800)]
ceph: remove useless CEPHFS_FEATURES_CLIENT_REQUIRED

This macro was added but never be used. And check the ceph code
there has another CEPHFS_FEATURES_MDS_REQUIRED but always be empty.

We should clean up all this related code, which make no sense but
introducing confusion.

Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Luís Henriques <lhenriques@suse.de>
Signed-off-by: Xiubo Li <xiubli@redhat.com>
3 years agoceph: use correct index when encoding client supported features
Luís Henriques [Tue, 24 May 2022 16:06:27 +0000 (17:06 +0100)]
ceph: use correct index when encoding client supported features

Feature bits have to be encoded into the correct locations.  This hasn't
been an issue so far because the only hole in the feature bits was in bit
10 (CEPHFS_FEATURE_RECLAIM_CLIENT), which is located in the 2nd byte.  When
adding more bits that go beyond the this 2nd byte, the bug will show up.

[xiubli: remove the incorrect comment for CEPHFS_FEATURES_CLIENT_SUPPORTED]

Fixes: 9ba1e224538a ("ceph: allocate the correct amount of extra bytes for the session features")
Signed-off-by: Luís Henriques <lhenriques@suse.de>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Xiubo Li <xiubli@redhat.com>
3 years ago[DO NOT MERGE] mm: BUG if filemap_alloc_folio gives us a folio with a non-NULL -...
Jeff Layton [Fri, 13 May 2022 14:23:25 +0000 (10:23 -0400)]
[DO NOT MERGE] mm: BUG if filemap_alloc_folio gives us a folio with a non-NULL ->private

We've seen some instances where we call __filemap_get_folio and get back
one with a ->private value that is non-NULL. Let's have the allocator
bug if that happens.

For now, let's just put this into the testing kernel. We can let Willy
decide if he wants it in mainline.

URL: https://tracker.ceph.com/issues/55421
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Xiubo Li <xiubli@redhat.com>
Cc: Luís Henriques <lhenriques@suse.de>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years ago[DO NOT MERGE] ceph: dump info about cap flushes when we're waiting too long for...
Jeff Layton [Wed, 7 Jul 2021 13:20:34 +0000 (09:20 -0400)]
[DO NOT MERGE] ceph: dump info about cap flushes when we're waiting too long for them

We've had some cases of hung umounts in teuthology testing. It looks
like client is waiting for cap flushes to complete, but they aren't.

Add a field to the inode to track the highest cap flush tid seen for
that inode. Also, add a backpointer to the inode to the ceph_cap_flush
struct.

Change wait_caps_flush to wait 60s, and then dump info about the
condition of the list.

Also, print pr_info messages if we end up dropping a FLUSH_ACK for an
inode onto the floor, or if we get a message on an unregistered
session.

Reported-by: Patrick Donnelly <pdonnell@redhat.com>
URL: https://tracker.ceph.com/issues/51279
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years ago[DO NOT MERGE] rbd: bump RBD_MAX_PARENT_CHAIN_LEN to 128
Ilya Dryomov [Sat, 20 Feb 2016 17:26:57 +0000 (18:26 +0100)]
[DO NOT MERGE] rbd: bump RBD_MAX_PARENT_CHAIN_LEN to 128

Bump RBD_MAX_PARENT_CHAIN_LEN from 16 to 128 to avoid fsx failures.

(The alternative is changing fsx to flatten unconditionally when the
limit of 16 is reached, which is ugly and not needed for librbd.)

3 years agofscrypt: add fscrypt_context_for_new_inode
Jeff Layton [Tue, 1 Sep 2020 16:56:42 +0000 (12:56 -0400)]
fscrypt: add fscrypt_context_for_new_inode

Most filesystems just call fscrypt_set_context on new inodes, which
usually causes a setxattr. That's a bit late for ceph, which can send
along a full set of attributes with the create request.

Doing so allows it to avoid race windows that where the new inode could
be seen by other clients without the crypto context attached. It also
avoids the separate round trip to the server.

Refactor the fscrypt code a bit to allow us to create a new crypto
context, attach it to the inode, and write it to the buffer, but without
calling set_context on it. ceph can later use this to marshal the
context into the attributes we send along with the create request.

Acked-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agofscrypt: export fscrypt_fname_encrypt and fscrypt_fname_encrypted_size
Jeff Layton [Fri, 8 Jan 2021 20:34:38 +0000 (15:34 -0500)]
fscrypt: export fscrypt_fname_encrypt and fscrypt_fname_encrypted_size

For ceph, we want to use our own scheme for handling filenames that are
are longer than NAME_MAX after encryption and Base64 encoding. This
allows us to have a consistent view of the encrypted filenames for
clients that don't support fscrypt and clients that do but that don't
have the key.

Currently, fs/crypto only supports encrypting filenames using
fscrypt_setup_filename, but that also handles encoding nokey names. Ceph
can't use that because it handles nokey names in a different way.

Export fscrypt_fname_encrypt. Rename fscrypt_fname_encrypted_size to
__fscrypt_fname_encrypted_size and add a new wrapper called
fscrypt_fname_encrypted_size that takes an inode argument rather than a
pointer to a fscrypt_policy union.

Acked-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
3 years agofs: change test in inode_insert5 for adding to the sb list
Jeff Layton [Thu, 31 Mar 2022 20:29:00 +0000 (16:29 -0400)]
fs: change test in inode_insert5 for adding to the sb list

inode_insert5 currently looks at I_CREATING to decide whether to insert
the inode into the sb list. This test is a bit ambiguous, as I_CREATING
state is not directly related to that list.

This test is also problematic for some upcoming ceph changes to add
fscrypt support. We need to be able to allocate an inode using new_inode
and insert it into the hash later iff we end up using it, and doing that
now means that we double add it and corrupt the list.

What we really want to know in this test is whether the inode is already
in its superblock list, and then add it if it isn't. Have it test for
list_empty instead and ensure that we always initialize the list by
doing it in inode_init_once. It's only ever removed from the list with
list_del_init, so that should be sufficient.

Suggested-by: Al Viro <viro@zeniv.linux.org.uk>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoMAINTAINERS: move myself from ceph "Maintainer" to "Reviewer" ceph-for-5.19-rc1
Jeff Layton [Mon, 23 May 2022 17:22:09 +0000 (13:22 -0400)]
MAINTAINERS: move myself from ceph "Maintainer" to "Reviewer"

Xiubo has graciously volunteered to take over for me as the Linux cephfs
client maintainer. Make it official by changing myself to be a
"Reviewer" for libceph and ceph.

Signed-off-by: Jeff Layton <jlayton@kernel.org>
Acked-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: fix decoding of client session messages flags
Luís Henriques [Mon, 23 May 2022 16:09:51 +0000 (17:09 +0100)]
ceph: fix decoding of client session messages flags

The cephfs kernel client started to show  the message:

 ceph: mds0 session blocklisted

when mounting a filesystem.  This is due to the fact that the session
messages are being incorrectly decoded: the skip needs to take into
account the 'len'.

While there, fixed some whitespaces too.

Cc: stable@vger.kernel.org
Fixes: e1c9788cb397 ("ceph: don't rely on error_string to validate blocklisted session.")
Signed-off-by: Luís Henriques <lhenriques@suse.de>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: switch TASK_INTERRUPTIBLE to TASK_KILLABLE
Xiubo Li [Wed, 18 May 2022 14:49:26 +0000 (22:49 +0800)]
ceph: switch TASK_INTERRUPTIBLE to TASK_KILLABLE

If the task is placed in the TASK_INTERRUPTIBLE state it will sleep
until either something explicitly wakes it up, or a non-masked signal
is received. Switch to TASK_KILLABLE to avoid the noises.

Cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: remove redundant variable ino
Colin Ian King [Wed, 18 May 2022 08:55:08 +0000 (09:55 +0100)]
ceph: remove redundant variable ino

Variable ino is being assigned a value that is never read. The variable
and assignment are redundant, remove it.

Cleans up clang scan build warning:
warning: Although the value stored to 'ino' is used in the enclosing
expression, the value is never actually read from 'ino'
[deadcode.DeadStores]

Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: try to queue a writeback if revoking fails
Xiubo Li [Wed, 27 Apr 2022 06:14:41 +0000 (14:14 +0800)]
ceph: try to queue a writeback if revoking fails

If the pagecaches writeback just finished and the i_wrbuffer_ref
reaches zero it will try to trigger ceph_check_caps(). But if just
before ceph_check_caps() the i_wrbuffer_ref could be increased
again by mmap/cache write, then the Fwb revoke will fail.

We need to try to queue a writeback in this case instead of
triggering the writeback by BDI's delayed work per 5 seconds.

URL: https://tracker.ceph.com/issues/46904
URL: https://tracker.ceph.com/issues/55377
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: fix statfs for subdir mounts
Luís Henriques [Wed, 27 Apr 2022 15:57:04 +0000 (16:57 +0100)]
ceph: fix statfs for subdir mounts

When doing a mount using as base a directory that has 'max_bytes' quotas
statfs uses that value as the total; if a subdirectory is used instead,
the same 'max_bytes' too in statfs, unless there is another quota set.

Unfortunately, if this subdirectory only has the 'max_files' quota set,
then statfs uses the filesystem total.  Fix this by making sure we only
lookup realms that contain the 'max_bytes' quota.

Cc: Ryan Taylor <rptaylor@uvic.ca>
URL: https://tracker.ceph.com/issues/55090
Signed-off-by: Luís Henriques <lhenriques@suse.de>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: fix possible deadlock when holding Fwb to get inline_data
Xiubo Li [Mon, 25 Apr 2022 08:08:24 +0000 (16:08 +0800)]
ceph: fix possible deadlock when holding Fwb to get inline_data

1, mount with wsync.
2, create a file with O_RDWR, and the request was sent to mds.0:

   ceph_atomic_open()-->
     ceph_mdsc_do_request(openc)
     finish_open(file, dentry, ceph_open)-->
       ceph_open()-->
         ceph_init_file()-->
           ceph_init_file_info()-->
             ceph_uninline_data()-->
             {
               ...
               if (inline_version == 1 || /* initial version, no data */
                   inline_version == CEPH_INLINE_NONE)
                     goto out_unlock;
               ...
             }

The inline_version will be 1, which is the initial version for the
new create file. And here the ci->i_inline_version will keep with 1,
it's buggy.

