The IPMI driver was recently modified to use SRCU, but it turns out
this uses a chunk of percpu memory, even if IPMI is never used.
So modify thing to on initialize on the first use. There was already
code to sort of handle this for handling init races, so piggy back
on top of that, and simplify it in the process.
Signed-off-by: Corey Minyard <cminyard@mvista.com> Reported-by: Tejun Heo <tj@kernel.org> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: stable@vger.kernel.org # 4.18 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some IPMI modules (e.g. ibmpex_msg_handler()) will have ipmi_usr_hdlr
handlers that call ipmi_free_recv_msg() directly. This will essentially
kfree(msg), leading to use-after-free.
This does not happen in the ipmi_devintf module, which will queue the
message and run ipmi_free_recv_msg() later.
BUG: KASAN: use-after-free in deliver_response+0x12f/0x1b0
Read of size 8 at addr ffff888a7bf20018 by task ksoftirqd/3/27
CPU: 3 PID: 27 Comm: ksoftirqd/3 Tainted: G O 4.19.11-amd64-ani99-debug #12.0.1.601133+pv
Hardware name: AppNeta r1000/X11SPW-TF, BIOS 2.1a-AP 09/17/2018
Call Trace:
dump_stack+0x92/0xeb
print_address_description+0x73/0x290
kasan_report+0x258/0x380
deliver_response+0x12f/0x1b0
? ipmi_free_recv_msg+0x50/0x50
deliver_local_response+0xe/0x50
handle_one_recv_msg+0x37a/0x21d0
handle_new_recv_msgs+0x1ce/0x440
...
Allocated by task 9885:
kasan_kmalloc+0xa0/0xd0
kmem_cache_alloc_trace+0x116/0x290
ipmi_alloc_recv_msg+0x28/0x70
i_ipmi_request+0xb4a/0x1640
ipmi_request_settime+0x1b8/0x1e0
...
Fix this by sanitizing channel and addr->channel before using them to
index user->intf->addrinfo and intf->addrinfo, correspondingly.
Notice that given that speculation windows are large, the policy is
to kill the speculation on the first load and not worry if it can be
completed with a dependent load/store [1].
Because the user->release_barrier.rda is freed in ipmi_destroy_user(), but
the refcount is not zero, when acquire_ipmi_user() uses user->release_barrier.rda
in __srcu_read_lock(), it causes oops.
Fix this by calling cleanup_srcu_struct() when the refcount is zero.
Fixes: e86ee2d44b44 ("ipmi: Rework locking and shutdown for hot remove") Cc: stable@vger.kernel.org # 4.18 Signed-off-by: Yang Yingliang <yangyingliang@huawei.com> Signed-off-by: Corey Minyard <cminyard@mvista.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In case a command which completes in Command Status was sent using the
hci_cmd_send-family of APIs there would be a misleading error in the
hci_get_cmd_complete function, since the code would be trying to fetch
the Command Complete parameters when there are none.
Avoid the misleading error and silently bail out from the function in
case the received event is a command status.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Acked-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Tested-by Adam Ford <aford173@gmail.com> #4.19.16 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The parameter that indicated whether the LQ command should be sent
as sync or async was removed, causing the LQ command to be sent as
sync from interrupt context (e.g. from the RX path). This resulted
in a kernel warning: "scheduling while atomic" and failing to send
the LQ command, which ultimately leads to a queue hang.
Fix it by adding back the required parameter to send the command as
sync only when it is allowed.
Fixes: d94c5a820d10 ("iwlwifi: mvm: open BA session only when sta is authorized") Signed-off-by: Avraham Stern <avraham.stern@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Patch series "THP eligibility reporting via proc".
This series of three patches aims at making THP eligibility reporting much
more robust and long term sustainable. The trigger for the change is a
regression report [2] and the long follow up discussion. In short the
specific application didn't have good API to query whether a particular
mapping can be backed by THP so it has used VMA flags to workaround that.
These flags represent a deep internal state of VMAs and as such they
should be used by userspace with a great deal of caution.
A similar has happened for [3] when users complained that VM_MIXEDMAP is
no longer set on DAX mappings. Again a lack of a proper API led to an
abuse.
The first patch in the series tries to emphasise that that the semantic of
flags might change and any application consuming those should be really
careful.
The remaining two patches provide a more suitable interface to address [2]
and provide a consistent API to query the THP status both for each VMA and
process wide as well. [1]
Even though vma flags exported via /proc/<pid>/smaps are explicitly
documented to be not guaranteed for future compatibility the warning
doesn't go far enough because it doesn't mention semantic changes to those
flags. And they are important as well because these flags are a deep
implementation internal to the MM code and the semantic might change at
any time.
Let's consider two recent examples:
http://lkml.kernel.org/r/20181002100531.GC4135@quack2.suse.cz
: commit e1fb4a086495 "dax: remove VM_MIXEDMAP for fsdax and device dax" has
: removed VM_MIXEDMAP flag from DAX VMAs. Now our testing shows that in the
: mean time certain customer of ours started poking into /proc/<pid>/smaps
: and looks at VMA flags there and if VM_MIXEDMAP is missing among the VMA
: flags, the application just fails to start complaining that DAX support is
: missing in the kernel.
http://lkml.kernel.org/r/alpine.DEB.2.21.1809241054050.224429@chino.kir.corp.google.com
: Commit 1860033237d4 ("mm: make PR_SET_THP_DISABLE immediately active")
: introduced a regression in that userspace cannot always determine the set
: of vmas where thp is ineligible.
: Userspace relies on the "nh" flag being emitted as part of /proc/pid/smaps
: to determine if a vma is eligible to be backed by hugepages.
: Previous to this commit, prctl(PR_SET_THP_DISABLE, 1) would cause thp to
: be disabled and emit "nh" as a flag for the corresponding vmas as part of
: /proc/pid/smaps. After the commit, thp is disabled by means of an mm
: flag and "nh" is not emitted.
: This causes smaps parsing libraries to assume a vma is eligible for thp
: and ends up puzzling the user on why its memory is not backed by thp.
In both cases userspace was relying on a semantic of a specific VMA flag.
The primary reason why that happened is a lack of a proper interface.
While this has been worked on and it will be fixed properly, it seems that
our wording could see some refinement and be more vocal about semantic
aspect of these flags as well.
Link: http://lkml.kernel.org/r/20181211143641.3503-2-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Jan Kara <jack@suse.cz> Acked-by: Dan Williams <dan.j.williams@intel.com> Acked-by: David Rientjes <rientjes@google.com> Acked-by: Mike Rapoport <rppt@linux.ibm.com> Acked-by: Vlastimil Babka <vbabka@suse.cz> Cc: Dan Williams <dan.j.williams@intel.com> Cc: David Rientjes <rientjes@google.com> Cc: Paul Oppenheimer <bepvte@gmail.com> Cc: William Kucharski <william.kucharski@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When the process being tracked does mremap() without
UFFD_FEATURE_EVENT_REMAP on the corresponding tracking uffd file handle,
we should not generate the remap event, and at the same time we should
clear all the uffd flags on the new VMA. Without this patch, we can still
have the VM_UFFD_MISSING|VM_UFFD_WP flags on the new VMA even the fault
handling process does not even know the existance of the VMA.
Link: http://lkml.kernel.org/r/20181211053409.20317-1-peterx@redhat.com Signed-off-by: Peter Xu <peterx@redhat.com> Reviewed-by: Andrea Arcangeli <aarcange@redhat.com> Acked-by: Mike Rapoport <rppt@linux.vnet.ibm.com> Reviewed-by: William Kucharski <william.kucharski@oracle.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Mike Rapoport <rppt@linux.vnet.ibm.com> Cc: Kirill A. Shutemov <kirill@shutemov.name> Cc: Hugh Dickins <hughd@google.com> Cc: Pavel Emelyanov <xemul@virtuozzo.com> Cc: Pravin Shedge <pravin.shedge4linux@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Since a2468cc9bfdf ("swap: choose swap device according to numa node"),
avail_lists field of swap_info_struct is changed to an array with
MAX_NUMNODES elements. This made swap_info_struct size increased to 40KiB
and needs an order-4 page to hold it.
This is not optimal in that:
1 Most systems have way less than MAX_NUMNODES(1024) nodes so it
is a waste of memory;
2 It could cause swapon failure if the swap device is swapped on
after system has been running for a while, due to no order-4
page is available as pointed out by Vasily Averin.
Solve the above two issues by using nr_node_ids(which is the actual
possible node number the running system has) for avail_lists instead of
MAX_NUMNODES.
nr_node_ids is unknown at compile time so can't be directly used when
declaring this array. What I did here is to declare avail_lists as zero
element array and allocate space for it when allocating space for
swap_info_struct. The reason why keep using array but not pointer is
plist_for_each_entry needs the field to be part of the struct, so pointer
will not work.
This patch is on top of Vasily Averin's fix commit. I think the use of
kvzalloc for swap_info_struct is still needed in case nr_node_ids is
really big on some systems.
