We need to use test_and_set_bit() when changing xprt state flags to
avoid potentially getting xps->xps_nactive out of sync.
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
The refcount leak issues take place in an error handling path. When the
3rd argument buf doesn't match with "offline", "online" or "remove", the
function simply returns -EINVAL and forgets to decrease the reference
count of a rpc_xprt object and a rpc_xprt_switch object increased by
rpc_sysfs_xprt_kobj_get_xprt() and
rpc_sysfs_xprt_kobj_get_xprt_switch(), causing reference count leaks of
both unused objects.
Fix this issue by jumping to the error handling path labelled with
out_put when buf matches none of "offline", "online" or "remove".
Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn>
Signed-off-by: Xin Xiong <xiongx18@fudan.edu.cn>
Signed-off-by: Xin Tan <tanxin.ctf@gmail.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
For each location returned in FS_LOCATION query, establish a
transport to the server, send EXCHANGE_ID and test for trunking,
if successful, add the transport to the exiting client.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
If the supplied argument doesn't specify the transport type, use the
type of the existing rpc clnt and its existing transport.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
An fs_location attribute returns a string that can be ipv4, ipv6,
or DNS name. An ip location can have a port appended to it and if
no port is present a default port needs to be set. If rpc_pton()
fails to parse, try calling rpc_uaddr2socaddr() that can convert
an universal address.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Make nfs_parse_server_name available outside of nfs4namespace.c.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Query the server for other possible trunkable locations for a given
file system on a 4.1+ mount.
v2:
-- added missing static to nfs4_discover_trunking,
reported by the kernel test robot
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Define and store if server returns it supports fs_locations attribute
as a capability.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Remove the check for the zero length fs_locations reply in the
xdr decoding, and instead check for that in the migration code.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
If the file type changes back to being a regular file on the server
between the failed OPEN and our LOOKUP, then we need to re-run the OPEN.
Fixes: 0dd2b474d0 ("nfs: implement i_op->atomic_open()")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
If the application sets the O_DIRECTORY flag, and tries to open a
regular file, nfs_atomic_open() will punt to doing a regular lookup.
If the server then returns a regular file, we will happily return a
file descriptor with uninitialised open state.
The fix is to return the expected ENOTDIR error in these cases.
Reported-by: Lyu Tao <tao.lyu@epfl.ch>
Fixes: 0dd2b474d0 ("nfs: implement i_op->atomic_open()")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
When doing a non-pNFS write, allow the writeback code to specify that it
also needs to update 'blocks used'.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
There are currently 2 ways to create a set of sysfs files for a
kobj_type, through the default_attrs field, and the default_groups
field. Move the NFS code to use default_groups field which has been the
preferred way since aa30f47cf6 ("kobject: Add support for default
attribute groups to kobj_type") so that we can soon get rid of the
obsolete default_attrs field.
Cc: Trond Myklebust <trond.myklebust@hammerspace.com>
Cc: Anna Schumaker <anna.schumaker@netapp.com>
Cc: linux-nfs@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
When dealing with case insensitive names, the client has no idea how the
server performs the mapping, so cannot collapse the dentries into a
single representative. So both rename and unlink need to deal with the
fact that there could be several dentries representing the file, and
have to somehow force them to be revalidated. Use d_prune_aliases() as a
big hammer approach.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
If we create a file, rename it, or hardlink it, then we need to assume
that cached negative dentries need to be revalidated.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Add capabilities to allow the NFS client to recognise when it is dealing
with case insensitive and case preserving filesystems.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
When decode_devicenotify_args() exits with no entries, we need to
ensure that the struct cb_devicenotifyargs is initialised to
{ 0, NULL } in order to avoid problems in
nfs4_callback_devicenotify().
Reported-by: <rtm@csail.mit.edu>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
kstrdup() returns NULL when some internal memory errors happen, it is
better to check the return value of it so to catch the memory error in
time.
Signed-off-by: Xiaoke Wang <xkernel.wang@foxmail.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
When the bitmask of the attributes doesn't include the security label,
don't bother printing it. Since the label might not be null terminated,
adjust the printing format accordingly.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
In g_verify_token_header, the null check of 'ret'
is unneeded to be done twice.
Signed-off-by: Xu Wang <vulab@iscas.ac.cn>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Renaming a file is required by POSIX to update the file ctime, so
ensure that the file data is synced to disk so that we don't clobber the
updated ctime by writing back after creating the hard link.
Fixes: f2c2c552f1 ("NFS: Move delegation recall into the NFSv4 callback for rename_setup()")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Creating a hard link is required by POSIX to update the file ctime, so
ensure that the file data is synced to disk so that we don't clobber the
updated ctime by writing back after creating the hard link.