3, buffer write to the file immediately:

   ceph_write_iter()-->
     ceph_get_caps(file, need=Fw, want=Fb, ...);
     generic_perform_write()-->
       a_ops->write_begin()-->
         ceph_write_begin()-->
           netfs_write_begin()-->
             netfs_begin_read()-->
               netfs_rreq_submit_slice()-->
                 netfs_read_from_server()-->
                   rreq->netfs_ops->issue_read()-->
                     ceph_netfs_issue_read()-->
                     {
                       ...
                       if (ci->i_inline_version != CEPH_INLINE_NONE &&
                           ceph_netfs_issue_op_inline(subreq))
                         return;
                       ...
                     }
     ceph_put_cap_refs(ci, Fwb);

The ceph_netfs_issue_op_inline() will send a getattr(Fsr) request to
mds.1.

4, then the mds.1 will request the rd lock for CInode::filelock from
the auth mds.0, the mds.0 will do the CInode::filelock state transation
from excl --> sync, but it need to revoke the Fxwb caps back from the
clients.

While the kernel client has aleady held the Fwb caps and waiting for
the getattr(Fsr).

It's deadlock!

URL: https://tracker.ceph.com/issues/55377
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: redirty the page for writepage on failure
Xiubo Li [Sun, 24 Apr 2022 09:35:53 +0000 (17:35 +0800)]
ceph: redirty the page for writepage on failure

When run out of memories we should redirty the page before failing
the writepage. Or we will hit BUG_ON(folio_get_private(folio)) in
ceph_dirty_folio().

URL: https://tracker.ceph.com/issues/55421
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: try to choose the auth MDS if possible for getattr
Xiubo Li [Thu, 21 Apr 2022 03:26:40 +0000 (11:26 +0800)]
ceph: try to choose the auth MDS if possible for getattr

If any 'x' caps is issued we can just choose the auth MDS instead
of the random replica MDSes. Because only when the Locker is in
LOCK_EXEC state will the loner client could get the 'x' caps. And
if we send the getattr requests to any replica MDS it must auth pin
and tries to rdlock from the auth MDS, and then the auth MDS need
to do the Locker state transition to LOCK_SYNC. And after that the
lock state will change back.

This cost much when doing the Locker state transition and usually
will need to revoke caps from clients.

URL: https://tracker.ceph.com/issues/55240
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: disable updating the atime since cephfs won't maintain it
Xiubo Li [Wed, 20 Apr 2022 05:13:02 +0000 (13:13 +0800)]
ceph: disable updating the atime since cephfs won't maintain it

Since CephFS makes no attempt to maintain atime, we shouldn't
try to update it in mmap and generic read cases and ignore updating
it in direct and sync read cases.

And even we update it in mmap and generic read cases we will drop
it and won't sync it to MDS. And we are seeing the atime will be
updated and then dropped to the floor again and again.

URL: https://lists.ceph.io/hyperkitty/list/ceph-users@ceph.io/thread/VSJM7T4CS5TDRFF6XFPIYMHP75K73PZ6/
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Acked-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: flush the mdlog for filesystem sync
Xiubo Li [Tue, 19 Apr 2022 00:58:49 +0000 (08:58 +0800)]
ceph: flush the mdlog for filesystem sync

Before waiting for a request's safe reply, we will send the mdlog flush
request to the relevant MDS. And this will also flush the mdlog for all
the other unsafe requests in the same session, so we can record the last
session and no need to flush mdlog again in the next loop. But there
still have cases that it may send the mdlog flush requst twice or more,
but that should be not often.

Rename wait_unsafe_requests() to
flush_mdlog_and_wait_mdsc_unsafe_requests() to make it more
descriptive.

[xiubli: fold in MDS request refcount leak fix from Jeff]

URL: https://tracker.ceph.com/issues/55284
URL: https://tracker.ceph.com/issues/55411
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: rename unsafe_request_wait()
Xiubo Li [Mon, 18 Apr 2022 12:04:51 +0000 (20:04 +0800)]
ceph: rename unsafe_request_wait()

Rename it to flush_mdlog_and_wait_inode_unsafe_requests() to make
it more descriptive.

Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agolibceph: use swap() macro instead of taking tmp variable
Guo Zhengkui [Tue, 12 Apr 2022 06:46:20 +0000 (14:46 +0800)]
libceph: use swap() macro instead of taking tmp variable

Fix the following coccicheck warning:
net/ceph/crush/mapper.c:1077:8-9: WARNING opportunity for swap()

by using swap() for the swapping of variable values and drop
the tmp variable that is not needed any more.

Signed-off-by: Guo Zhengkui <guozhengkui@vivo.com>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: fix statx AT_STATX_DONT_SYNC vs AT_STATX_FORCE_SYNC check
Xiubo Li [Mon, 11 Apr 2022 01:59:09 +0000 (09:59 +0800)]
ceph: fix statx AT_STATX_DONT_SYNC vs AT_STATX_FORCE_SYNC check

From the posix and the initial statx supporting commit comments,
the AT_STATX_DONT_SYNC is a lightweight stat and the
AT_STATX_FORCE_SYNC is a heaverweight one. And also checked all
the other current usage about these two flags they are all doing
the same, that is only when the AT_STATX_FORCE_SYNC is not set
and the AT_STATX_DONT_SYNC is set will they skip sync retriving
the attributes from storage.

Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: David Howells <dhowells@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: no need to invalidate the fscache twice
Xiubo Li [Thu, 7 Apr 2022 05:12:42 +0000 (13:12 +0800)]
ceph: no need to invalidate the fscache twice

Fixes: 400e1286c0ec3 ("ceph: conversion to new fscache API")
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: replace usage of found with dedicated list iterator variable
Jakob Koschel [Thu, 31 Mar 2022 21:53:29 +0000 (23:53 +0200)]
ceph: replace usage of found with dedicated list iterator variable

To move the list iterator variable into the list_for_each_entry_*()
macro in the future it should be avoided to use the list iterator
variable after the loop body.

To *never* use the list iterator variable after the loop it was
concluded to use a separate iterator variable instead of a
found boolean.

This removes the need to use a found variable and simply checking if
the variable was set, can determine if the break/goto was hit.

Link: https://lore.kernel.org/all/CAHk-=wgRr_D8CB-D9Kg-c=EHreAsk5SqXPwr9Y7k9sA6cWXJ6w@mail.gmail.com/
Signed-off-by: Jakob Koschel <jakobkoschel@gmail.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: use dedicated list iterator variable
Jakob Koschel [Thu, 31 Mar 2022 21:53:28 +0000 (23:53 +0200)]
ceph: use dedicated list iterator variable

To move the list iterator variable into the list_for_each_entry_*()
macro in the future it should be avoided to use the list iterator
variable after the loop body.

To *never* use the list iterator variable after the loop it was
concluded to use a separate iterator variable.

Link: https://lore.kernel.org/all/CAHk-=wgRr_D8CB-D9Kg-c=EHreAsk5SqXPwr9Y7k9sA6cWXJ6w@mail.gmail.com/
Signed-off-by: Jakob Koschel <jakobkoschel@gmail.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: update the dlease for the hashed dentry when removing
Xiubo Li [Wed, 30 Mar 2022 04:21:12 +0000 (12:21 +0800)]
ceph: update the dlease for the hashed dentry when removing

The MDS will always refresh the dentry lease when removing the files
or directories. And if the dentry is still hashed, we can update
the dentry lease and no need to do the lookup from the MDS later.

Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: stop retrying the request when exceeding 256 times
Xiubo Li [Wed, 30 Mar 2022 06:39:33 +0000 (14:39 +0800)]
ceph: stop retrying the request when exceeding 256 times

The type of 'r_attempts' in kernel 'ceph_mds_request' is 'int',
while in 'ceph_mds_request_head' the type of 'num_retry' is '__u8'.
So in case the request retries exceeding 256 times, the MDS will
receive a incorrect retry seq.

In this case it's ususally a bug in MDS and continue retrying the
request makes no sense. For now let's limit it to 256. In future
this could be fixed in ceph code, so avoid using the hardcode here.

Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: stop forwarding the request when exceeding 256 times
Xiubo Li [Tue, 29 Mar 2022 04:48:01 +0000 (12:48 +0800)]
ceph: stop forwarding the request when exceeding 256 times

The type of 'num_fwd' in ceph 'MClientRequestForward' is 'int32_t',
while in 'ceph_mds_request_head' the type is '__u8'. So in case
the request bounces between MDSes exceeding 256 times, the client
will get stuck.

In this case it's ususally a bug in MDS and continue bouncing the
request makes no sense.

URL: https://tracker.ceph.com/issues/55130
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Luís Henriques <lhenriques@suse.de>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: remove unused CEPH_MDS_LEASE_RELEASE related code
Xiubo Li [Mon, 28 Mar 2022 02:25:35 +0000 (10:25 +0800)]
ceph: remove unused CEPH_MDS_LEASE_RELEASE related code

The ceph_mdsc_lease_release() has been removed by commit 8aa152c77890
(ceph: remove ceph_mdsc_lease_release). ceph_mdsc_lease_send_msg will
never be called with CEPH_MDS_LEASE_RELEASE.

Signed-off-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agorbd: replace usage of found with dedicated list iterator variable
Jakob Koschel [Thu, 24 Mar 2022 07:20:50 +0000 (08:20 +0100)]
rbd: replace usage of found with dedicated list iterator variable

To move the list iterator variable into the list_for_each_entry_*()
macro in the future it should be avoided to use the list iterator
variable after the loop body.

To *never* use the list iterator variable after the loop it was
concluded to use a separate iterator variable instead of a
found boolean.

This removes the need to use a found variable and simply checking if
the variable was set, can determine if the break/goto was hit.

Link: https://lore.kernel.org/all/CAHk-=wgRr_D8CB-D9Kg-c=EHreAsk5SqXPwr9Y7k9sA6cWXJ6w@mail.gmail.com/
Signed-off-by: Jakob Koschel <jakobkoschel@gmail.com>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoceph: allow ceph.dir.rctime xattr to be updatable
Venky Shankar [Thu, 10 Mar 2022 14:34:19 +0000 (09:34 -0500)]
ceph: allow ceph.dir.rctime xattr to be updatable

`rctime' has been a pain point in cephfs due to its buggy
nature - inconsistent values reported and those sorts.
Fixing rctime is non-trivial needing an overall redesign
of the entire nested statistics infrastructure.

As a workaround, PR

     http://github.com/ceph/ceph/pull/37938

allows this extended attribute to be manually set. This allows
users to "fixup" inconsistent rctime values. While this sounds
messy, its probably the wisest approach allowing users/scripts
to workaround buggy rctime values.

The above PR enables Ceph MDS to allow manually setting
rctime extended attribute with the corresponding user-land
changes. We may as well allow the same to be done via kclient
for parity.