Link: http://lkml.kernel.org/r/20181115083847.GA11129@intel.com Signed-off-by: Aaron Lu <aaron.lu@intel.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vasily Averin <vvs@virtuozzo.com> Cc: Huang Ying <ying.huang@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
write_cache_pages() is used in both background and integrity writeback
scenarios by various filesystems. Background writeback is mostly
concerned with cleaning a certain number of dirty pages based on various
mm heuristics. It may not write the full set of dirty pages or wait for
I/O to complete. Integrity writeback is responsible for persisting a set
of dirty pages before the writeback job completes. For example, an
fsync() call must perform integrity writeback to ensure data is on disk
before the call returns.
write_cache_pages() unconditionally breaks out of its processing loop in
the event of a ->writepage() error. This is fine for background
writeback, which had no strict requirements and will eventually come
around again. This can cause problems for integrity writeback on
filesystems that might need to clean up state associated with failed page
writeouts. For example, XFS performs internal delayed allocation
accounting before returning a ->writepage() error, where applicable. If
the current writeback happens to be associated with an unmount and
write_cache_pages() completes the writeback prematurely due to error, the
filesystem is unmounted in an inconsistent state if dirty+delalloc pages
still exist.
To handle this problem, update write_cache_pages() to always process the
full set of pages for integrity writeback regardless of ->writepage()
errors. Save the first encountered error and return it to the caller once
complete. This facilitates XFS (or any other fs that expects integrity
writeback to process the entire set of dirty pages) to clean up its
internal state completely in the event of persistent mapping errors.
Background writeback continues to exit on the first error encountered.
[akpm@linux-foundation.org: fix typo in comment] Link: http://lkml.kernel.org/r/20181116134304.32440-1-bfoster@redhat.com Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
mount.ocfs2 ignore the inconsistent error that journal is clean but
local alloc is unrecovered. After mount, local alloc not empty, then
reserver cluster didn't alloc a new local alloc window, reserveration
map is empty(ocfs2_reservation_map.m_bitmap_len = 0), that triggered the
following panic.
and was advised to fixed during mount. But this is a very unusual
inconsistent state, usually journal dirty flag should be cleared at the
last stage of umount until every other things go right. We may need do
further debug to check that. Any way to avoid possible futher
corruption, mount should be abort and fsck should be run.
iomap_is_partially_uptodate() is intended to check wither blocks within
the selected range of a not-uptodate page are uptodate; if the range we
care about is up to date, it's an optimization.
However, the iomap implementation continues to check all blocks up to
from+count, which is beyond the page, and can even be well beyond the
iop->uptodate bitmap.
I think the worst that will happen is that we may eventually find a zero
bit and return "not partially uptodate" when it would have otherwise
returned true, and skip the optimization. Still, it's clearly an invalid
memory access that must be fixed.
So: fix this by limiting the search to within the page as is done in the
non-iomap variant, block_is_partially_uptodate().
Zorro noticed thiswhen KASAN went off for 512 byte blocks on a 64k
page system:
BUG: KASAN: slab-out-of-bounds in iomap_is_partially_uptodate+0x1a0/0x1e0
Read of size 8 at addr ffff800120c3a318 by task fsstress/22337
Reported-by: Zorro Lang <zlang@redhat.com> Signed-off-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
This is because when populating Driver Map using firmware raid map, all
non-existing VDs set their ldTgtIdToLd to 0xff, so it can be skipped later.
From drivers/scsi/megaraid/megaraid_sas_base.c ,
memset(instance->ld_ids, 0xff, MEGASAS_MAX_LD_IDS);
From drivers/scsi/megaraid/megaraid_sas_fp.c ,
/* For non existing VDs, iterate to next VD*/
if (ld >= (MAX_LOGICAL_DRIVES_EXT - 1))
continue;
However, there are a few places that failed to skip those non-existing VDs
due to off-by-one errors. Then, those 0xff leaked into MR_LdRaidGet(0xff,
map) and triggered the out-of-bound accesses.
Fixes: 51087a8617fe ("megaraid_sas : Extended VD support") Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Sumit Saxena <sumit.saxena@broadcom.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
1. smartpqi is built as module, not built-in;
2. We have a disk connected to smartpqi card;
3. Both kexec's 1st and 2nd kernels use this disk as Rootfs' mount point.
Signed-off-by: Yanjiang Jin <yanjiang.jin@hxt-semitech.com> Acked-by: Don Brace <don.brace@microsemi.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
There was a race condition in SMP that an ath10k_peer was created but its
member sta was null. Following are procedures of ath10k_peer creation and
member sta access in peer statistics path.
Any access of peer->sta after peer was added to peer_map but before sta was
assigned could cause a null pointer issue. And because these two steps are
asynchronous, no proper lock can protect them. So both peer and sta need to
be checked before access.
Tested: QCA9984 with firmware ver 10.4-3.9.0.1-00005 Signed-off-by: Zhi Chen <zhichen@codeaurora.org> Signed-off-by: Kalle Valo <kvalo@codeaurora.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Problem:
The Linux kernel takes a logical volume offline after a LUN reset. This is
generally accompanied by this message in the dmesg output:
Device offlined - not ready after error recovery
Root Cause:
The root cause is a "quirk" in the timeout handling in the Linux SCSI
layer. The Linux kernel places a 30-second timeout on most media access
commands (reads and writes) that it send to device drivers. When a media
access command times out, the Linux kernel goes into error recovery mode
for the LUN that was the target of the command that timed out. Every
command that timed out is kept on a list inside of the Linux kernel to be
retried later. The kernel attempts to recover the command(s) that timed out
by issuing a LUN reset followed by a TEST UNIT READY. If the LUN reset and
TEST UNIT READY commands are successful, the kernel retries the command(s)
that timed out.
Each SCSI command issued by the kernel has a result field associated with
it. This field indicates the final result of the command (success or
error). When a command times out, the kernel places a value in this result
field indicating that the command timed out.
The "quirk" is that after the LUN reset and TEST UNIT READY commands are
completed, the kernel checks each command on the timed-out command list
before retrying it. If the result field is still "timed out", the kernel
treats that command as not having been successfully recovered for a
retry. If the number of commands that are in this state are greater than
two, the kernel takes the LUN offline.
Fix:
When our RAIDStack receives a LUN reset, it simply waits until all
outstanding commands complete. Generally, all of these outstanding commands
complete successfully. Therefore, the fix in the smartpqi driver is to
always set the command result field to indicate success when a request
completes successfully. This normally isn’t necessary because the result
field is always initialized to success when the command is submitted to the
driver. So when the command completes successfully, the result field is
left untouched. But in this case, the kernel changes the result field
behind the driver’s back and then expects the field to be changed by the
driver as the commands that timed-out complete.
Reviewed-by: Dave Carroll <david.carroll@microsemi.com> Reviewed-by: Scott Teel <scott.teel@microsemi.com> Signed-off-by: Kevin Barnett <kevin.barnett@microsemi.com> Signed-off-by: Don Brace <don.brace@microsemi.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
With commit 09c2f95ad404 ("scsi: mpt3sas: Swap I/O memory read value back
to cpu endianness"), 64bit writes in _base_writeq() were rewritten to use
__raw_writeq() instad of writeq().
This introduced a bug apparent on powerpc64 systems such as the Raptor
Talos II that causes the HBA to drop from the PCIe bus under heavy load and
being reinitialized after a couple of seconds.
It can easily be triggered on affacted systems by using something like
a couple of times. In my case I tested it on both a ZFS raidz2 and a btrfs
raid6 using LSI 9300-8i and 9400-8i controllers.
The fix consists in resembling the write ordering of writeq() by adding a
mandatory write memory barrier before device access and a compiler barrier
afterwards. The additional MMIO barrier is superfluous.
Signed-off-by: Stephan Günther <moepi@moepi.net> Reported-by: Matt Corallo <linux@bluematt.me> Acked-by: Sreekanth Reddy <Sreekanth.Reddy@broadcom.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
This commit moves the usdev_lock acquisition after the rtnl lock has been
released.
This is safe to do because usdev_lock is not protecting anything being
accessed in ib_get_eth_speed(). Hence, the correct order of holding locks
(rtnl -> usdev_lock) is not violated.
This is the much more correct fix for my earlier attempt at:
https://lkml.org/lkml/2018/12/10/118
Short recap:
- There's not actually a locking issue, it's just lockdep being a bit
too eager to complain about a possible deadlock.
- Contrary to what I claimed the real problem is recursion on
kn->count. Greg pointed me at sysfs_break_active_protection(), used
by the scsi subsystem to allow a sysfs file to unbind itself. That
would be a real deadlock, which isn't what's happening here. Also,
breaking the active protection means we'd need to manually handle
all the lifetime fun.
- With Rafael we discussed the task_work approach, which kinda works,
but has two downsides: It's a functional change for a lockdep
annotation issue, and it won't work for the bind file (which needs
to get the errno from the driver load function back to userspace).
- Greg also asked why this never showed up: To hit this you need to
unregister a 2nd driver from the unload code of your first driver. I
guess only gpus do that. The bug has always been there, but only
with a recent patch series did we add more locks so that lockdep
built a chain from unbinding the snd-hda driver to the
acpi_video_unregister call.