Fixes: 9f76827287 ("NFS: Move the delegation return down into nfs4_proc_link()")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Storing the 'struct cred *' in nfs_access_entry is problematic.
An active 'cred' can keep a 'struct key *' active, and a quota is
imposed on the number of such keys that a user can maintain.
Cached 'nfs_access_entry' structs have indefinite lifetime, and having
these keep 'struct key's alive imposes on that quota.
So remove the 'struct cred *' and replace it with the fields we need:
kuid_t, kgid_t, and struct group_info *
This makes the 'struct nfs_access_entry' 64 bits larger.
New function "access_cmp" is introduced which is identical to
cred_fscmp() except that the second arg is an 'nfs_access_entry', rather
than a 'cred'
Fixes: b68572e07c ("NFS: change access cache to use 'struct cred'.")
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Storing the 'struct cred *' in nfs_access_entry is problematic.
An active 'cred' can keep a 'struct key *' active, and a quota is
imposed on the number of such keys that a user can maintain.
Cached 'nfs_access_entry' structs have indefinite lifetime, and having
these keep 'struct key's alive imposes on that quota.
So a future patch will remove the ->cred ref from nfs_access_entry.
To prepare, change various functions to not assume there is a 'cred' in
the nfs_access_entry, but to pass the cred around explicitly.
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Currently the nfs_access_get_cached family of functions report a
'struct nfs_access_entry' as the result, with both .mask and .cred set.
However the .cred is never used. This is probably good and there is no
guarantee that it won't be freed before use.
Change to only report the 'mask' - as this is all that is used or needed.
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Pull perf tools fixes from Arnaldo Carvalho de Melo:
- Fix TUI exit screen refresh race condition in 'perf top'.
- Fix parsing of Intel PT VM time correlation arguments.
- Honour CPU filtering command line request of a script's switch events
in 'perf script'.
- Fix printing of switch events in Intel PT python script.
- Fix duplicate alias events list printing in 'perf list', noticed on
heterogeneous arm64 systems.
- Fix return value of ids__new(), users expect NULL for failure, not
ERR_PTR(-ENOMEM).
* tag 'perf-tools-fixes-for-v5.16-2022-01-02' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux:
perf top: Fix TUI exit screen refresh race condition
perf pmu: Fix alias events list
perf scripts python: intel-pt-events.py: Fix printing of switch events
perf script: Fix CPU filtering of a script's switch events
perf intel-pt: Fix parsing of VM time correlation arguments
perf expr: Fix return value of ids__new()
Pull i2c fixes from Wolfram Sang:
"Better input validation for compat ioctls and a documentation bugfix
for 5.16"
* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
Docs: Fixes link to I2C specification
i2c: validate user data in compat ioctl
Pull x86 fix from Borislav Petkov:
- Use the proper CONFIG symbol in a preprocessor check.
* tag 'x86_urgent_for_v5.16_rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/build: Use the proper name CONFIG_FW_LOADER
When the following command is executed several times, a coredump file is
generated.
$ timeout -k 9 5 perf top -e task-clock
*******
*******
*******
0.01% [kernel] [k] __do_softirq
0.01% libpthread-2.28.so [.] __pthread_mutex_lock
0.01% [kernel] [k] __ll_sc_atomic64_sub_return
double free or corruption (!prev) perf top --sort comm,dso
timeout: the monitored command dumped core
When we terminate "perf top" using sending signal method,
SLsmg_reset_smg() called. SLsmg_reset_smg() resets the SLsmg screen
management routines by freeing all memory allocated while it was active.
However SLsmg_reinit_smg() maybe be called by another thread.
SLsmg_reinit_smg() will free the same memory accessed by
SLsmg_reset_smg(), thus it results in a double free.
SLsmg_reinit_smg() is called already protected by ui__lock, so we fix
the problem by adding pthread_mutex_trylock of ui__lock when calling
SLsmg_reset_smg().
Signed-off-by: Wenyu Liu <liuwenyu7@huawei.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: wuxu.wu@huawei.com
Link: http://lore.kernel.org/lkml/a91e3943-7ddc-f5c0-a7f5-360f073c20e6@huawei.com
Signed-off-by: Hewenliang <hewenliang4@huawei.com>
Signed-off-by: yaowenbin <yaowenbin1@huawei.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Commit 0e0ae87422 ("perf list: Display hybrid PMU events with cpu
type") changes the event list for uncore PMUs or arm64 heterogeneous CPU
systems, such that duplicate aliases are incorrectly listed per PMU
(which they should not be), like:
# perf list
...
unc_cbo_cache_lookup.any_es
[Unit: uncore_cbox L3 Lookup any request that access cache and found
line in E or S-state]
unc_cbo_cache_lookup.any_es
[Unit: uncore_cbox L3 Lookup any request that access cache and found
line in E or S-state]
unc_cbo_cache_lookup.any_i
[Unit: uncore_cbox L3 Lookup any request that access cache and found
line in I-state]
unc_cbo_cache_lookup.any_i
[Unit: uncore_cbox L3 Lookup any request that access cache and found
line in I-state]
...