Signed-off-by: Venky Shankar <vshankar@redhat.com>
Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
3 years agoLinux 5.18
Linus Torvalds [Sun, 22 May 2022 19:52:31 +0000 (09:52 -1000)]
Linux 5.18

3 years agoafs: Fix afs_getattr() to refetch file status if callback break occurred
David Howells [Sat, 21 May 2022 07:18:28 +0000 (08:18 +0100)]
afs: Fix afs_getattr() to refetch file status if callback break occurred

If a callback break occurs (change notification), afs_getattr() needs to
issue an FS.FetchStatus RPC operation to update the status of the file
being examined by the stat-family of system calls.

Fix afs_getattr() to do this if AFS_VNODE_CB_PROMISED has been cleared
on a vnode by a callback break.  Skip this if AT_STATX_DONT_SYNC is set.

This can be tested by appending to a file on one AFS client and then
using "stat -L" to examine its length on a machine running kafs.  This
can also be watched through tracing on the kafs machine.  The callback
break is seen:

     kworker/1:1-46      [001] .....   978.910812: afs_cb_call: c=0000005f YFSCB.CallBack
     kworker/1:1-46      [001] ...1.   978.910829: afs_cb_break: 100058:23b4c:242d2c2 b=2 s=1 break-cb
     kworker/1:1-46      [001] .....   978.911062: afs_call_done:    c=0000005f ret=0 ab=0 [0000000082994ead]

And then the stat command generated no traffic if unpatched, but with
this change a call to fetch the status can be observed:

            stat-4471    [000] .....   986.744122: afs_make_fs_call: c=000000ab 100058:023b4c:242d2c2 YFS.FetchStatus
            stat-4471    [000] .....   986.745578: afs_call_done:    c=000000ab ret=0 ab=0 [0000000087fc8c84]

Fixes: 08e0e7c82eea ("[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC.")
Reported-by: Markus Suvanto <markus.suvanto@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Tested-by: Markus Suvanto <markus.suvanto@gmail.com>
Tested-by: kafs-testing+fedora34_64checkkafs-build-496@auristor.com
Link: https://bugzilla.kernel.org/show_bug.cgi?id=216010
Link: https://lore.kernel.org/r/165308359800.162686.14122417881564420962.stgit@warthog.procyon.org.uk/
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
3 years agoMerge branch 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Sun, 22 May 2022 18:04:38 +0000 (08:04 -1000)]
Merge branch 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux

Pull i2c fixes from Wolfram Sang:
 "Some I2C driver bugfixes for 5.18. Nothing spectacular but worth
  fixing"

* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  drivers: i2c: thunderx: Allow driver to work with ACPI defined TWSI controllers
  i2c: ismt: Provide a DMA buffer for Interrupt Cause Logging
  i2c: mt7621: fix missing clk_disable_unprepare() on error in mtk_i2c_probe()

3 years agoMerge tag 'perf-tools-fixes-for-v5.18-2022-05-21' of git://git.kernel.org/pub/scm...
Linus Torvalds [Sun, 22 May 2022 00:14:02 +0000 (14:14 -1000)]
Merge tag 'perf-tools-fixes-for-v5.18-2022-05-21' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux

Pull perf tools fixes from Arnaldo Carvalho de Melo:

 - Fix and validate CPU map inputs in synthetic PERF_RECORD_STAT events
   in 'perf stat'.

 - Fix x86's arch__intr_reg_mask() for the hybrid platform.

 - Address 'perf bench numa' compiler error on s390.

 - Fix check for btf__load_from_kernel_by_id() in libbpf.

 - Fix "all PMU test" 'perf test' to skip hv_24x7/hv_gpci tests on
   powerpc.

 - Fix session topology test to skip the test in guest environment.

 - Skip BPF 'perf test' if clang is not present.

 - Avoid shell test description infinite loop in 'perf test'.

 - Fix Intel LBR callstack entries and nr print message.

* tag 'perf-tools-fixes-for-v5.18-2022-05-21' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux:
  perf session: Fix Intel LBR callstack entries and nr print message
  perf test bpf: Skip test if clang is not present
  perf test session topology: Fix test to skip the test in guest environment
  perf bench numa: Address compiler error on s390
  perf test: Avoid shell test description infinite loop
  perf regs x86: Fix arch__intr_reg_mask() for the hybrid platform
  perf test: Fix "all PMU test" to skip hv_24x7/hv_gpci tests on powerpc
  perf stat: Fix and validate CPU map inputs in synthetic PERF_RECORD_STAT events
  perf build: Fix check for btf__load_from_kernel_by_id() in libbpf

3 years agoMerge tag 'input-for-v5.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 21 May 2022 23:58:43 +0000 (13:58 -1000)]
Merge tag 'input-for-v5.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input

Pull input fixes from Dmitry Torokhov:
 "A small fixup to ili210x touchscreen driver, and updated maintainer
  entry for the device tree binding of Mediatek 6779 keypad:

   - fix reset timing of Ilitek touchscreens

   - update maintainer entry of DT binding of Mediatek 6779 keypad"

* tag 'input-for-v5.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: ili210x - use one common reset implementation
  Input: ili210x - fix reset timing
  dt-bindings: input: mediatek,mt6779-keypad: update maintainer

3 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Sat, 21 May 2022 23:31:50 +0000 (13:31 -1000)]
Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "Two patches, both in drivers.

  The iscsi one is fixing the cpumask issue you commented on and the ufs
  one is a late arriving fix for conditions that can occur in Host
  Performance Booster reads"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: ufs: core: Fix referencing invalid rsp field
  scsi: target: Fix incorrect use of cpumask_t

3 years agoperf session: Fix Intel LBR callstack entries and nr print message
Chengdong Li [Tue, 17 May 2022 01:57:26 +0000 (09:57 +0800)]
perf session: Fix Intel LBR callstack entries and nr print message

When generating callstack information from branch_stack(Intel LBR), the
actual number of callstack entry should be bigger than the number of
branch_stack, for example:

branch_stack records:
B() -> C()
A() -> B()
converted callstack records should be:
C()
B()
A()
though, the number of callstack equals
to the number of branch stack plus 1.

This patch fixes above issue in branch_stack__printf(). For example,

# echo 'scale=2000; 4*a(1)' > cmd
# perf record --call-graph lbr bc -l < cmd

Before applying this patch, `perf script -D` output:

1220022677386876 0x2a40 [0xd8]: PERF_RECORD_SAMPLE(IP, 0x4002): 17990/17990: 0x40a6d6 period: 894172 addr: 0
... LBR call chain: nr:8
.....  0: fffffffffffffe00
.....  1: 000000000040a410
.....  2: 000000000040573c
.....  3: 0000000000408650
.....  4: 00000000004022f2
.....  5: 00000000004015f5
.....  6: 00007f5ed6dcb553
.....  7: 0000000000401698
... FP chain: nr:2
.....  0: fffffffffffffe00
.....  1: 000000000040a6d8
... branch callstack: nr:6    # which is not consistent with LBR records.
.....  0: 000000000040a410
.....  1: 0000000000408650    # ditto
.....  2: 00000000004022f2
.....  3: 00000000004015f5
.....  4: 00007f5ed6dcb553
.....  5: 0000000000401698
 ... thread: bc:17990
 ...... dso: /usr/bin/bc
bc 17990 1220022.677386:     894172 cycles:
  40a410 [unknown] (/usr/bin/bc)
  40573c [unknown] (/usr/bin/bc)
  408650 [unknown] (/usr/bin/bc)
  4022f2 [unknown] (/usr/bin/bc)
  4015f5 [unknown] (/usr/bin/bc)
    7f5ed6dcb553 __libc_start_main+0xf3 (/usr/lib64/libc-2.17.so)
  401698 [unknown] (/usr/bin/bc)

After applied:

1220022677386876 0x2a40 [0xd8]: PERF_RECORD_SAMPLE(IP, 0x4002): 17990/17990: 0x40a6d6 period: 894172 addr: 0
... LBR call chain: nr:8
.....  0: fffffffffffffe00
.....  1: 000000000040a410
.....  2: 000000000040573c
.....  3: 0000000000408650
.....  4: 00000000004022f2
.....  5: 00000000004015f5
.....  6: 00007f5ed6dcb553
.....  7: 0000000000401698
... FP chain: nr:2
.....  0: fffffffffffffe00
.....  1: 000000000040a6d8
... branch callstack: nr:7
.....  0: 000000000040a410
.....  1: 000000000040573c
.....  2: 0000000000408650
.....  3: 00000000004022f2
.....  4: 00000000004015f5
.....  5: 00007f5ed6dcb553
.....  6: 0000000000401698
 ... thread: bc:17990
 ...... dso: /usr/bin/bc
bc 17990 1220022.677386:     894172 cycles:
  40a410 [unknown] (/usr/bin/bc)
  40573c [unknown] (/usr/bin/bc)
  408650 [unknown] (/usr/bin/bc)
  4022f2 [unknown] (/usr/bin/bc)
  4015f5 [unknown] (/usr/bin/bc)
    7f5ed6dcb553 __libc_start_main+0xf3 (/usr/lib64/libc-2.17.so)
  401698 [unknown] (/usr/bin/bc)

Change from v1:
- refined code style according to Jiri's review comments.

Signed-off-by: Chengdong Li <chengdongli@tencent.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexey Bayduraev <alexey.v.bayduraev@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: German Gomez <german.gomez@arm.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Riccardo Mancini <rickyman7@gmail.com>
Cc: likexu@tencent.com
Link: https://lore.kernel.org/r/20220517015726.96131-1-chengdongli@tencent.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
3 years agoperf test bpf: Skip test if clang is not present
Athira Rajeev [Wed, 11 May 2022 11:54:38 +0000 (17:24 +0530)]
perf test bpf: Skip test if clang is not present

Perf BPF filter test fails in environment where "clang" is not
installed.

Test failure logs:

<<>>
 42: BPF filter                    :
 42.1: Basic BPF filtering         : Skip
 42.2: BPF pinning                 : FAILED!
 42.3: BPF prologue generation     : FAILED!
<<>>

Enabling verbose option provided debug logs which says clang/llvm needs
to be installed. Snippet of verbose logs:

<<>>
 42.2: BPF pinning                  :
 --- start ---
test child forked, pid 61423
ERROR: unable to find clang.
Hint: Try to install latest clang/llvm to support BPF.
        Check your $PATH

<<logs_here>>

Failed to compile test case: 'Basic BPF llvm compile'
Unable to get BPF object, fix kbuild first
test child finished with -1
 ---- end ----
BPF filter subtest 2: FAILED!
<<>>

Here subtests, "BPF pinning" and "BPF prologue generation" failed and
logs shows clang/llvm is needed. After installing clang, testcase
passes.

Reason on why subtest failure happens though logs has proper debug
information:

Main function __test__bpf calls test_llvm__fetch_bpf_obj by
passing 4th argument as true ( 4th arguments maps to parameter
"force" in test_llvm__fetch_bpf_obj ). But this will cause
test_llvm__fetch_bpf_obj to skip the check for clang/llvm.