Looking around I've noticed that usb and i2c already handle similar
recursion problems, where a sysfs file can unbind the same type of
sysfs somewhere else in the hierarchy. Relevant commits are:
This commit fixes hard-coded model-id for an unit of Apogee Ensemble with
a correct value. This unit uses DM1500 ASIC produced ArchWave AG (formerly
known as BridgeCo AG).
I note that this model supports three modes in the number of data channels
in tx/rx streams; 8 ch pairs, 10 ch pairs, 18 ch pairs. The mode is
switched by Vendor-dependent AV/C command, like:
When switching between different mode, the unit disappears from IEEE 1394
bus, then appears on the bus with different combination of stream formats.
In a mode of 18 ch pairs, available sampling rate is up to 96.0 kHz, else
up to 192.0 kHz.
Reference to a device in device-mapper table contains offset in sectors.
If the sector_t is 32bit integer (CONFIG_LBDAF is not set), then
several device-mapper targets can overflow this offset and validity
check is then performed on a wrong offset and a wrong table is activated.
See for example (on 32bit without CONFIG_LBDAF) this overflow:
# dmsetup create test --table "0 2048 linear /dev/sdg 4294967297"
# dmsetup table test
0 2048 linear 8:96 1
This patch adds explicit check for overflow if the offset is sector_t type.
Signed-off-by: Milan Broz <gmazyland@gmail.com> Reviewed-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Mike Snitzer <snitzer@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The function of_find_node_by_path() acquires a reference to the node
returned by it and that reference needs to be dropped by its caller.
integrator_ap_timer_init_of() doesn't do that. The pri_node and the
sec_node are used as an identifier to compare against the current
node, so we can directly drop the refcount after getting the node from
the path as it is not used as pointer.
By dropping the refcount right after getting it, a single variable is
needed instead of two.
Fix this by use a single variable and drop the refcount right after
of_find_node_by_path().
Signed-off-by: Yangtao Li <tiny.windzz@gmail.com> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 1fa5efe3622db58cb8c7b9a50665e9eb9a6c7e97 (ext4: Use generic helpers for quotaon
and quotaoff) made possible to call quotactl(Q_XQUOTAON/OFF) on ext4 filesystems
with sysfile quota support. This leads to calling dquot_enable/disable without s_umount
held in excl. mode, because quotactl_cmd_onoff checks only for Q_QUOTAON/OFF.
The following WARN_ON_ONCE triggers (in this case for dquot_enable, ext4, latest Linus' tree):
[ 117.807056] EXT4-fs (dm-0): mounted filesystem with ordered data mode. Opts: quota,prjquota
and then the syscall proceeds without s_umount locking.
This patch locks the superblock ->s_umount sem. in exclusive mode for all Q_XQUOTAON/OFF
quotactls too in addition to Q_QUOTAON/OFF.
AFAICT, other than ext4, only xfs and ocfs2 are affected by this change.
The VFS will now call in xfs_quota_* functions with s_umount held, which wasn't the case
before. This looks good to me but I can not say for sure. Ext4 and ocfs2 where already
beeing called with s_umount exclusive via quota_quotaon/off which is basically the same.
Signed-off-by: Javier Barrio <javier.barrio.mart@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Sasha Levin <sashal@kernel.org>
There are systems such as the Android NDK API level 24 has the
open_memstream() function but doesn't provide a prototype, adding noise
to the build:
builtin-timechart.c: In function 'cat_backtrace':
builtin-timechart.c:486:2: warning: implicit declaration of function 'open_memstream' [-Wimplicit-function-declaration]
FILE *f = open_memstream(&p, &p_len);
^
builtin-timechart.c:486:2: warning: nested extern declaration of 'open_memstream' [-Wnested-externs]
builtin-timechart.c:486:12: warning: initialization makes pointer from integer without a cast
FILE *f = open_memstream(&p, &p_len);
^
Define a LACKS_OPEN_MEMSTREAM_PROTOTYPE define so that code needing that
can get a prototype.
Checked in the bionic git repo to be available since level 23:
There are systems such as the Android NDK API level 24 has the
sigqueue() function but doesn't provide a prototype, adding noise to the
build:
util/evlist.c: In function 'perf_evlist__prepare_workload':
util/evlist.c:1494:4: warning: implicit declaration of function 'sigqueue' [-Wimplicit-function-declaration]
if (sigqueue(getppid(), SIGUSR1, val))
^
util/evlist.c:1494:4: warning: nested extern declaration of 'sigqueue' [-Wnested-externs]
Define a LACKS_SIGQUEUE_PROTOTYPE define so that code needing that can
get a prototype.
Checked in the bionic git repo to be available since level 23:
The structure cs_etm_queue uses 'prev_packet' to point to previous
packet, this can be used to combine with new coming packet to generate
samples.
In function cs_etm__flush() it swaps packets only when the flag
'etm->synth_opts.last_branch' is true, this means that it will not swap
packets if without option '--itrace=il' to generate last branch entries;
thus for this case the 'prev_packet' doesn't point to the correct
previous packet and the stale packet still will be used to generate
sequential sample. Thus if dump trace with 'perf script' command we can
see the incorrect flow with the stale packet's address info.
This patch corrects packets swapping in cs_etm__flush(); except using
the flag 'etm->synth_opts.last_branch' it also checks the another flag
'etm->sample_branches', if any flag is true then it swaps packets so can
save correct content to 'prev_packet'. Finally this can fix the wrong
program flow dumping issue.
The patch has a minor refactoring to use 'etm->synth_opts.last_branch'
instead of 'etmq->etm->synth_opts.last_branch' for condition checking,
this is consistent with that is done in cs_etm__sample().
Signed-off-by: Leo Yan <leo.yan@linaro.org> Reviewed-by: Mathieu Poirier <mathieu.poirier@linaro.org> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Leach <mike.leach@linaro.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Robert Walker <robert.walker@arm.com> Cc: coresight@lists.linaro.org Cc: linux-arm-kernel@lists.infradead.org Link: http://lkml.kernel.org/r/1544513908-16805-2-git-send-email-leo.yan@linaro.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
kcopyd has no upper limit to the number of jobs one can allocate and
issue. Under certain workloads this can lead to excessive memory usage
and workqueue stalls. For example, when creating multiple dm-snapshot
targets with a 4K chunk size and then writing to the origin through the
page cache. Syncing the page cache causes a large number of BIOs to be
issued to the dm-snapshot origin target, which itself issues an even
larger (because of the BIO splitting taking place) number of kcopyd
jobs.
Running the following test, from the device mapper test suite [1],
dmtest run --suite snapshot -n many_snapshots_of_same_volume_N
, with 8 active snapshots, results in the kcopyd job slab cache growing
to 10G. Depending on the available system RAM this can lead to the OOM
killer killing user processes:
Moreover, issuing a large number of kcopyd jobs results in kcopyd
hogging the CPU, while processing them. As a result, processing of work
items, queued for execution on the same CPU as the currently running
kcopyd thread, is stalled for long periods of time, hurting performance.
Running the aforementioned test we get, in dmesg, messages like the
following:
The root cause for these issues is the way dm-snapshot uses kcopyd. In
particular, the lack of an explicit or implicit limit to the maximum
number of in-flight COW jobs. The merging path is not affected because
it implicitly limits the in-flight kcopyd jobs to one.
Fix these issues by using a semaphore to limit the maximum number of
in-flight kcopyd jobs. We grab the semaphore before allocating a new
kcopyd job in start_copy() and start_full_bio() and release it after the
job finishes in copy_callback().
The initial semaphore value is configurable through a module parameter,
to allow fine tuning the maximum number of in-flight COW jobs. Setting
this parameter to zero initializes the semaphore to INT_MAX.
A default value of 2048 maximum in-flight kcopyd jobs was chosen. This
value was decided experimentally as a trade-off between memory
consumption, stalling the kernel's workqueues and maintaining a high
enough throughput.
Re-running the aforementioned test:
* Workqueue stalls are eliminated
* kcopyd's job slab cache uses a maximum of 130MB
* The time taken by the test to write to the snapshot-origin target is
reduced from 05m20.48s to 03m26.38s
At some point we decided not to directly include kernel sources files
when building tools/perf/, but when tools/lib/subcmd/ was forked from
tools/perf it somehow ended up adding it via these two lines in its
Makefile:
This fixes the build when building from the sources in environments such
as the Android NDK crossbuilding from a fedora:26 system:
subcmd-util.h:11:15: error: expected ',' or ';' before 'void'
static inline void report(const char *prefix, const char *err, va_list params)
^
In file included from /git/perf/include/uapi/linux/stddef.h:2:0,
from /git/perf/include/uapi/linux/posix_types.h:5,
from /opt/android-ndk-r12b/platforms/android-24/arch-arm/usr/include/sys/types.h:36,
from /opt/android-ndk-r12b/platforms/android-24/arch-arm/usr/include/unistd.h:33,
from run-command.c:2:
subcmd-util.h:18:17: error: '__no_instrument_function__' attribute applies only to functions
The /opt/android-ndk-r12b/platforms/android-24/arch-arm/usr/include/sys/types.h
file that includes linux/posix_types.h ends up getting the one in the kernel
sources causing the breakage. Fix it.