Notice how the events are listed twice.
The named commit changed how we remove duplicate events, in that events
for different PMUs are not treated as duplicates. I suppose this is to
handle how "Each hybrid pmu event has been assigned with a pmu name".
Fix PMU alias listing by restoring behaviour to remove duplicates for
non-hybrid PMUs.
Fixes: 0e0ae87422 ("perf list: Display hybrid PMU events with cpu type")
Signed-off-by: John Garry <john.garry@huawei.com>
Tested-by: Zhengjun Xing <zhengjun.xing@linux.intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/1640103090-140490-1-git-send-email-john.garry@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Pull input fixes from Dmitry Torokhov:
"Two small fixups for spaceball joystick driver and appletouch touchpad
driver"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
Input: spaceball - fix parsing of movement data packets
Input: appletouch - initialize work before device registration
Hugh Dickins reported the following
My tmpfs swapping load (tweaked to use huge pages more heavily
than in real life) is far from being a realistic load: but it was
notably slowed down by your throttling mods in 5.16-rc, and this
patch makes it well again - thanks.
But: it very quickly hit NULL pointer until I changed that last
line to
if (first_pgdat)
consider_reclaim_throttle(first_pgdat, sc);
The likely issue is that huge pages are a major component of the test
workload. When this is the case, first_pgdat may never get set if
compaction is ready to continue due to this check
if (IS_ENABLED(CONFIG_COMPACTION) &&
sc->order > PAGE_ALLOC_COSTLY_ORDER &&
compaction_ready(zone, sc)) {
sc->compaction_ready = true;
continue;
}
If this was true for every zone in the zonelist, first_pgdat would never
get set resulting in a NULL pointer exception.
Link: https://lkml.kernel.org/r/20211209095453.GM3366@techsingularity.net
Fixes: 1b4e3f26f9 ("mm: vmscan: Reduce throttling due to a failure to make progress")
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Reported-by: Hugh Dickins <hughd@google.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Rik van Riel <riel@surriel.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Darrick J. Wong <djwong@kernel.org>
Cc: Shakeel Butt <shakeelb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mike Galbraith, Alexey Avramov and Darrick Wong all reported similar
problems due to reclaim throttling for excessive lengths of time. In
Alexey's case, a memory hog that should go OOM quickly stalls for
several minutes before stalling. In Mike and Darrick's cases, a small
memcg environment stalled excessively even though the system had enough
memory overall.
Commit 69392a403f ("mm/vmscan: throttle reclaim when no progress is
being made") introduced the problem although commit a19594ca4a
("mm/vmscan: increase the timeout if page reclaim is not making
progress") made it worse. Systems at or near an OOM state that cannot
be recovered must reach OOM quickly and memcg should kill tasks if a
memcg is near OOM.
To address this, only stall for the first zone in the zonelist, reduce
the timeout to 1 tick for VMSCAN_THROTTLE_NOPROGRESS and only stall if
the scan control nr_reclaimed is 0, kswapd is still active and there
were excessive pages pending for writeback. If kswapd has stopped
reclaiming due to excessive failures, do not stall at all so that OOM
triggers relatively quickly. Similarly, if an LRU is simply congested,
only lightly throttle similar to NOPROGRESS.
Alexey's original case was the most straight forward
for i in {1..3}; do tail /dev/zero; done
On vanilla 5.16-rc1, this test stalled heavily, after the patch the test
completes in a few seconds similar to 5.15.
Alexey's second test case added watching a youtube video while tail runs
10 times. On 5.15, playback only jitters slightly, 5.16-rc1 stalls a
lot with lots of frames missing and numerous audio glitches. With this
patch applies, the video plays similarly to 5.15.
[lkp@intel.com: Fix W=1 build warning]
Link: https://lore.kernel.org/r/99e779783d6c7fce96448a3402061b9dc1b3b602.camel@gmx.de
Link: https://lore.kernel.org/r/20211124011954.7cab9bb4@mail.inbox.lv
Link: https://lore.kernel.org/r/20211022144651.19914-1-mgorman@techsingularity.net
Link: https://lore.kernel.org/r/20211202150614.22440-1-mgorman@techsingularity.net
Link: https://linux-regtracking.leemhuis.info/regzbot/regression/20211124011954.7cab9bb4@mail.inbox.lv/
Reported-and-tested-by: Alexey Avramov <hakavlad@inbox.lv>
Reported-and-tested-by: Mike Galbraith <efault@gmx.de>
Reported-and-tested-by: Darrick J. Wong <djwong@kernel.org>
Reported-by: kernel test robot <lkp@intel.com>
Acked-by: Hugh Dickins <hughd@google.com>
Tracked-by: Thorsten Leemhuis <regressions@leemhuis.info>
Fixes: 69392a403f ("mm/vmscan: throttle reclaim when no progress is being made")
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge misc mm fixes from Andrew Morton:
"2 patches.