Snippet of code part which checks for clang based on
parameter "force" in test_llvm__fetch_bpf_obj:

<<>>
if (!force && (!llvm_param.user_set_param &&
<<>>

Since force is set to "false", test won't get skipped and fails to
compile test case. The BPF code compilation needs clang, So pass the
fourth argument as "false" and also skip the test if reason for return
is "TEST_SKIP"

After the patch:

<<>>
 42: BPF filter                    :
 42.1: Basic BPF filtering         : Skip
 42.2: BPF pinning                 : Skip
 42.3: BPF prologue generation     : Skip
<<>>

Fixes: ba1fae431e74bb42 ("perf test: Add 'perf test BPF'")
Reviewed-by: Kajol Jain <kjain@linux.ibm.com>
Signed-off-by: Athira Jajeev <atrajeev@linux.vnet.ibm.com>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Disha Goel <disgoel@linux.vnet.ibm.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Nageswara R Sastry <rnsastry@linux.ibm.com>
Cc: Wang Nan <wangnan0@huawei.com>
Link: https://lore.kernel.org/r/20220511115438.84032-1-atrajeev@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
3 years agoperf test session topology: Fix test to skip the test in guest environment
Athira Rajeev [Wed, 11 May 2022 11:49:59 +0000 (17:19 +0530)]
perf test session topology: Fix test to skip the test in guest environment

The session topology test fails in powerpc pSeries platform.

Test logs:

  <<>>
  Session topology : FAILED!
  <<>>

This testcases tests cpu topology by checking the core_id and socket_id
stored in perf_env from perf session. The data from perf session is
compared with the cpu topology information from
"/sys/devices/system/cpu/cpuX/topology" like core_id,
physical_package_id.

In case of virtual environment, detail like physical_package_id is
restricted to be exposed. Hence physical_package_id is set to -1. The
testcase fails on such platforms since socket_id can't be fetched from
topology info.

Skip the testcase in powerpc if physical_package_id returns -1.

Reviewed-by: Kajol Jain <kjain@linux.ibm.com>
Signed-off-by: Athira Rajeev <atrajeev@linux.vnet.ibm.com>---
Tested-by: Disha Goel <disgoel@linux.vnet.ibm.com>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Nageswara R Sastry <rnsastry@linux.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Link: https://lore.kernel.org/r/20220511114959.84002-1-atrajeev@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
3 years agoperf bench numa: Address compiler error on s390
Thomas Richter [Fri, 20 May 2022 08:11:58 +0000 (10:11 +0200)]
perf bench numa: Address compiler error on s390

The compilation on s390 results in this error:

  # make DEBUG=y bench/numa.o
  ...
  bench/numa.c: In function ‘__bench_numa’:
  bench/numa.c:1749:81: error: ‘%d’ directive output may be truncated
              writing between 1 and 11 bytes into a region of size between
              10 and 20 [-Werror=format-truncation=]
  1749 |        snprintf(tname, sizeof(tname), "process%d:thread%d", p, t);
                                                               ^~
  ...
  bench/numa.c:1749:64: note: directive argument in the range
                 [-21474836472147483646]
  ...
  #

The maximum length of the %d replacement is 11 characters because of the
negative sign.  Therefore extend the array by two more characters.

Output after:

  # make  DEBUG=y bench/numa.o > /dev/null 2>&1; ll bench/numa.o
  -rw-r--r-- 1 root root 418320 May 19 09:11 bench/numa.o
  #

Fixes: 3aff8ba0a4c9c919 ("perf bench numa: Avoid possible truncation when using snprintf()")
Suggested-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: Sumanth Korikkar <sumanthk@linux.ibm.com>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Link: https://lore.kernel.org/r/20220520081158.2990006-1-tmricht@linux.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
3 years agoperf test: Avoid shell test description infinite loop
Ian Rogers [Tue, 17 May 2022 20:41:44 +0000 (13:41 -0700)]
perf test: Avoid shell test description infinite loop

for_each_shell_test() is already strict in expecting tests to be files
and executable. It is sometimes possible when it iterates over all files
that it finds one that is executable and lacks a newline character. When
this happens the loop never terminates as it doesn't check for EOF.

Add the EOF check to make this loop at least bounded by the file size.

If the description is returned as NULL then also skip the test.

Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Marco Elver <elver@google.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Riccardo Mancini <rickyman7@gmail.com>
Cc: Sohaib Mohamed <sohaib.amhmd@gmail.com>
Cc: Stephane Eranian <eranian@google.com>
Link: https://lore.kernel.org/r/20220517204144.645913-1-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
3 years agoperf regs x86: Fix arch__intr_reg_mask() for the hybrid platform
Kan Liang [Wed, 18 May 2022 14:51:25 +0000 (07:51 -0700)]
perf regs x86: Fix arch__intr_reg_mask() for the hybrid platform

The X86 specific arch__intr_reg_mask() is to check whether the kernel
and hardware can collect XMM registers. But it doesn't work on some
hybrid platform.

Without the patch on ADL-N:

  $ perf record -I?
  available registers: AX BX CX DX SI DI BP SP IP FLAGS CS SS R8 R9 R10
  R11 R12 R13 R14 R15

The config of the test event doesn't contain the PMU information. The
kernel may fail to initialize it on the correct hybrid PMU and return
the wrong non-supported information.

Add the PMU information into the config for the hybrid platform. The
same register set is supported among different hybrid PMUs. Checking
the first available one is good enough.

With the patch on ADL-N:

  $ perf record -I?
  available registers: AX BX CX DX SI DI BP SP IP FLAGS CS SS R8 R9 R10
  R11 R12 R13 R14 R15 XMM0 XMM1 XMM2 XMM3 XMM4 XMM5 XMM6 XMM7 XMM8 XMM9
  XMM10 XMM11 XMM12 XMM13 XMM14 XMM15

Fixes: 6466ec14aaf44ff1 ("perf regs x86: Add X86 specific arch__intr_reg_mask()")
Reported-by: Ammy Yi <ammy.yi@intel.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Xing Zhengjun <zhengjun.xing@linux.intel.com>
Link: https://lore.kernel.org/r/20220518145125.1494156-1-kan.liang@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
3 years agoperf test: Fix "all PMU test" to skip hv_24x7/hv_gpci tests on powerpc
Athira Rajeev [Fri, 20 May 2022 10:12:36 +0000 (15:42 +0530)]
perf test: Fix "all PMU test" to skip hv_24x7/hv_gpci tests on powerpc

"perf all PMU test" picks the input events from "perf list --raw-dump
pmu" list and runs "perf stat -e" for each of the event in the list. In
case of powerpc, the PowerVM environment supports events from hv_24x7
and hv_gpci PMU which is of example format like below:

- hv_24x7/CPM_ADJUNCT_INST,domain=?,core=?/
- hv_gpci/event,partition_id=?/

The value for "?" needs to be filled in depending on system and
respective event. CPM_ADJUNCT_INST needs have core value and domain
value. hv_gpci event needs partition_id.  Similarly, there are other
events for hv_24x7 and hv_gpci having "?" in event format. Hence skip
these events on powerpc platform since values like partition_id, domain
is specific to system and event.

Fixes: 3d5ac9effcc640d5 ("perf test: Workload test of all PMUs")
Signed-off-by: Athira Jajeev <atrajeev@linux.vnet.ibm.com>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Disha Goel <disgoel@linux.vnet.ibm.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Nageswara R Sastry <rnsastry@linux.ibm.com>
Link: https://lore.kernel.org/r/20220520101236.17249-1-atrajeev@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
3 years agodrivers: i2c: thunderx: Allow driver to work with ACPI defined TWSI controllers
Piyush Malgujar [Wed, 11 May 2022 13:36:59 +0000 (06:36 -0700)]
drivers: i2c: thunderx: Allow driver to work with ACPI defined TWSI controllers

Due to i2c->adap.dev.fwnode not being set, ACPI_COMPANION() wasn't properly
found for TWSI controllers.

Signed-off-by: Szymon Balcerak <sbalcerak@marvell.com>
Signed-off-by: Piyush Malgujar <pmalgujar@marvell.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
3 years agoi2c: ismt: Provide a DMA buffer for Interrupt Cause Logging
Mika Westerberg [Wed, 27 Apr 2022 10:19:10 +0000 (13:19 +0300)]
i2c: ismt: Provide a DMA buffer for Interrupt Cause Logging

Before sending a MSI the hardware writes information pertinent to the
interrupt cause to a memory location pointed by SMTICL register. This
memory holds three double words where the least significant bit tells
whether the interrupt cause of master/target/error is valid. The driver
does not use this but we need to set it up because otherwise it will
perform DMA write to the default address (0) and this will cause an
IOMMU fault such as below:

  DMAR: DRHD: handling fault status reg 2
  DMAR: [DMA Write] Request device [00:12.0] PASID ffffffff fault addr 0
        [fault reason 05] PTE Write access is not set

To prevent this from happening, provide a proper DMA buffer for this
that then gets mapped by the IOMMU accordingly.

Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: From: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
3 years agoi2c: mt7621: fix missing clk_disable_unprepare() on error in mtk_i2c_probe()
Yang Yingliang [Sat, 14 May 2022 02:31:47 +0000 (10:31 +0800)]
i2c: mt7621: fix missing clk_disable_unprepare() on error in mtk_i2c_probe()

Fix the missing clk_disable_unprepare() before return
from mtk_i2c_probe() in the error handling case.