Test built tools/objtool/ too.
Reported-by: Jiri Olsa <jolsa@kernel.org> Tested-by: Jiri Olsa <jolsa@kernel.org> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Josh Poimboeuf <jpoimboe@redhat.com> Cc: Namhyung Kim <namhyung@kernel.org> Fixes: 4b6ab94eabe4 ("perf subcmd: Create subcmd library") Link: https://lkml.kernel.org/n/tip-5lhaoecrj12t0bqwvpiu14sm@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Some 'perf stat' options do not make sense to be negated (event,
cgroup), some do not have negated path implemented (metrics). Due to
that, it is better to disable the "no-" prefix for them, since
otherwise, the later opt-parsing segfaults.
Before:
$ perf stat --no-metrics -- ls
Segmentation fault (core dumped)
After:
$ perf stat --no-metrics -- ls
Error: option `no-metrics' isn't available
Usage: perf stat [<options>] [<command>]
Signed-off-by: Michael Petlan <mpetlan@redhat.com> Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
LPU-Reference: 1485912065.62416880.1544457604340.JavaMail.zimbra@redhat.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When using kcopyd to run callbacks through dm_kcopyd_do_callback() or
submitting copy jobs with a source size of 0, the jobs are pushed
directly to the complete_jobs list, which could be under processing by
the kcopyd thread. As a result, the kcopyd thread can continue running
completed jobs indefinitely, without releasing the CPU, as long as
someone keeps submitting new completed jobs through the aforementioned
paths. Processing of work items, queued for execution on the same CPU as
the currently running kcopyd thread, is thus stalled for excessive
amounts of time, hurting performance.
Running the following test, from the device mapper test suite [1],
dmtest run --suite snapshot -n parallel_io_to_many_snaps_N
, with 8 active snapshots, we get, in dmesg, messages like the
following:
Fix this by splitting the complete_jobs list into two parts: A user
facing part, named callback_jobs, and one used internally by kcopyd,
retaining the name complete_jobs. dm_kcopyd_do_callback() and
dispatch_job() now push their jobs to the callback_jobs list, which is
spliced to the complete_jobs list once, every time the kcopyd thread
wakes up. This prevents kcopyd from hogging the CPU indefinitely and
causing workqueue stalls.
Re-running the aforementioned test:
* Workqueue stalls are eliminated
* The maximum writing time among all targets is reduced from 09m37.10s
to 06m04.85s and the total run time of the test is reduced from
10m43.591s to 7m19.199s
The iv_offset in the mapping table of crypt target is a 64bit number when
IV algorithm is plain64, plain64be, essiv or benbi. It will be assigned to
iv_offset of struct crypt_config, cc_sector of struct convert_context and
iv_sector of struct dm_crypt_request. These structures members are defined
as a sector_t. But sector_t is 32bit when CONFIG_LBDAF is not set in 32bit
kernel. In this situation sector_t is not big enough to store the 64bit
iv_offset.
Here is a reproducer.
Prepare test image and device (loop is automatically allocated by cryptsetup):
# dd if=/dev/zero of=tst.img bs=1M count=1
# echo "tst"|cryptsetup open --type plain -c aes-xts-plain64 \
--skip 500000000000000000 tst.img test
On 32bit system (use IV offset value that overflows to 64bit; CONFIG_LBDAF if off)
and device checksum is wrong:
Signed-off-by: AliOS system security <alios_sys_security@linux.alibaba.com> Tested-and-Reviewed-by: Milan Broz <gmazyland@gmail.com> Signed-off-by: Mike Snitzer <snitzer@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
nr_cpu_ids can be limited on the command line via nr_cpus=. This can break the
logical package management because it results in a smaller number of packages
while in kdump kernel.
Check below case:
There is a two sockets system, each socket has 8 cores, which has 16 logical
cpus while HT was turn on.
While starting the kdump kernel with command line option nr_cpus=16 panic
was triggered on one of the cpus 24-31 eg. 26, then online cpu will be
1-15, 26(cpu 0 was disabled in kdump), ncpus will be 16 and
__max_logical_packages will be 1, but actually two packages were booted on.
This issue can reproduced by set kdump option nr_cpus=<real physical core
numbers>, and then trigger panic on last socket's thread, for example:
taskset -c 26 echo c > /proc/sysrq-trigger
Use total_cpus which will not be limited by nr_cpus command line to calculate
the value of __max_logical_packages.
When network namespace is destroyed, cleanup_net() is called.
cleanup_net() holds pernet_ops_rwsem then calls each ->exit callback.
So that clusterip_tg_destroy() is called by cleanup_net().
And clusterip_tg_destroy() calls unregister_netdevice_notifier().
But both cleanup_net() and clusterip_tg_destroy() hold same
lock(pernet_ops_rwsem). hence deadlock occurrs.
After this patch, only 1 notifier is registered when module is inserted.
And all of configs are added to per-net list.
test commands:
%ip netns add vm1
%ip netns exec vm1 bash
%ip link set lo up
%iptables -A INPUT -p tcp -i lo -d 192.168.0.5 --dport 80 \
-j CLUSTERIP --new --hashmode sourceip \
--clustermac 01:00:5e:00:00:20 --total-nodes 2 --local-node 1
%exit
%ip netns del vm1
When network namespace is destroyed, both clusterip_tg_destroy() and
clusterip_net_exit() are called. and clusterip_net_exit() is called
before clusterip_tg_destroy().
Hence cleanup check code in clusterip_net_exit() doesn't make sense.
test commands:
%ip netns add vm1
%ip netns exec vm1 bash
%ip link set lo up
%iptables -A INPUT -p tcp -i lo -d 192.168.0.5 --dport 80 \
-j CLUSTERIP --new --hashmode sourceip \
--clustermac 01:00:5e:00:00:20 --total-nodes 2 --local-node 1
%exit
%ip netns del vm1
If same destination IP address config is already existing, that config is
just used. MAC address also should be same.
However, there is no MAC address checking routine.
So that MAC address checking routine is added.
test commands:
%iptables -A INPUT -p tcp -i lo -d 192.168.0.5 --dport 80 \
-j CLUSTERIP --new --hashmode sourceip \
--clustermac 01:00:5e:00:00:20 --total-nodes 2 --local-node 1
%iptables -A INPUT -p tcp -i lo -d 192.168.0.5 --dport 80 \
-j CLUSTERIP --new --hashmode sourceip \
--clustermac 01:00:5e:00:00:21 --total-nodes 2 --local-node 1
The strncpy() function may leave the destination string buffer
unterminated, better use strlcpy() that we have a __weak fallback
implementation for systems without it.
This fixes this warning on an Alpine Linux Edge system with gcc 8.2:
util/parse-events.c: In function 'print_symbol_events':
util/parse-events.c:2465:4: error: 'strncpy' specified bound 100 equals destination size [-Werror=stringop-truncation]
strncpy(name, syms->symbol, MAX_NAME_LEN);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In function 'print_symbol_events.constprop',
inlined from 'print_events' at util/parse-events.c:2508:2:
util/parse-events.c:2465:4: error: 'strncpy' specified bound 100 equals destination size [-Werror=stringop-truncation]
strncpy(name, syms->symbol, MAX_NAME_LEN);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In function 'print_symbol_events.constprop',
inlined from 'print_events' at util/parse-events.c:2511:2:
util/parse-events.c:2465:4: error: 'strncpy' specified bound 100 equals destination size [-Werror=stringop-truncation]
strncpy(name, syms->symbol, MAX_NAME_LEN);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Fixes: 947b4ad1d198 ("perf list: Fix max event string size") Link: https://lkml.kernel.org/n/tip-b663e33bm6x8hrkie4uxh7u2@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The strncpy() function may leave the destination string buffer
unterminated, better use strlcpy() that we have a __weak fallback
implementation for systems without it.
In this specific case this would only happen if fgets() was buggy, as
its man page states that it should read one less byte than the size of
the destination buffer, so that it can put the nul byte at the end of
it, so it would never copy 255 non-nul chars, as fgets reads into the
orig buffer at most 254 non-nul chars and terminates it. But lets just
switch to strlcpy to keep the original intent and silence the gcc 8.2
warning.
This fixes this warning on an Alpine Linux Edge system with gcc 8.2:
In function 'cpu_model',
inlined from 'svg_cpu_box' at util/svghelper.c:378:2:
util/svghelper.c:337:5: error: 'strncpy' output may be truncated copying 255 bytes from a string of length 255 [-Werror=stringop-truncation]
strncpy(cpu_m, &buf[13], 255);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Arjan van de Ven <arjan@linux.intel.com> Fixes: f48d55ce7871 ("perf: Add a SVG helper library file") Link: https://lkml.kernel.org/n/tip-xzkoo0gyr56gej39ltivuh9g@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The breakpoint tests on the ARM 32-bit kernel are broken in several
ways.