Subsystems affected by this patch series: mm (userfaultfd and damon)"
* akpm:
mm/damon/dbgfs: fix 'struct pid' leaks in 'dbgfs_target_ids_write()'
userfaultfd/selftests: fix hugetlb area allocations
Pull SCSI fixes from James Bottomley:
"Three fixes, all in drivers. The lpfc one doesn't look exploitable,
but nasty things could happen in string operations if mybuf ends up
with an on stack unterminated string"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: vmw_pvscsi: Set residual data length conditionally
scsi: libiscsi: Fix UAF in iscsi_conn_get_param()/iscsi_conn_teardown()
scsi: lpfc: Terminate string in lpfc_debugfs_nvmeio_trc_write()
DAMON debugfs interface increases the reference counts of 'struct pid's
for targets from the 'target_ids' file write callback
('dbgfs_target_ids_write()'), but decreases the counts only in DAMON
monitoring termination callback ('dbgfs_before_terminate()').
Therefore, when 'target_ids' file is repeatedly written without DAMON
monitoring start/termination, the reference count is not decreased and
therefore memory for the 'struct pid' cannot be freed. This commit
fixes this issue by decreasing the reference counts when 'target_ids' is
written.
Link: https://lkml.kernel.org/r/20211229124029.23348-1-sj@kernel.org
Fixes: 4bc05954d0 ("mm/damon: implement a debugfs-based user space interface")
Signed-off-by: SeongJae Park <sj@kernel.org>
Cc: <stable@vger.kernel.org> [5.15+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently, userfaultfd selftest for hugetlb as run from run_vmtests.sh
or any environment where there are 'just enough' hugetlb pages will
always fail with:
testing events (fork, remap, remove):
ERROR: UFFDIO_COPY error: -12 (errno=12, line=616)
The ENOMEM error code implies there are not enough hugetlb pages.
However, there are free hugetlb pages but they are all reserved. There
is a basic problem with the way the test allocates hugetlb pages which
has existed since the test was originally written.
Due to the way 'cleanup' was done between different phases of the test,
this issue was masked until recently. The issue was uncovered by commit
8ba6e86408 ("userfaultfd/selftests: reinitialize test context in each
test").
For the hugetlb test, src and dst areas are allocated as PRIVATE
mappings of a hugetlb file. This means that at mmap time, pages are
reserved for the src and dst areas. At the start of event testing (and
other tests) the src area is populated which results in allocation of
huge pages to fill the area and consumption of reserves associated with
the area. Then, a child is forked to fault in the dst area. Note that
the dst area was allocated in the parent and hence the parent owns the
reserves associated with the mapping. The child has normal access to
the dst area, but can not use the reserves created/owned by the parent.
Thus, if there are no other huge pages available allocation of a page
for the dst by the child will fail.
Fix by not creating reserves for the dst area. In this way the child
can use free (non-reserved) pages.
Also, MAP_PRIVATE of a file only makes sense if you are interested in
the contents of the file before making a COW copy. The test does not do
this. So, just use MAP_ANONYMOUS | MAP_HUGETLB to create an anonymous
hugetlb mapping. There is no need to create a hugetlb file in the
non-shared case.
Link: https://lkml.kernel.org/r/20211217172919.7861-1-mike.kravetz@oracle.com
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Axel Rasmussen <axelrasmussen@google.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Mina Almasry <almasrymina@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The link to the I2C specification is broken. Although
"https://www.nxp.com" hosts Rev 7 (2021) of this specification, it is
behind a login-wall. Thus, an additional link has been added (which
doesn't require a login) and the NXP official docs link has been
updated.
Signed-off-by: Deep Majumder <deep@fastmail.in>
[wsa: minor updates to text and commit message]
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Wrong user data may cause warning in i2c_transfer(), ex: zero msgs.
Userspace should not be able to trigger warnings, so this patch adds
validation checks for user data in compact ioctl to prevent reported
warnings
Reported-and-tested-by: syzbot+e417648b303855b91d8a@syzkaller.appspotmail.com
Fixes: 7d5cb45655 ("i2c compat ioctls: move to ->compat_ioctl()")
Signed-off-by: Pavel Skripkin <paskripkin@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