Fixes: d04913ec5f89 ("i2c: mt7621: Add MediaTek MT7621/7628/7688 I2C driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
3 years agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Sat, 21 May 2022 06:34:59 +0000 (20:34 -1000)]
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm

Pull kvm fixes from Paolo Bonzini:
 "ARM:

   - Correctly expose GICv3 support even if no irqchip is created so
     that userspace doesn't observe it changing pointlessly (fixing a
     regression with QEMU)

   - Don't issue a hypercall to set the id-mapped vectors when protected
     mode is enabled (fix for pKVM in combination with CPUs affected by
     Spectre-v3a)

  x86 (five oneliners, of which the most interesting two are):

   - a NULL pointer dereference on INVPCID executed with paging
     disabled, but only if KVM is using shadow paging

   - an incorrect bsearch comparison function which could truncate the
     result and apply PMU event filtering incorrectly. This one comes
     with a selftests update too"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: x86/mmu: fix NULL pointer dereference on guest INVPCID
  KVM: x86: hyper-v: fix type of valid_bank_mask
  KVM: Free new dirty bitmap if creating a new memslot fails
  KVM: eventfd: Fix false positive RCU usage warning
  selftests: kvm/x86: Verify the pmu event filter matches the correct event
  selftests: kvm/x86: Add the helper function create_pmu_event_filter
  kvm: x86/pmu: Fix the compare function used by the pmu event filter
  KVM: arm64: Don't hypercall before EL2 init
  KVM: arm64: vgic-v3: Consistently populate ID_AA64PFR0_EL1.GIC
  KVM: x86/mmu: Update number of zapped pages even if page list is stable

3 years agoMerge tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 21 May 2022 05:07:28 +0000 (19:07 -1000)]
Merge tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux

Pull clk fixes from Stephen Boyd:
 "Three clk driver fixes to close out the release

   - Fix a divider calculation breaking boot on Broadcom bcm2835

   - Fix HDMI output on Tanix TX6 mini board by reverting a patch

   - Fix clk_set_rate_range() calls on at91 by considering the range
     while calculating the divisor"

* tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux:
  clk: at91: generated: consider range when calculating best rate
  Revert "clk: sunxi-ng: sun6i-rtc: Add support for H6"
  clk: bcm2835: fix bcm2835_clock_choose_div

3 years agoMerge tag 'drm-fixes-2022-05-21' of git://anongit.freedesktop.org/drm/drm
Linus Torvalds [Sat, 21 May 2022 04:58:37 +0000 (18:58 -1000)]
Merge tag 'drm-fixes-2022-05-21' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "Few final fixes for 5.18, one amdgpu, core dp mst leak fix, dma-buf
  two fixes, and i915 has a few fixes, one for a regression on older
  GM45 chipsets,

  dma-buf:
   - ioctl userspace use fix
   - fix dma-buf sysfs name generation

  core:
   - dp/mst leak fix

  amdgpu:
   - suspend/resume regression fix

  i915:
   - fix for #5806: GPU hangs and display artifacts on Intel GM45
   - reject DMC with out-of-spec MMIO
   - correctly mark guilty contexts on GuC reset"

* tag 'drm-fixes-2022-05-21' of git://anongit.freedesktop.org/drm/drm:
  drm/i915: Use i915_gem_object_ggtt_pin_ww for reloc_iomap
  drm/amd: Don't reset dGPUs if the system is going to s2idle
  drm/dp/mst: fix a possible memory leak in fetch_monitor_name()
  dma-buf: fix use of DMA_BUF_SET_NAME_{A,B} in userspace
  i915/guc/reset: Make __guc_reset_context aware of guilty engines
  drm/i915/dmc: Add MMIO range restrictions
  dma-buf: ensure unique directory name for dmabuf stats

3 years agoMerge tag 'drm-intel-fixes-2022-05-20' of git://anongit.freedesktop.org/drm/drm-intel...
Dave Airlie [Fri, 20 May 2022 20:00:48 +0000 (06:00 +1000)]
Merge tag 'drm-intel-fixes-2022-05-20' of git://anongit.freedesktop.org/drm/drm-intel into drm-fixes

- fix for #5806: GPU hangs and display artifacts on 5.18-rc3 on Intel GM45
- reject DMC with out-of-spec MMIO (Cc: stable)
- correctly mark guilty contexts on GuC reset.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/YocqqvG6PbYx3QgJ@jlahtine-mobl.ger.corp.intel.com
3 years agoMerge tag 'drm-misc-fixes-2022-05-20' of git://anongit.freedesktop.org/drm/drm-misc...
Dave Airlie [Fri, 20 May 2022 19:58:59 +0000 (05:58 +1000)]
Merge tag 'drm-misc-fixes-2022-05-20' of git://anongit.freedesktop.org/drm/drm-misc into drm-fixes

Fix for a memory leak in dp_mst, a (userspace) build fix for
DMA_BUF_SET_NAME defines and a directory name generation fix for dmabuf
stats

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20220520072408.cpjzy2taugagvrh7@houat
3 years agoperf: Fix sys_perf_event_open() race against self
Peter Zijlstra [Fri, 20 May 2022 18:38:06 +0000 (20:38 +0200)]
perf: Fix sys_perf_event_open() race against self

Norbert reported that it's possible to race sys_perf_event_open() such
that the looser ends up in another context from the group leader,
triggering many WARNs.

The move_group case checks for races against itself, but the
!move_group case doesn't, seemingly relying on the previous
group_leader->ctx == ctx check. However, that check is racy due to not
holding any locks at that time.

Therefore, re-check the result after acquiring locks and bailing
if they no longer match.

Additionally, clarify the not_move_group case from the
move_group-vs-move_group race.

Fixes: f63a8daa5812 ("perf: Fix event->ctx locking")
Reported-by: Norbert Slusarek <nslusarek@gmx.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
3 years agoMerge tag 'gpio-fixes-for-v5.18' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 20 May 2022 18:26:28 +0000 (08:26 -1000)]
Merge tag 'gpio-fixes-for-v5.18' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux

Pull gpio fixes from Bartosz Golaszewski:

 - fix bitops logic in gpio-vf610

 - return an error if the user tries to use inverted polarity in
   gpio-mvebu

* tag 'gpio-fixes-for-v5.18' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
  gpio: mvebu/pwm: Refuse requests with inverted polarity
  gpio: gpio-vf610: do not touch other bits when set the target bit

3 years agoMerge tag 'mmc-v5.18-rc4-2' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc
Linus Torvalds [Fri, 20 May 2022 18:20:28 +0000 (08:20 -1000)]
Merge tag 'mmc-v5.18-rc4-2' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc

Pull MMC fix from Ulf Hansson:
 "MMC core:

   - Fix busy polling for MMC_SEND_OP_COND again"

* tag 'mmc-v5.18-rc4-2' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc:
  mmc: core: Fix busy polling for MMC_SEND_OP_COND again

3 years agoMerge tag 'ceph-for-5.18-rc8' of https://github.com/ceph/ceph-client
Linus Torvalds [Fri, 20 May 2022 18:15:40 +0000 (08:15 -1000)]
Merge tag 'ceph-for-5.18-rc8' of https://github.com/ceph/ceph-client

Pull ceph fix from Ilya Dryomov:
 "A fix for a nasty use-after-free, marked for stable"

* tag 'ceph-for-5.18-rc8' of https://github.com/ceph/ceph-client:
  libceph: fix misleading ceph_osdc_cancel_request() comment
  libceph: fix potential use-after-free on linger ping and resends

3 years agoMerge tag 'riscv-for-linus-5.18-rc8' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 20 May 2022 18:13:32 +0000 (08:13 -1000)]
Merge tag 'riscv-for-linus-5.18-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux

Pull RISC-V fixes from Palmer Dabbelt:

 - fix the fu540-c000 device tree to avoid a schema check failure on the
   DMA node name

 - fix typo in the PolarFire SOC device tree

* tag 'riscv-for-linus-5.18-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux:
  riscv: dts: microchip: fix gpio1 reg property typo
  riscv: dts: sifive: fu540-c000: align dma node name with dtschema

3 years agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Fri, 20 May 2022 18:09:00 +0000 (08:09 -1000)]
Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux

Pull arm64 fixes from Will Deacon:
 "Three arm64 fixes for -rc8/final.

  The MTE and stolen time fixes have been doing the rounds for a little
  while, but review and testing feedback was ongoing until earlier this
  week. The kexec fix showed up on Monday and addresses a failure
  observed under Qemu.

  Summary:

   - Add missing write barrier to publish MTE tags before a pte update

   - Fix kexec relocation clobbering its own data structures

   - Fix stolen time crash if a timer IRQ fires during CPU hotplug"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: mte: Ensure the cleared tags are visible before setting the PTE
  arm64: kexec: load from kimage prior to clobbering
  arm64: paravirt: Use RCU read locks to guard stolen_time

3 years agoKVM: x86/mmu: fix NULL pointer dereference on guest INVPCID
Paolo Bonzini [Fri, 20 May 2022 17:48:11 +0000 (13:48 -0400)]
KVM: x86/mmu: fix NULL pointer dereference on guest INVPCID

With shadow paging enabled, the INVPCID instruction results in a call
to kvm_mmu_invpcid_gva.  If INVPCID is executed with CR0.PG=0, the
invlpg callback is not set and the result is a NULL pointer dereference.
Fix it trivially by checking for mmu->invlpg before every call.

There are other possibilities:

- check for CR0.PG, because KVM (like all Intel processors after P5)
  flushes guest TLB on CR0.PG changes so that INVPCID/INVLPG are a
  nop with paging disabled

- check for EFER.LMA, because KVM syncs and flushes when switching
  MMU contexts outside of 64-bit mode

All of these are tricky, go for the simple solution.  This is CVE-2022-1789.

Reported-by: Yongkang Jia <kangel@zju.edu.cn>
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agoKVM: x86: hyper-v: fix type of valid_bank_mask
Yury Norov [Thu, 19 May 2022 17:15:04 +0000 (10:15 -0700)]
KVM: x86: hyper-v: fix type of valid_bank_mask

In kvm_hv_flush_tlb(), valid_bank_mask is declared as unsigned long,
but is used as u64, which is wrong for i386, and has been spotted by
LKP after applying "KVM: x86: hyper-v: replace bitmap_weight() with
hweight64()"

https://lore.kernel.org/lkml/20220510154750.212913-12-yury.norov@gmail.com/

But it's wrong even without that patch because now bitmap_weight()
dereferences a word after valid_bank_mask on i386.