The breakpoint length requested does not necessarily match whether the
function address has the Thumb bit (bit 0) set or not, and this does
matter to the ARM kernel hw_breakpoint infrastructure. See [1] for
background.
[1]: https://lkml.org/lkml/2018/11/15/205
As Will indicated, the overflow handling would require single-stepping
which is not supported at the moment. Just disable those tests for the
ARM 32-bit platforms and update the comment above to explain these
limitations.
Co-developed-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Will Deacon <will.deacon@arm.com> Acked-by: Jiri Olsa <jolsa@redhat.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20181203191138.2419-1-f.fainelli@gmail.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Users should never use 'pt=0', but if they do it may give a meaningless
error:
$ perf record -e intel_pt/pt=0/u uname
Error:
The sys_perf_event_open() syscall returned with 22 (Invalid argument) for
event (intel_pt/pt=0/u).
Fix that by forcing 'pt=1'.
Committer testing:
# perf record -e intel_pt/pt=0/u uname
Error:
The sys_perf_event_open() syscall returned with 22 (Invalid argument) for event (intel_pt/pt=0/u).
/bin/dmesg | grep -i perf may provide additional information.
# perf record -e intel_pt/pt=0/u uname
pt=0 doesn't make sense, forcing pt=1
Linux
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.020 MB perf.data ]
#
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Jiri Olsa <jolsa@redhat.com> Link: http://lkml.kernel.org/r/b7c5b4e5-9497-10e5-fd43-5f3e4a0fe51d@intel.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
So there are two dependency chains:
uart_port->lock -> db->lock
And
db->lock -> uart_port->lock
This particular circular locking dependency can be addressed in several
ways:
a) One way would be to move debug_print_object() out of db->lock scope
and, thus, break the db->lock -> uart_port->lock chain.
b) Another one would be to free() transmit buffer page out of db->lock
in UART code; which is what this patch does.
It makes sense to apply a) and b) independently: there are too many things
going on behind free(), none of which depend on uart_port->lock.
The patch fixes transmit buffer page free() in uart_shutdown() and,
additionally, in uart_port_startup() (as was suggested by Dmitry Safonov).
[1] https://lore.kernel.org/lkml/20181211091154.GL23332@shao2-debian/T/#u Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Reviewed-by: Petr Mladek <pmladek@suse.com> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Jiri Slaby <jslaby@suse.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Waiman Long <longman@redhat.com> Cc: Dmitry Safonov <dima@arista.com> Cc: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In the error handling block, err holds the return value of either
btrfs_del_root_ref() or btrfs_del_inode_ref() but it hasn't been checked
since it's introduction with commit fe66a05a0679 (Btrfs: improve error
handling for btrfs_insert_dir_item callers) in 2012.
If the error handling in the error handling fails, there's not much left
to do and the abort either happened earlier in the callees or is
necessary here.
So if one of btrfs_del_root_ref() or btrfs_del_inode_ref() failed, abort
the transaction, but still return the original code of the failure
stored in 'ret' as this will be reported to the user.
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The device replace cancel thread can race with the replace start thread
and if fs_info::scrubs_running is not yet set, btrfs_scrub_cancel() will
fail to stop the scrub thread.
The scrub thread continues with the scrub for replace which then will
try to write to the target device and which is already freed by the
cancel thread.
Fix this by letting the btrfs_dev_replace_finishing() to do the job of
cleaning after the cancel, including freeing of the target device.
btrfs_dev_replace_finishing() is called when btrfs_scub_dev() returns
along with the scrub return status.
Signed-off-by: Anand Jain <anand.jain@oracle.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 92e222df7b "btrfs: alloc_chunk: fix DUP stripe size handling"
fixed calculating the stripe_size for a new DUP chunk.
However, the same calculation reappears a bit later, and that one was
not changed yet. The resulting bug that is exposed is that the newly
allocated device extents ('stripes') can have a few MiB overlap with the
next thing stored after them, which is another device extent or the end
of the disk.
The scenario in which this can happen is:
* The block device for the filesystem is less than 10GiB in size.
* The amount of contiguous free unallocated disk space chosen to use for
chunk allocation is 20% of the total device size, or a few MiB more or
less.
An example:
- The filesystem device is 7880MiB (max_chunk_size gets set to 788MiB)
- There's 1578MiB unallocated raw disk space left in one contiguous
piece.
In this case stripe_size is first calculated as 789MiB, (half of
1578MiB).
Since 789MiB (stripe_size * data_stripes) > 788MiB (max_chunk_size), we
enter the if block. Now stripe_size value is immediately overwritten
while calculating an adjusted value based on max_chunk_size, which ends
up as 788MiB.
Next, the value is rounded up to a 16MiB boundary, 800MiB, which is
actually more than the value we had before. However, the last comparison
fails to detect this, because it's comparing the value with the total
amount of free space, which is about twice the size of stripe_size.
In the example above, this means that the resulting raw disk space being
allocated is 1600MiB, while only a gap of 1578MiB has been found. The
second device extent object for this DUP chunk will overlap for 22MiB
with whatever comes next.
The underlying problem here is that the stripe_size is reused all the
time for different things. So, when entering the code in the if block,
stripe_size is immediately overwritten with something else. If later we
decide we want to have the previous value back, then the logic to
compute it was copy pasted in again.
With this change, the value in stripe_size is not unnecessarily
destroyed, so the duplicated calculation is not needed any more.
Signed-off-by: Hans van Kranenburg <hans.van.kranenburg@mendix.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Enhance btrfs_verify_dev_extents() to remember previous checked dev
extents, so it can verify no dev extents can overlap.
Analysis from Hans:
"Imagine allocating a DATA|DUP chunk.
In the chunk allocator, we first set...
max_stripe_size = SZ_1G;
max_chunk_size = BTRFS_MAX_DATA_CHUNK_SIZE
... which is 10GiB.
Then...
/* we don't want a chunk larger than 10% of writeable space */
max_chunk_size = min(div_factor(fs_devices->total_rw_bytes, 1),
max_chunk_size);
Imagine we only have one 7880MiB block device in this filesystem. Now
max_chunk_size is down to 788MiB.
The next step in the code is to search for max_stripe_size * dev_stripes
amount of free space on the device, which is in our example 1GiB * 2 =
2GiB. Imagine the device has exactly 1578MiB free in one contiguous
piece. This amount of bytes will be put in devices_info[ndevs - 1].max_avail
Next we recalculate the stripe_size (which is actually the device extent
length), based on the actual maximum amount of available raw disk space:
stripe_size = div_u64(devices_info[ndevs - 1].max_avail, dev_stripes);
stripe_size is now 789MiB
Next we do...
data_stripes = num_stripes / ncopies
...where data_stripes ends up as 1, because num_stripes is 2 (the amount
of device extents we're going to have), and DUP has ncopies 2.
Next there's a check...
if (stripe_size * data_stripes > max_chunk_size)
...which matches because 789MiB * 1 > 788MiB.
We go into the if code, and next is...
stripe_size = div_u64(max_chunk_size, data_stripes);
...which resets stripe_size to max_chunk_size: 788MiB
Next is a fun one...
/* bump the answer up to a 16MB boundary */
stripe_size = round_up(stripe_size, SZ_16M);
...which changes stripe_size from 788MiB to 800MiB.
We're not done changing stripe_size yet...
/* But don't go higher than the limits we found while searching
* for free extents
*/
stripe_size = min(devices_info[ndevs - 1].max_avail,
stripe_size);
This is bad. max_avail is twice the stripe_size (we need to fit 2 device
extents on the same device for DUP).
The result here is that 800MiB < 1578MiB, so it's unchanged. However,
the resulting DUP chunk will need 1600MiB disk space, which isn't there,
and the second dev_extent might extend into the next thing (next
dev_extent? end of device?) for 22MiB.
The last shown line of code relies on a situation where there's twice
the value of stripe_size present as value for the variable stripe_size
when it's DUP. This was actually the case before commit 92e222df7b
"btrfs: alloc_chunk: fix DUP stripe size handling", from which I quote:
"[...] in the meantime there's a check to see if the stripe_size does
not exceed max_chunk_size. Since during this check stripe_size is twice
the amount as intended, the check will reduce the stripe_size to
max_chunk_size if the actual correct to be used stripe_size is more than
half the amount of max_chunk_size."
In the previous version of the code, the 16MiB alignment (why is this
done, by the way?) would result in a 50% chance that it would actually
do an 8MiB alignment for the individual dev_extents, since it was
operating on double the size. Does this matter?
Does it matter that stripe_size can be set to anything which is not
16MiB aligned because of the amount of remaining available disk space
which is just taken?
What is the main purpose of this round_up?
The most straightforward thing to do seems something like...
stripe_size = min(
div_u64(devices_info[ndevs - 1].max_avail, dev_stripes),
stripe_size
)
..just putting half of the max_avail into stripe_size."