>> include/asm-generic/bitops/const_hweight.h:21:76: warning: right shift count >= width of type
+[-Wshift-count-overflow]
      21 | #define __const_hweight64(w) (__const_hweight32(w) + __const_hweight32((w) >> 32))
         |                                                                            ^~
   include/asm-generic/bitops/const_hweight.h:10:16: note: in definition of macro '__const_hweight8'
      10 |          ((!!((w) & (1ULL << 0))) +     \
         |                ^
   include/asm-generic/bitops/const_hweight.h:20:31: note: in expansion of macro '__const_hweight16'
      20 | #define __const_hweight32(w) (__const_hweight16(w) + __const_hweight16((w) >> 16))
         |                               ^~~~~~~~~~~~~~~~~
   include/asm-generic/bitops/const_hweight.h:21:54: note: in expansion of macro '__const_hweight32'
      21 | #define __const_hweight64(w) (__const_hweight32(w) + __const_hweight32((w) >> 32))
         |                                                      ^~~~~~~~~~~~~~~~~
   include/asm-generic/bitops/const_hweight.h:29:49: note: in expansion of macro '__const_hweight64'
      29 | #define hweight64(w) (__builtin_constant_p(w) ? __const_hweight64(w) : __arch_hweight64(w))
         |                                                 ^~~~~~~~~~~~~~~~~
   arch/x86/kvm/hyperv.c:1983:36: note: in expansion of macro 'hweight64'
    1983 |                 if (hc->var_cnt != hweight64(valid_bank_mask))
         |                                    ^~~~~~~~~

CC: Borislav Petkov <bp@alien8.de>
CC: Dave Hansen <dave.hansen@linux.intel.com>
CC: H. Peter Anvin <hpa@zytor.com>
CC: Ingo Molnar <mingo@redhat.com>
CC: Jim Mattson <jmattson@google.com>
CC: Joerg Roedel <joro@8bytes.org>
CC: Paolo Bonzini <pbonzini@redhat.com>
CC: Sean Christopherson <seanjc@google.com>
CC: Thomas Gleixner <tglx@linutronix.de>
CC: Vitaly Kuznetsov <vkuznets@redhat.com>
CC: Wanpeng Li <wanpengli@tencent.com>
CC: kvm@vger.kernel.org
CC: linux-kernel@vger.kernel.org
CC: x86@kernel.org
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Yury Norov <yury.norov@gmail.com>
Message-Id: <20220519171504.1238724-1-yury.norov@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agoKVM: Free new dirty bitmap if creating a new memslot fails
Sean Christopherson [Wed, 18 May 2022 00:38:42 +0000 (00:38 +0000)]
KVM: Free new dirty bitmap if creating a new memslot fails

Fix a goof in kvm_prepare_memory_region() where KVM fails to free the
new memslot's dirty bitmap during a CREATE action if
kvm_arch_prepare_memory_region() fails.  The logic is supposed to detect
if the bitmap was allocated and thus needs to be freed, versus if the
bitmap was inherited from the old memslot and thus needs to be kept.  If
there is no old memslot, then obviously the bitmap can't have been
inherited

The bug was exposed by commit 86931ff7207b ("KVM: x86/mmu: Do not create
SPTEs for GFNs that exceed host.MAXPHYADDR"), which made it trivally easy
for syzkaller to trigger failure during kvm_arch_prepare_memory_region(),
but the bug can be hit other ways too, e.g. due to -ENOMEM when
allocating x86's memslot metadata.

The backtrace from kmemleak:

  __vmalloc_node_range+0xb40/0xbd0 mm/vmalloc.c:3195
  __vmalloc_node mm/vmalloc.c:3232 [inline]
  __vmalloc+0x49/0x50 mm/vmalloc.c:3246
  __vmalloc_array mm/util.c:671 [inline]
  __vcalloc+0x49/0x70 mm/util.c:694
  kvm_alloc_dirty_bitmap virt/kvm/kvm_main.c:1319
  kvm_prepare_memory_region virt/kvm/kvm_main.c:1551
  kvm_set_memslot+0x1bd/0x690 virt/kvm/kvm_main.c:1782
  __kvm_set_memory_region+0x689/0x750 virt/kvm/kvm_main.c:1949
  kvm_set_memory_region virt/kvm/kvm_main.c:1962
  kvm_vm_ioctl_set_memory_region virt/kvm/kvm_main.c:1974
  kvm_vm_ioctl+0x377/0x13a0 virt/kvm/kvm_main.c:4528
  vfs_ioctl fs/ioctl.c:51
  __do_sys_ioctl fs/ioctl.c:870
  __se_sys_ioctl fs/ioctl.c:856
  __x64_sys_ioctl+0xfc/0x140 fs/ioctl.c:856
  do_syscall_x64 arch/x86/entry/common.c:50
  do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
  entry_SYSCALL_64_after_hwframe+0x44/0xae

And the relevant sequence of KVM events:

  ioctl(3, KVM_CREATE_VM, 0)              = 4
  ioctl(4, KVM_SET_USER_MEMORY_REGION, {slot=0,
                                        flags=KVM_MEM_LOG_DIRTY_PAGES,
                                        guest_phys_addr=0x10000000000000,
                                        memory_size=4096,
                                        userspace_addr=0x20fe8000}
       ) = -1 EINVAL (Invalid argument)

Fixes: 244893fa2859 ("KVM: Dynamically allocate "new" memslots from the get-go")
Cc: stable@vger.kernel.org
Reported-by: syzbot+8606b8a9cc97a63f1c87@syzkaller.appspotmail.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20220518003842.1341782-1-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agogpio: mvebu/pwm: Refuse requests with inverted polarity
Uwe Kleine-König [Wed, 11 May 2022 07:58:56 +0000 (09:58 +0200)]
gpio: mvebu/pwm: Refuse requests with inverted polarity

The driver doesn't take struct pwm_state::polarity into account when
configuring the hardware, so refuse requests for inverted polarity.

Fixes: 757642f9a584 ("gpio: mvebu: Add limited PWM support")
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Bartosz Golaszewski <brgl@bgdev.pl>
3 years agogpio: gpio-vf610: do not touch other bits when set the target bit
Haibo Chen [Wed, 11 May 2022 02:15:04 +0000 (10:15 +0800)]
gpio: gpio-vf610: do not touch other bits when set the target bit

For gpio controller contain register PDDR, when set one target bit,
current logic will clear all other bits, this is wrong. Use operator
'|=' to fix it.

Fixes: 659d8a62311f ("gpio: vf610: add imx7ulp support")
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Signed-off-by: Bartosz Golaszewski <brgl@bgdev.pl>
3 years agoperf stat: Fix and validate CPU map inputs in synthetic PERF_RECORD_STAT events
Ian Rogers [Thu, 19 May 2022 03:20:01 +0000 (20:20 -0700)]
perf stat: Fix and validate CPU map inputs in synthetic PERF_RECORD_STAT events

Stat events can come from disk and so need a degree of validation. They
contain a CPU which needs looking up via CPU map to access a counter.

Add the CPU to index translation, alongside validity checking.

Discussion thread:

  https://lore.kernel.org/linux-perf-users/CAP-5=fWQR=sCuiSMktvUtcbOLidEpUJLCybVF6=BRvORcDOq+g@mail.gmail.com/

Fixes: 7ac0089d138f80dc ("perf evsel: Pass cpu not cpu map index to synthesize")
Reported-by: Michael Petlan <mpetlan@redhat.com>
Suggested-by: Michael Petlan <mpetlan@redhat.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Andrii Nakryiko <andrii@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Dave Marchevsky <davemarchevsky@fb.com>
Cc: Ian Rogers <irogers@google.com>
Cc: James Clark <james.clark@arm.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: KP Singh <kpsingh@kernel.org>
Cc: Lv Ruyi <lv.ruyi@zte.com.cn>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Martin KaFai Lau <kafai@fb.com>
Cc: Michael Petlan <mpetlan@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: netdev@vger.kernel.org
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Quentin Monnet <quentin@isovalent.com>
Cc: Song Liu <songliubraving@fb.com>
Cc: Stephane Eranian <eranian@google.com>
Cc: Xing Zhengjun <zhengjun.xing@linux.intel.com>
Cc: Yonghong Song <yhs@fb.com>
Link: http://lore.kernel.org/lkml/20220519032005.1273691-2-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
3 years agoKVM: eventfd: Fix false positive RCU usage warning
Wanpeng Li [Thu, 19 May 2022 08:49:13 +0000 (01:49 -0700)]
KVM: eventfd: Fix false positive RCU usage warning

The splat below can be seen when running kvm-unit-test:

     =============================
     WARNING: suspicious RCU usage
     5.18.0-rc7 #5 Tainted: G          IOE
     -----------------------------
     /home/kernel/linux/arch/x86/kvm/../../../virt/kvm/eventfd.c:80 RCU-list traversed in non-reader section!!

     other info that might help us debug this:

     rcu_scheduler_active = 2, debug_locks = 1
     4 locks held by qemu-system-x86/35124:
      #0: ffff9725391d80b8 (&vcpu->mutex){+.+.}-{4:4}, at: kvm_vcpu_ioctl+0x77/0x710 [kvm]
      #1: ffffbd25cfb2a0b8 (&kvm->srcu){....}-{0:0}, at: vcpu_enter_guest+0xdeb/0x1900 [kvm]
      #2: ffffbd25cfb2b920 (&kvm->irq_srcu){....}-{0:0}, at: kvm_hv_notify_acked_sint+0x79/0x1e0 [kvm]
      #3: ffffbd25cfb2b920 (&kvm->irq_srcu){....}-{0:0}, at: irqfd_resampler_ack+0x5/0x110 [kvm]

     stack backtrace:
     CPU: 2 PID: 35124 Comm: qemu-system-x86 Tainted: G          IOE     5.18.0-rc7 #5
     Call Trace:
      <TASK>
      dump_stack_lvl+0x6c/0x9b
      irqfd_resampler_ack+0xfd/0x110 [kvm]
      kvm_notify_acked_gsi+0x32/0x90 [kvm]
      kvm_hv_notify_acked_sint+0xc5/0x1e0 [kvm]
      kvm_hv_set_msr_common+0xec1/0x1160 [kvm]
      kvm_set_msr_common+0x7c3/0xf60 [kvm]
      vmx_set_msr+0x394/0x1240 [kvm_intel]
      kvm_set_msr_ignored_check+0x86/0x200 [kvm]
      kvm_emulate_wrmsr+0x4f/0x1f0 [kvm]
      vmx_handle_exit+0x6fb/0x7e0 [kvm_intel]
      vcpu_enter_guest+0xe5a/0x1900 [kvm]
      kvm_arch_vcpu_ioctl_run+0x16e/0xac0 [kvm]
      kvm_vcpu_ioctl+0x279/0x710 [kvm]
      __x64_sys_ioctl+0x83/0xb0
      do_syscall_64+0x3b/0x90
      entry_SYSCALL_64_after_hwframe+0x44/0xae

resampler-list is protected by irq_srcu (see kvm_irqfd_assign), so fix
the false positive by using list_for_each_entry_srcu().

Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
Message-Id: <1652950153-12489-1-git-send-email-wanpengli@tencent.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agoperf build: Fix check for btf__load_from_kernel_by_id() in libbpf
Arnaldo Carvalho de Melo [Fri, 20 May 2022 00:25:12 +0000 (21:25 -0300)]
perf build: Fix check for btf__load_from_kernel_by_id() in libbpf

Avi Kivity reported a problem where the __weak
btf__load_from_kernel_by_id() in tools/perf/util/bpf-event.c was being
used and it called btf__get_from_id() in tools/lib/bpf/btf.c that in
turn called back to btf__load_from_kernel_by_id(), resulting in an
endless loop.

Fix this by adding a feature test to check if
btf__load_from_kernel_by_id() is available when building perf with
LIBBPF_DYNAMIC=1, and if not then provide the fallback to the old
btf__get_from_id(), that doesn't call back to btf__load_from_kernel_by_id()
since at that time it didn't exist at all.