On our AT91SAM9260 board we use the same sdio bus for wifi and for the
sd card slot. This caused the atmel-mci to give the following splat on
the serial console:
------------[ cut here ]------------
WARNING: CPU: 0 PID: 538 at drivers/mmc/host/atmel-mci.c:859 atmci_send_command+0x24/0x44
Modules linked in:
CPU: 0 PID: 538 Comm: mmcqd/0 Not tainted 4.14.76 #14
Hardware name: Atmel AT91SAM9
[<c000fccc>] (unwind_backtrace) from [<c000d3dc>] (show_stack+0x10/0x14)
[<c000d3dc>] (show_stack) from [<c0017644>] (__warn+0xd8/0xf4)
[<c0017644>] (__warn) from [<c0017704>] (warn_slowpath_null+0x1c/0x24)
[<c0017704>] (warn_slowpath_null) from [<c033bb9c>] (atmci_send_command+0x24/0x44)
[<c033bb9c>] (atmci_send_command) from [<c033e984>] (atmci_start_request+0x1f4/0x2dc)
[<c033e984>] (atmci_start_request) from [<c033f3b4>] (atmci_request+0xf0/0x164)
[<c033f3b4>] (atmci_request) from [<c0327108>] (mmc_start_request+0x280/0x2d0)
[<c0327108>] (mmc_start_request) from [<c032800c>] (mmc_start_areq+0x230/0x330)
[<c032800c>] (mmc_start_areq) from [<c03366f8>] (mmc_blk_issue_rw_rq+0xc4/0x310)
[<c03366f8>] (mmc_blk_issue_rw_rq) from [<c03372c4>] (mmc_blk_issue_rq+0x118/0x5ac)
[<c03372c4>] (mmc_blk_issue_rq) from [<c033781c>] (mmc_queue_thread+0xc4/0x118)
[<c033781c>] (mmc_queue_thread) from [<c002daf8>] (kthread+0x100/0x118)
[<c002daf8>] (kthread) from [<c000a580>] (ret_from_fork+0x14/0x34)
---[ end trace 594371ddfa284bd6 ]---
This is:
WARN_ON(host->cmd);
This was fixed on our board by letting atmci_request_end determine what
state we are in. Instead of unconditionally setting it to STATE_IDLE on
STATE_END_REQUEST.
An unterminated string literal followed by new line is passed to the
parser (with "multi-line strings not supported" warning shown), then
handled properly there.
On the other hand, an unterminated string literal at end of file is
never passed to the parser, then results in memory leak.
Currently, warn_ignore_character() displays invalid file name and
line number.
The lexer should use current_file->name and yylineno, while the parser
should use zconf_curname() and zconf_lineno().
This difference comes from that the lexer is always going ahead
of the parser. The parser needs to look ahead one token to make a
shift/reduce decision, so the lexer is requested to scan more text
from the input file.
This commit fixes the warning message from warn_ignored_character().
[Test Code]
----(Kconfig begin)----
/
-----(Kconfig end)-----
[Output]
Before the fix:
<none>:0:warning: ignoring unsupported character '/'
After the fix:
Kconfig:1:warning: ignoring unsupported character '/'
Currently, the destination register is marked as unknown for 32-bit
sub-register move (BPF_MOV | BPF_ALU) whenever the source register type is
SCALAR_VALUE.
This is too conservative that some valid cases will be rejected.
Especially, this may turn a constant scalar value into unknown value that
could break some assumptions of verifier.
For example, test_l4lb_noinline.c has the following C code:
struct real_definition *dst
1: if (!get_packet_dst(&dst, &pckt, vip_info, is_ipv6))
2: return TC_ACT_SHOT;
3:
4: if (dst->flags & F_IPV6) {
get_packet_dst is responsible for initializing "dst" into valid pointer and
return true (1), otherwise return false (0). The compiled instruction
sequence using alu32 will be:
insn 413, a BPF_MOV | BPF_ALU, however will turn r0 into unknown value even
r7 contains SCALAR_VALUE 1.
This causes trouble when verifier is walking the code path that hasn't
initialized "dst" inside get_packet_dst, for which case 0 is returned and
we would then expect verifier concluding line 1 in the above C code pass
the "if" check, therefore would skip fall through path starting at line 4.
Now, because r0 returned from callee has became unknown value, so verifier
won't skip analyzing path starting at line 4 and "dst->flags" requires
dereferencing the pointer "dst" which actually hasn't be initialized for
this path.
This patch relaxed the code marking sub-register move destination. For a
SCALAR_VALUE, it is safe to just copy the value from source then truncate
it into 32-bit.
A unit test also included to demonstrate this issue. This test will fail
before this patch.
This relaxation could let verifier skipping more paths for conditional
comparison against immediate. It also let verifier recording a more
accurate/strict value for one register at one state, if this state end up
with going through exit without rejection and it is used for state
comparison later, then it is possible an inaccurate/permissive value is
better. So the real impact on verifier processed insn number is complex.
But in all, without this fix, valid program could be rejected.
>From real benchmarking on kernel selftests and Cilium bpf tests, there is
no impact on processed instruction number when tests ares compiled with
default compilation options. There is slightly improvements when they are
compiled with -mattr=+alu32 after this patch.
Also, test_xdp_noinline/-mattr=+alu32 now passed verification. It is
rejected before this fix.
NOTE:
- bpf_lxc.o and bpf_netdev.o compiled by -mattr=+alu32 are rejected by
verifier due to another issue inside verifier on supporting alu32
binary.
- Each cilium bpf program could generate several processed insn number,
above number is sum of them.
v1->v2:
- Restrict the change on SCALAR_VALUE.
- Update benchmark numbers on Cilium bpf tests.
The dcache_by_line_op macro suffers from a couple of small problems:
First, the GAS directives that are currently being used rely on
assembler behavior that is not documented, and probably not guaranteed
to produce the correct behavior going forward. As a result, we end up
with some undefined symbols in cache.o:
$ nm arch/arm64/mm/cache.o
...
U civac
...
U cvac
U cvap
U cvau
This is due to the fact that the comparisons used to select the
operation type in the dcache_by_line_op macro are comparing symbols
not strings, and even though it seems that GAS is doing the right
thing here (undefined symbols by the same name are equal to each
other), it seems unwise to rely on this.
Second, when patching in a DC CVAP instruction on CPUs that support it,
the fallback path consists of a DC CVAU instruction which may be
affected by CPU errata that require ARM64_WORKAROUND_CLEAN_CACHE.
Solve these issues by unrolling the various maintenance routines and
using the conditional directives that are documented as operating on
strings. To avoid the complexity of nested alternatives, we move the
DC CVAP patching to __clean_dcache_area_pop, falling back to a branch
to __clean_dcache_area_poc if DCPOP is not supported by the CPU.
Reported-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Suggested-by: Robin Murphy <robin.murphy@arm.com> Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The exclusive gates may be set up in the wrong way by software running
before the clock driver comes up. In that case the exclusive setup is
locked in its initial state, as the complementary function can't be
activated without disabling the initial setup first.
To avoid this lock situation, reset the exclusive gates to the off
state and allow the kernel to provide the proper setup.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de> Reviewed-by: Dong Aisheng <Aisheng.dong@nxp.com> Signed-off-by: Stephen Boyd <sboyd@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
If the kernel is configured with KASAN_EXTRA, the stack size is
increased significantly due to setting the GCC -fstack-reuse option to
"none" [1]. As a result, it can trigger a stack overrun quite often with
32k stack size compiled using GCC 8. For example, this reproducer
There are just too many functions that could have a large stack with
KASAN_EXTRA due to large local variables that have been called over and
over again without being able to reuse the stacks. Some noticiable ones
are,
There are other 49 functions over 2k in size while compiling kernel with
"-Wframe-larger-than=" on this machine. Hence, it is too much work to
change Makefiles for each object to compile without
-fsanitize-address-use-after-scope individually.
A session must only be released after all code that accesses the session
structure has finished. Make sure that this is the case by introducing a
new command counter per session that is only decremented after the
.release_cmd() callback has finished. This patch fixes the following crash:
Cc: Nicholas Bellinger <nab@linux-iscsi.org> Cc: Mike Christie <mchristi@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: David Disseldorp <ddiss@suse.de> Cc: Hannes Reinecke <hare@suse.de> Signed-off-by: Bart Van Assche <bvanassche@acm.org> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
4.3.1 ASCII data field requirements
ASCII data fields shall contain only ASCII printable characters (i.e.,
code values 20h to 7Eh) and may be terminated with one or more ASCII null
(00h) characters. ASCII data fields described as being left-aligned
shall have any unused bytes at the end of the field (i.e., highest
offset) and the unused bytes shall be filled with ASCII space characters
(20h).
LIO currently space-pads the T10 VENDOR IDENTIFICATION and PRODUCT
IDENTIFICATION fields in the standard INQUIRY data. However, the PRODUCT
REVISION LEVEL field in the standard INQUIRY data as well as the T10 VENDOR
IDENTIFICATION field in the INQUIRY Device Identification VPD Page are
zero-terminated/zero-padded.
Fix this inconsistency by using space-padding for all of the above fields.