Tested on Fedora 35 where we have libbpf-devel 0.4.0 with LIBBPF_DYNAMIC
where we don't have btf__load_from_kernel_by_id() and thus its feature
test fail, not defining HAVE_LIBBPF_BTF__LOAD_FROM_KERNEL_BY_ID:

  $ cat /tmp/build/perf-urgent/feature/test-libbpf-btf__load_from_kernel_by_id.make.output
  test-libbpf-btf__load_from_kernel_by_id.c: In function ‘main’:
  test-libbpf-btf__load_from_kernel_by_id.c:6:16: error: implicit declaration of function ‘btf__load_from_kernel_by_id’ [-Werror=implicit-function-declaration]
      6 |         return btf__load_from_kernel_by_id(20151128, NULL);
        |                ^~~~~~~~~~~~~~~~~~~~~~~~~~~
  cc1: all warnings being treated as errors
  $

  $ nm /tmp/build/perf-urgent/perf | grep btf__load_from_kernel_by_id
  00000000005ba180 T btf__load_from_kernel_by_id
  $

  $ objdump --disassemble=btf__load_from_kernel_by_id -S /tmp/build/perf-urgent/perf

  /tmp/build/perf-urgent/perf:     file format elf64-x86-64
  <SNIP>
  00000000005ba180 <btf__load_from_kernel_by_id>:
  #include "record.h"
  #include "util/synthetic-events.h"

  #ifndef HAVE_LIBBPF_BTF__LOAD_FROM_KERNEL_BY_ID
  struct btf *btf__load_from_kernel_by_id(__u32 id)
  {
    5ba180: 55                    push   %rbp
    5ba181: 48 89 e5              mov    %rsp,%rbp
    5ba184: 48 83 ec 10           sub    $0x10,%rsp
    5ba188: 64 48 8b 04 25 28 00  mov    %fs:0x28,%rax
    5ba18f: 00 00
    5ba191: 48 89 45 f8           mov    %rax,-0x8(%rbp)
    5ba195: 31 c0                 xor    %eax,%eax
         struct btf *btf;
  #pragma GCC diagnostic push
  #pragma GCC diagnostic ignored "-Wdeprecated-declarations"
         int err = btf__get_from_id(id, &btf);
    5ba197: 48 8d 75 f0           lea    -0x10(%rbp),%rsi
    5ba19b: e8 a0 57 e5 ff        call   40f940 <btf__get_from_id@plt>
    5ba1a0: 89 c2                 mov    %eax,%edx
  #pragma GCC diagnostic pop

         return err ? ERR_PTR(err) : btf;
    5ba1a2: 48 98                 cltq
    5ba1a4: 85 d2                 test   %edx,%edx
    5ba1a6: 48 0f 44 45 f0        cmove  -0x10(%rbp),%rax
  }
  <SNIP>

Fixes: 218e7b775d368f38 ("perf bpf: Provide a weak btf__load_from_kernel_by_id() for older libbpf versions")
Reported-by: Avi Kivity <avi@scylladb.com>
Link: https://lore.kernel.org/linux-perf-users/f0add43b-3de5-20c5-22c4-70aff4af959f@scylladb.com
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/linux-perf-users/YobjjFOblY4Xvwo7@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
3 years agoselftests: kvm/x86: Verify the pmu event filter matches the correct event
Aaron Lewis [Tue, 17 May 2022 05:12:38 +0000 (05:12 +0000)]
selftests: kvm/x86: Verify the pmu event filter matches the correct event

Add a test to demonstrate that when the guest programs an event select
it is matched correctly in the pmu event filter and not inadvertently
filtered.  This could happen on AMD if the high nybble[1] in the event
select gets truncated away only leaving the bottom byte[2] left for
matching.

This is a contrived example used for the convenience of demonstrating
this issue, however, this can be applied to event selects 0x28A (OC
Mode Switch) and 0x08A (L1 BTB Correction), where 0x08A could end up
being denied when the event select was only set up to deny 0x28A.

[1] bits 35:32 in the event select register and bits 11:8 in the event
    select.
[2] bits 7:0 in the event select register and bits 7:0 in the event
    select.

Signed-off-by: Aaron Lewis <aaronlewis@google.com>
Message-Id: <20220517051238.2566934-3-aaronlewis@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agoselftests: kvm/x86: Add the helper function create_pmu_event_filter
Aaron Lewis [Tue, 17 May 2022 05:12:37 +0000 (05:12 +0000)]
selftests: kvm/x86: Add the helper function create_pmu_event_filter

Add a helper function that creates a pmu event filter given an event
list.  Currently, a pmu event filter can only be created with the same
hard coded event list.  Add a way to create one given a different event
list.

Also, rename make_pmu_event_filter to alloc_pmu_event_filter to clarify
it's purpose given the introduction of create_pmu_event_filter.

No functional changes intended.

Signed-off-by: Aaron Lewis <aaronlewis@google.com>
Message-Id: <20220517051238.2566934-2-aaronlewis@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agokvm: x86/pmu: Fix the compare function used by the pmu event filter
Aaron Lewis [Tue, 17 May 2022 05:12:36 +0000 (05:12 +0000)]
kvm: x86/pmu: Fix the compare function used by the pmu event filter

When returning from the compare function the u64 is truncated to an
int.  This results in a loss of the high nybble[1] in the event select
and its sign if that nybble is in use.  Switch from using a result that
can end up being truncated to a result that can only be: 1, 0, -1.

[1] bits 35:32 in the event select register and bits 11:8 in the event
    select.

Fixes: 7ff775aca48ad ("KVM: x86/pmu: Use binary search to check filtered events")
Signed-off-by: Aaron Lewis <aaronlewis@google.com>
Reviewed-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20220517051238.2566934-1-aaronlewis@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agoMerge tag 'v5.18-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Fri, 20 May 2022 06:04:17 +0000 (20:04 -1000)]
Merge tag 'v5.18-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6

Pull crypto fix from Herbert Xu:
 "Fix a regression in a recent fix to qcom-rng"

* tag 'v5.18-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: qcom-rng - fix infinite loop on requests not multiple of WORD_SZ

3 years agoscsi: ufs: core: Fix referencing invalid rsp field
Daejun Park [Thu, 19 May 2022 06:05:29 +0000 (15:05 +0900)]
scsi: ufs: core: Fix referencing invalid rsp field

Fix referencing sense data when it is invalid. When the length of the data
segment is 0, there is no valid information in the rsp field, so
ufshpb_rsp_upiu() is returned without additional operation.

Link: https://lore.kernel.org/r/252651381.41652940482659.JavaMail.epsvc@epcpadp4
Fixes: 4b5f49079c52 ("scsi: ufs: ufshpb: L2P map management for HPB read")
Acked-by: Avri Altman <avri.altman@wdc.com>
Signed-off-by: Daejun Park <daejun7.park@samsung.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
3 years agoriscv: dts: microchip: fix gpio1 reg property typo
Conor Paxton [Tue, 17 May 2022 10:40:58 +0000 (11:40 +0100)]
riscv: dts: microchip: fix gpio1 reg property typo

Fix reg address typo in the gpio1 stanza.

Signed-off-by: Conor Paxton <conor.paxton@microchip.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
Fixes: 528a5b1f2556 ("riscv: dts: microchip: add new peripherals to icicle kit device tree")
Link: https://lore.kernel.org/r/20220517104058.2004734-1-conor.paxton@microchip.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
3 years agoriscv: dts: sifive: fu540-c000: align dma node name with dtschema
Krzysztof Kozlowski [Thu, 7 Apr 2022 19:38:56 +0000 (21:38 +0200)]
riscv: dts: sifive: fu540-c000: align dma node name with dtschema

Fixes dtbs_check warnings like:

  dma@3000000: $nodename:0: 'dma@3000000' does not match '^dma-controller(@.*)?$'

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20220407193856.18223-1-krzysztof.kozlowski@linaro.org
Fixes: c5ab54e9945b ("riscv: dts: add support for PDMA device of HiFive Unleashed Rev A00")
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
3 years agoMerge tag 'for-5.18/parisc-4' of git://git.kernel.org/pub/scm/linux/kernel/git/deller...
Linus Torvalds [Thu, 19 May 2022 16:10:09 +0000 (06:10 -1000)]
Merge tag 'for-5.18/parisc-4' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux

Pull parisc architecture fixes from Helge Deller:
 "We had two big outstanding issues after v5.18-rc6:

   a) 32-bit kernels on 64-bit machines (e.g. on a C3700 which is able
      to run 32- and 64-bit kernels) failed early in userspace.

   b) 64-bit kernels on PA8800/PA8900 CPUs (e.g. in a C8000) showed
      random userspace segfaults. We assumed that those problems were
      caused by the tmpalias flushes.

  Dave did a lot of testing and reorganization of the current flush code
  and fixed the 32-bit cache flushing. For PA8800/PA8900 CPUs he
  switched the code to flush using the virtual address of user and
  kernel pages instead of using tmpalias flushes. The tmpalias flushes
  don't seem to work reliable on such CPUs.

  We tested the patches on a wide range machines (715/64, B160L, C3000,
  C3700, C8000, rp3440) and they have been in for-next without any
  conflicts.

  Summary:

   - Rewrite the cache flush code for PA8800/PA8900 CPUs to flush using
     the virtual address of user and kernel pages instead of using
     tmpalias flushes. Testing showed, that tmpalias flushes don't work
     reliably on PA8800/PA8900 CPUs

   - Fix flush code to allow 32-bit kernels to run on 64-bit capable
     machines, e.g. a 32-bit kernel on C3700 machines"

* tag 'for-5.18/parisc-4' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: Fix patch code locking and flushing
  parisc: Rewrite cache flush code for PA8800/PA8900
  parisc: Disable debug code regarding cache flushes in handle_nadtlb_fault()

3 years agoMerge tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm
Linus Torvalds [Thu, 19 May 2022 16:08:29 +0000 (06:08 -1000)]
Merge tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm

Pull ARM fixes from Russell King:
 "Two further fixes for Spectre-BHB from Ard for Cortex A15 and to use
  the wide branch instruction for Thumb2"

* tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm:
  ARM: 9197/1: spectre-bhb: fix loop8 sequence for Thumb2
  ARM: 9196/1: spectre-bhb: enable for Cortex-A15

3 years agoMerge tag 'pinctrl-v5.18-3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw...
Linus Torvalds [Thu, 19 May 2022 16:02:41 +0000 (06:02 -1000)]
Merge tag 'pinctrl-v5.18-3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl

Pull pin control fixes from Linus Walleij:

 - Fix an altmode in the Ocelot driver

 - Fix the IES control pins in the Mediatek MT8365 driver

 - Sunxi (AMLogic) driver:
    - Fix the UART2 function pin assignments
    - Fix the signal name of the PA2 SPI pin

* tag 'pinctrl-v5.18-3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:
  pinctrl: sunxi: f1c100s: Fix signal name comment for PA2 SPI pin
  pinctrl: sunxi: fix f1c100s uart2 function
  pinctrl: mediatek: mt8365: fix IES control pins
  pinctrl: ocelot: Fix for lan966x alt mode

3 years agoMerge tag 'net-5.18-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 19 May 2022 15:50:29 +0000 (05:50 -1000)]
Merge tag 'net-5.18-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

Pull networking fixes from Paolo Abeni:
 "Including fixes from can, xfrm and netfilter subtrees.