Signed-off-by: David Disseldorp <ddiss@suse.de> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Bryant G. Ly <bly@catalogicsoftware.com> Reviewed-by: Lee Duncan <lduncan@suse.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
after set SO_DONTROUTE to 1, the IP layer should not route packets if
the dest IP address is not in link scope. But if the socket has cached
the dst_entry, such packets would be routed until the sk_dst_cache
expires. So we should clean the sk_dst_cache when a user set
SO_DONTROUTE option. Below are server/client python scripts which
could reprodue this issue:
server side code:
==========================================================================
import socket
import struct
import time
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.bind(('0.0.0.0', 9000))
s.listen(1)
sock, addr = s.accept()
sock.setsockopt(socket.SOL_SOCKET, socket.SO_DONTROUTE, struct.pack('i', 1))
while True:
sock.send(b'foo')
time.sleep(1)
==========================================================================
client side code:
==========================================================================
import socket
import time
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.connect(('server_address', 9000))
while True:
data = s.recv(1024)
print(data)
==========================================================================
Signed-off-by: yupeng <yupeng0921@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
>From above warning one would deduce that the sg segment will overflow
the device's capacity. In reality, the hardware can accommodate larger
sg segments.
So, initialize the max segment size properly to weed out this warning.
Based on a similar patch sent by Sean Paul for mdss:
https://patchwork.kernel.org/patch/10671457/
AMD platform device acp_audio_dma can only be created by parent PCI
device driver (drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c). Pass struct
device of the parent to snd_pcm_lib_preallocate_pages() so
dma_alloc_coherent() can use correct dma_ops. Otherwise, it will
use default dma_ops which is nommu_dma_ops on x86_64 even when
IOMMU is enabled and set to non passthrough mode.
Though platform device inherits some dma related fields during its
creation in mfd_add_device(), we can't simply pass its struct device
to snd_pcm_lib_preallocate_pages() because dma_ops is not among the
inherited fields. Even it were, drivers/iommu/amd_iommu.c would
ignore it because get_device_id() doesn't handle platform device.
This change shouldn't give us any trouble even struct device of the
parent becomes null or represents some non PCI device in the future,
because get_dma_ops() correctly handles null struct device or uses
the default dma_ops if struct device doesn't have it set.
Signed-off-by: Yu Zhao <yuzhao@google.com> Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
drivers/media/firewire/firedtv-avc.c:999:45: warning: implicit
conversion from 'int' to 'char' changes value from 159 to -97
[-Wconstant-conversion]
app_info[0] = (EN50221_TAG_APP_INFO >> 16) & 0xff;
~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~
drivers/media/firewire/firedtv-avc.c:1000:45: warning: implicit
conversion from 'int' to 'char' changes value from 128 to -128
[-Wconstant-conversion]
app_info[1] = (EN50221_TAG_APP_INFO >> 8) & 0xff;
~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~
drivers/media/firewire/firedtv-avc.c:1040:44: warning: implicit
conversion from 'int' to 'char' changes value from 159 to -97
[-Wconstant-conversion]
app_info[0] = (EN50221_TAG_CA_INFO >> 16) & 0xff;
~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~
drivers/media/firewire/firedtv-avc.c:1041:44: warning: implicit
conversion from 'int' to 'char' changes value from 128 to -128
[-Wconstant-conversion]
app_info[1] = (EN50221_TAG_CA_INFO >> 8) & 0xff;
~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~
4 warnings generated.
Change app_info's type to unsigned char to match the type of the
member msg in struct ca_msg, which is the only thing passed into the
app_info parameter in this function.
This happens because the code calls get_lppaca() which calls
get_paca() and it checks if preemption is disabled through
check_preemption_disabled().
Preemption should be disabled because the per CPU variable may make no
sense if there is a preemption (and a CPU switch) after it reads the
per CPU data and when it is used.
In this device driver specifically, it is not a problem, because this
code just needs to have access to one lppaca struct, and it does not
matter if it is the current per CPU lppaca struct or not (i.e. when
there is a preemption and a CPU migration).
That said, the most appropriate fix seems to be related to avoiding
the debug_smp_processor_id() call at get_paca(), instead of calling
preempt_disable() before get_paca().
Currently xmon needs to get devtree_lock (through rtas_token()) during its
invocation (at crash time). If there is a crash while devtree_lock is being
held, then xmon tries to get the lock but spins forever and never get into
the interactive debugger, as in the following case:
int *ptr = NULL;
raw_spin_lock_irqsave(&devtree_lock, flags);
*ptr = 0xdeadbeef;
This patch avoids calling rtas_token(), thus trying to get the same lock,
at crash time. This new mechanism proposes getting the token at
initialization time (xmon_init()) and just consuming it at crash time.
This would allow xmon to be possible invoked independent of devtree_lock
being held or not.
Currently, disconnecting a USB webcam while it is in use prints out a
number of warnings, such as:
WARNING: CPU: 2 PID: 3118 at /build/linux-ezBi1T/linux-4.8.0/fs/sysfs/group.c:237 sysfs_remove_group+0x8b/0x90
sysfs group ffffffffa7cd0780 not found for kobject 'event13'
This has been noticed before. [0]
This is because of the order in which things are torn down.
If there are no streams active during a USB disconnect:
- uvc_disconnect() is invoked via device_del() through the bus
notifier mechanism.
- this calls uvc_unregister_video().
- uvc_unregister_video() unregisters the video device for each
stream,
- because there are no streams open, it calls uvc_delete()
- uvc_delete() calls uvc_status_cleanup(), which cleans up the status
input device.
- uvc_delete() calls media_device_unregister(), which cleans up the
media device
- uvc_delete(), uvc_unregister_video() and uvc_disconnect() all
return, and we end up back in device_del().
- device_del() then cleans up the sysfs folder for the camera with
dpm_sysfs_remove(). Because uvc_status_cleanup() and
media_device_unregister() have already been called, this all works
nicely.
If, on the other hand, there *are* streams active during a USB disconnect:
- uvc_disconnect() is invoked
- this calls uvc_unregister_video()
- uvc_unregister_video() unregisters the video device for each
stream,
- uvc_unregister_video() and uvc_disconnect() return, and we end up
back in device_del().
- device_del() then cleans up the sysfs folder for the camera with
dpm_sysfs_remove(). Because the status input device and the media
device are children of the USB device, this also deletes their
sysfs folders.
- Sometime later, the final stream is closed, invoking uvc_release().
- uvc_release() calls uvc_delete()
- uvc_delete() calls uvc_status_cleanup(), which cleans up the status
input device. Because the sysfs directory has already been removed,
this causes a WARNing.
- uvc_delete() calls media_device_unregister(), which cleans up the
media device. Because the sysfs directory has already been removed,
this causes another WARNing.
To fix this, we need to make sure the devices are always unregistered
before the end of uvc_disconnect(). To this, move the unregistration
into the disconnect path:
- split uvc_status_cleanup() into two parts, one on disconnect that
unregisters and one on delete that frees.
- move v4l2_device_unregister() and media_device_unregister() into
the disconnect path.
[0]: https://lkml.org/lkml/2016/12/8/657
[Renamed uvc_input_cleanup() to uvc_input_unregister()]
The ramoops backend currently calls persistent_ram_save_old() even
if a buffer is empty. While this appears to work, it is does not seem
like the right thing to do and could lead to future bugs so lets avoid
that. It also prevents misleading prints in the logs which claim the
buffer is valid.
I got something like:
found existing buffer, size 0, start 0
When I was expecting:
no valid data in buffer (sig = ...)
This bails out early (and reports with pr_debug()), since it's an
acceptable state.
As the commit 2893c379461a ("clk: make strings in parent name arrays
const"), let's make the parent strings const, otherwise we may meet
the following warning when compiling:
drivers/clk/imx/clk-imx7ulp.c: In function 'imx7ulp_clocks_init':
drivers/clk/imx/clk-imx7ulp.c:73:35: warning: passing argument 5 of
'imx_clk_mux_flags' discards 'const' qualifier from pointer target type
clks[IMX7ULP_CLK_APLL_PRE_SEL] = imx_clk_mux_flags("apll_pre_sel", base + 0x508, 0,
1, pll_pre_sels, ARRAY_SIZE(pll_pre_sels), CLK_SET_PARENT_GATE);
^
In file included from drivers/clk/imx/clk-imx7ulp.c:23:0:
drivers/clk/imx/clk.h:200:27: note: expected 'const char **' but argument is
of type 'const char * const*'
...
Cc: Stephen Boyd <sboyd@codeaurora.org> Cc: Michael Turquette <mturquette@baylibre.com> Cc: Shawn Guo <shawnguo@kernel.org> Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com> Signed-off-by: Stephen Boyd <sboyd@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
jffs2_sync_fs makes the assumption that if CONFIG_JFFS2_FS_WRITEBUFFER
is defined then a write buffer is available and has been initialized.
However, this does is not the case when the mtd device has no
out-of-band buffer:
int jffs2_nand_flash_setup(struct jffs2_sb_info *c)
{
if (!c->mtd->oobsize)
return 0;
...