  Notably this reverts a recent TCP/DCCP netns-related change to address
  a possible UaF.

  Current release - regressions:

   - tcp: revert "tcp/dccp: get rid of inet_twsk_purge()"

   - xfrm: set dst dev to blackhole_netdev instead of loopback_dev in
     ifdown

  Previous releases - regressions:

   - netfilter: flowtable: fix TCP flow teardown

   - can: revert "can: m_can: pci: use custom bit timings for Elkhart
     Lake"

   - xfrm: check encryption module availability consistency

   - eth: vmxnet3: fix possible use-after-free bugs in
     vmxnet3_rq_alloc_rx_buf()

   - eth: mlx5: initialize flow steering during driver probe

   - eth: ice: fix crash when writing timestamp on RX rings

  Previous releases - always broken:

   - mptcp: fix checksum byte order

   - eth: lan966x: fix assignment of the MAC address

   - eth: mlx5: remove HW-GRO from reported features

   - eth: ftgmac100: disable hardware checksum on AST2600"

* tag 'net-5.18-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (50 commits)
  net: bridge: Clear offload_fwd_mark when passing frame up bridge interface.
  ptp: ocp: change sysfs attr group handling
  selftests: forwarding: fix missing backslash
  netfilter: nf_tables: disable expression reduction infra
  netfilter: flowtable: move dst_check to packet path
  netfilter: flowtable: fix TCP flow teardown
  net: ftgmac100: Disable hardware checksum on AST2600
  igb: skip phy status check where unavailable
  nfc: pn533: Fix buggy cleanup order
  mptcp: Do TCP fallback on early DSS checksum failure
  mptcp: fix checksum byte order
  net: af_key: check encryption module availability consistency
  net: af_key: add check for pfkey_broadcast in function pfkey_process
  net/mlx5: Drain fw_reset when removing device
  net/mlx5e: CT: Fix setting flow_source for smfs ct tuples
  net/mlx5e: CT: Fix support for GRE tuples
  net/mlx5e: Remove HW-GRO from reported features
  net/mlx5e: Properly block HW GRO when XDP is enabled
  net/mlx5e: Properly block LRO when XDP is enabled
  net/mlx5e: Block rx-gro-hw feature in switchdev mode
  ...

3 years agommc: core: Fix busy polling for MMC_SEND_OP_COND again
Ulf Hansson [Tue, 17 May 2022 10:10:46 +0000 (12:10 +0200)]
mmc: core: Fix busy polling for MMC_SEND_OP_COND again

It turned out that polling period for MMC_SEND_OP_COND, that currently is
set to 1ms, still isn't sufficient. In particular a Micron eMMC on a
Beaglebone platform, is reported to sometimes fail to initialize.

Additional test, shows that extending the period to 4ms is working fine, so
let's make that change.

Reported-by: Jean Rene Dawin <jdawin@math.uni-bielefeld.de>
Tested-by: Jean Rene Dawin <jdawin@math.uni-bielefeld.de>
Fixes: 1760fdb6fe9f (mmc: core: Restore (almost) the busy polling for MMC_SEND_OP_COND")
Fixes: 76bfc7ccc2fa ("mmc: core: adjust polling interval for CMD1")
Cc: stable@vger.kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20220517101046.27512-1-ulf.hansson@linaro.org
3 years agodrm/i915: Use i915_gem_object_ggtt_pin_ww for reloc_iomap
Maarten Lankhorst [Wed, 11 May 2022 11:52:19 +0000 (13:52 +0200)]
drm/i915: Use i915_gem_object_ggtt_pin_ww for reloc_iomap

When removing short term pins, I've changed the the batch buffer
pinning for relocation to use __i915_vma_pin, because
i915_gem_object_ggtt_pin_ww was destroying the old vma. This
caused regressions, because the functions are not identical.

Fix the regressions by calling i915_gem_object_ggtt_pin_ww() again
on ggtt-only platforms, but only if the batch can be pinned without
being moved.

Fixes: b5cfe6f7a6e1 ("drm/i915: Remove short-term pins from execbuf, v6.")
Cc: Matthew Auld <matthew.auld@intel.com>
Reported-by: Mateusz Jończyk <mat.jonczyk@o2.pl>
Tested-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Acked-by: Matthew Auld <matthew.auld@intel.com>
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/5806
Link: https://patchwork.freedesktop.org/patch/msgid/20220511115219.46507-1-maarten.lankhorst@linux.intel.com
(cherry picked from commit 451374eef622fca6f00eeeda89aaccb45a30a149)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
3 years agonet: bridge: Clear offload_fwd_mark when passing frame up bridge interface.
Andrew Lunn [Wed, 18 May 2022 00:58:40 +0000 (02:58 +0200)]
net: bridge: Clear offload_fwd_mark when passing frame up bridge interface.

It is possible to stack bridges on top of each other. Consider the
following which makes use of an Ethernet switch:

       br1
     /    \
    /      \
   /        \
 br0.11    wlan0
   |
   br0
 /  |  \
p1  p2  p3

br0 is offloaded to the switch. Above br0 is a vlan interface, for
vlan 11. This vlan interface is then a slave of br1. br1 also has a
wireless interface as a slave. This setup trunks wireless lan traffic
over the copper network inside a VLAN.

A frame received on p1 which is passed up to the bridge has the
skb->offload_fwd_mark flag set to true, indicating that the switch has
dealt with forwarding the frame out ports p2 and p3 as needed. This
flag instructs the software bridge it does not need to pass the frame
back down again. However, the flag is not getting reset when the frame
is passed upwards. As a result br1 sees the flag, wrongly interprets
it, and fails to forward the frame to wlan0.

When passing a frame upwards, clear the flag. This is the Rx
equivalent of br_switchdev_frame_unmark() in br_dev_xmit().

Fixes: f1c2eddf4cb6 ("bridge: switchdev: Use an helper to clear forward mark")
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Ido Schimmel <idosch@nvidia.com>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Link: https://lore.kernel.org/r/20220518005840.771575-1-andrew@lunn.ch
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
3 years agoptp: ocp: change sysfs attr group handling
Jonathan Lemon [Tue, 17 May 2022 21:46:00 +0000 (14:46 -0700)]
ptp: ocp: change sysfs attr group handling

In the detach path, the driver calls sysfs_remove_group() for the
groups it believes has been registered.  However, if the group was
never previously registered, then this causes a splat.

Instead, compute the groups that should be registered in advance,
and then call sysfs_create_groups(), which registers them all at once.

Update the error handling appropriately.

Fixes: c205d53c4923 ("ptp: ocp: Add firmware capability bits for feature gating")
Reported-by: Zheyu Ma <zheyuma97@gmail.com>
Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Link: https://lore.kernel.org/r/20220517214600.10606-1-jonathan.lemon@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge tag 'amd-drm-fixes-5.18-2022-05-18' of https://gitlab.freedesktop.org/agd5f...
Dave Airlie [Thu, 19 May 2022 04:03:02 +0000 (14:03 +1000)]
Merge tag 'amd-drm-fixes-5.18-2022-05-18' of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes

amd-drm-fixes-5.18-2022-05-18:

amdgpu:
- Suspend/resume regression fix

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220518202045.9123-1-alexander.deucher@amd.com
3 years agoselftests: forwarding: fix missing backslash
Joachim Wiberg [Wed, 18 May 2022 15:16:30 +0000 (17:16 +0200)]
selftests: forwarding: fix missing backslash

Fix missing backslash, introduced in f62c5acc800ee.  Causes all tests to
not be installed.

Fixes: f62c5acc800e ("selftests/net/forwarding: add missing tests to Makefile")
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Acked-by: Hangbin Liu <liuhangbin@gmail.com>
Link: https://lore.kernel.org/r/20220518151630.2747773-1-troglobit@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf
Jakub Kicinski [Thu, 19 May 2022 02:34:25 +0000 (19:34 -0700)]
Merge git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf

Pablo Neira Ayuso says:

====================
Netfilter fixes for net

1) Reduce number of hardware offload retries from flowtable datapath
   which might hog system with retries, from Felix Fietkau.

2) Skip neighbour lookup for PPPoE device, fill_forward_path() already
   provides this and set on destination address from fill_forward_path for
   PPPoE device, also from Felix.

4) When combining PPPoE on top of a VLAN device, set info->outdev to the
   PPPoE device so software offload works, from Felix.

5) Fix TCP teardown flowtable state, races with conntrack gc might result
   in resetting the state to ESTABLISHED and the time to one day. Joint
   work with Oz Shlomo and Sven Auhagen.

6) Call dst_check() from flowtable datapath to check if dst is stale
   instead of doing it from garbage collector path.

7) Disable register tracking infrastructure, either user-space or
   kernel need to pre-fetch keys inconditionally, otherwise register
   tracking assumes data is already available in register that might
   not well be there, leading to incorrect reductions.

* git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
  netfilter: nf_tables: disable expression reduction infra
  netfilter: flowtable: move dst_check to packet path
  netfilter: flowtable: fix TCP flow teardown
  netfilter: nft_flow_offload: fix offload with pppoe + vlan
  net: fix dev_fill_forward_path with pppoe + bridge
  netfilter: nft_flow_offload: skip dst neigh lookup for ppp devices
  netfilter: flowtable: fix excessive hw offload attempts after failure
====================

Link: https://lore.kernel.org/r/20220518213841.359653-1-pablo@netfilter.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge tag 'block-5.18-2022-05-18' of git://git.kernel.dk/linux-block
Linus Torvalds [Thu, 19 May 2022 00:32:27 +0000 (14:32 -1000)]
Merge tag 'block-5.18-2022-05-18' of git://git.kernel.dk/linux-block

Pull block fix from Jens Axboe:
 "Just a small fix for a missing fifo time assigment for the head
  insertion case in mq-deadline"

* tag 'block-5.18-2022-05-18' of git://git.kernel.dk/linux-block:
  block/mq-deadline: Set the fifo_time member also if inserting at head