The resulting call to cancel_delayed_work_sync passing a uninitialized
(but zeroed) delayed_work struct forces lockdep to become disabled.
When building the kernel with Clang, some disabled warnings appear
because this Makefile overrides KBUILD_CFLAGS for x86{,_64}. Add them to
this list so that the build is clean again.
-Wpointer-sign was disabled for the whole kernel before the beginning of Git history.
-Waddress-of-packed-member was disabled for the whole kernel and for
the early boot code in these commits:
-Wgnu was disabled for the whole kernel and for the early boot code in
these commits:
61163efae020 ("kbuild: LLVMLinux: Add Kbuild support for building kernel with Clang") 6c3b56b19730 ("x86/boot: Disable Clang warnings about GNU extensions").
[ mingo: Made the changelog more readable. ]
Tested-by: Sedat Dilek <sedat.dilek@gmail.com> Signed-off-by: Nathan Chancellor <natechancellor@gmail.com> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Reviewed-by: Sedat Dilek <sedat.dilek@gmail.com> Cc: Andy Lutomirski <luto@kernel.org> Cc: Arend van Spriel <arend.vanspriel@broadcom.com> Cc: Bhupesh Sharma <bhsharma@redhat.com> Cc: Borislav Petkov <bp@alien8.de> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Eric Snowberg <eric.snowberg@oracle.com> Cc: Hans de Goede <hdegoede@redhat.com> Cc: Joe Perches <joe@perches.com> Cc: Jon Hunter <jonathanh@nvidia.com> Cc: Julien Thierry <julien.thierry@arm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Marc Zyngier <marc.zyngier@arm.com> Cc: Matt Fleming <matt@codeblueprint.co.uk> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Sai Praneeth Prakhya <sai.praneeth.prakhya@intel.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: YiFei Zhu <zhuyifei1999@gmail.com> Cc: linux-efi@vger.kernel.org Link: http://lkml.kernel.org/r/20181129171230.18699-8-ard.biesheuvel@linaro.org Link: https://github.com/ClangBuiltLinux/linux/issues/112 Signed-off-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
FRWR memory registration is done with a series of calls and WRs.
1. ULP invokes ib_dma_map_sg()
2. ULP invokes ib_map_mr_sg()
3. ULP posts an IB_WR_REG_MR on the Send queue
Step 2 generates an iova. It is permissible for ULPs to change this
iova (with certain restrictions) between steps 2 and 3.
rxe_map_mr_sg captures the MR's iova but later when rxe processes the
REG_MR WR, it ignores the MR's iova field. If a ULP alters the MR's iova
after step 2 but before step 3, rxe never captures that change.
When the remote sends an RDMA Read targeting that MR, rxe looks up the
R_key, but the altered iova does not match the iova stored in the MR,
causing the RDMA Read request to fail.
Reported-by: Anna Schumaker <schumaker.anna@gmail.com> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Jason Gunthorpe <jgg@mellanox.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108709 Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Alex Deucher <alexdeucher@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
pcm3168 codec support runtime_[resume|suspend], whenever it
is not active, it enters suspend mode, and it's clock and regulators
will be disabled. so there is no need to disable them again in
remove callback. Otherwise we got following kernel warnings,
when unload pcm3168a driver
If a superblock has the MS_SUBMOUNT flag set, we should always allow
mounting it. These mounts are done automatically by the kernel either as
part of mounting some parent mount (e.g. debugfs always mounts tracefs
under "tracing" for compatibility) or they are mounted automatically as
needed on subdirectory accesses (e.g. NFS crossmnt mounts). Since such
automounts are either an implicit consequence of the parent mount (which
is already checked) or they can happen during regular accesses (where it
doesn't make sense to check against the current task's context), the
mount permission check should be skipped for them.
Without this patch, attempts to access contents of an automounted
directory can cause unexpected SELinux denials.
In the current kernel tree, the MS_SUBMOUNT flag is set only via
vfs_submount(), which is called only from the following places:
- AFS, when automounting special "symlinks" referencing other cells
- CIFS, when automounting "referrals"
- NFS, when automounting subtrees
- debugfs, when automounting tracefs
In all cases the submounts are meant to be transparent to the user and
it makes sense that if mounting the master is allowed, then so should be
the automounts. Note that CAP_SYS_ADMIN capability checking is already
skipped for (SB_KERNMOUNT|SB_SUBMOUNT) in:
- sget_userns() in fs/super.c:
if (!(flags & (SB_KERNMOUNT|SB_SUBMOUNT)) &&
!(type->fs_flags & FS_USERNS_MOUNT) &&
!capable(CAP_SYS_ADMIN))
return ERR_PTR(-EPERM);
- sget() in fs/super.c:
/* Ensure the requestor has permissions over the target filesystem */
if (!(flags & (SB_KERNMOUNT|SB_SUBMOUNT)) && !ns_capable(user_ns, CAP_SYS_ADMIN))
return ERR_PTR(-EPERM);
Verified internally on patched RHEL 7.6 with a reproducer using
NFS+httpd and selinux-tesuite.
Fixes: 93faccbbfa95 ("fs: Better permission checking for submounts") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
This patch adds a safety connection way for "forced_b_device" with
"workaround_for_vbus" like below:
< Example for R-Car E3 Ebisu >
# modprobe <any usb gadget driver>
# echo 1 > /sys/kernel/debug/ee020000.usb/b_device
(connect a usb cable to host side.)
# echo 2 > /sys/kernel/debug/ee020000.usb/b_device
Previous code should have connected a usb cable before the "b_device"
is set to 1 on the Ebisu board. However, if xHCI driver on the board
is probed, it causes some troubles:
- Conflicts USB VBUS/signals between the board and another host.
- "Cannot enable. Maybe the USB cable is bad?" might happen on
both the board and another host with a usb hub.
- Cannot enumerate a usb gadget correctly because an interruption
of VBUS change happens unexpectedly.
Currently, kprobe_events failure won't be handled properly.
Due to calling system() indirectly to write to kprobe_events,
it can't be identified whether an error is derived from kprobe or system.
For example, running ./tracex7 sample in ext4 partition,
"echo p:open_ctree open_ctree >> /s/k/d/t/kprobe_events"
gets 256 error code system() failure.
=> The error comes from kprobe, but it's not handled correctly.
According to man of system(3), it's return value
just passes the termination status of the child shell
rather than treating the error as -1. (don't care success)
Which means, currently it's not working as desired.
(According to the upper code snippet)
ex) running ./tracex7 with ext4 env.
# Current Output
sh: echo: I/O error
failed to open event open_ctree
# Desired Output
failed to create kprobe 'open_ctree' error 'No such file or directory'
The problem is, error can't be verified whether from child ps
or system. But using write() directly can verify the command
failure, and it will treat all error as -1. So I suggest using
write() directly to 'kprobe_events' rather than calling system().
Signed-off-by: Daniel T. Lee <danieltimlee@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
The public S805 datasheet only mentions that
HHI_SYS_CPU_CLK_CNTL1[20:29] contains a divider called "cpu_scale_div".
Unfortunately it does not mention how to use the register contents.
The Amlogic 3.10 GPL kernel sources are using the following code to
calculate the CPU clock based on that register (taken from
arch/arm/mach-meson8/clock.c in the 3.10 Amlogic kernel, shortened to
make it easier to read):
N = (aml_read_reg32(P_HHI_SYS_CPU_CLK_CNTL1) >> 20) & 0x3FF;
if (sel == 3) /* use cpu_scale_div */
div = 2 * N;
else
div = ... /* not relevant for this example */
cpu_clk = parent_clk / div;
This suggests that the formula is: parent_rate / 2 * register_value
However, running perf (which can measure the CPU clock rate thanks to
the ARM PMU) shows that this formula is not correct.
This can be reproduced with the following steps:
1. boot into u-boot
2. let the CPU clock run off the XTAL clock:
mw.l 0xC110419C 0x30 1
3. set the cpu_scale_div register:
to value 0x1: mw.l 0xC110415C 0x801016A2 1
to value 0x2: mw.l 0xC110415C 0x802016A2 1
to value 0x5: mw.l 0xC110415C 0x805016A2 1
4. let the CPU clock run off cpu_scale_div:
mw.l 0xC110419C 0xbd 1
5. boot Linux
6. run: perf stat -aB stress --cpu 4 --timeout 10
7. check the "cycles" value
I get the following results depending on the cpu_scale_div value:
- (cpu_in_sel - this is the input clock for cpu_scale_div - runs at
1.2GHz)
- 0x1 = 300MHz
- 0x2 = 200MHz
- 0x5 = 100MHz
This means that the actual formula to calculate the output of the
cpu_scale_div clock is: parent_rate / 2 * (register value + 1).
The register value 0x0 is reserved. When letting the CPU clock run off
the cpu_scale_div while the value is 0x0 the whole board hangs (even in
u-boot).
I also verified this with the TWD timer: when adding this to the .dts
without specifying it's clock it will auto-detect the PERIPH (which is
the input clock of the TWD) clock rate (and the result is shown in the
kernel log). On Meson8, Meson8b and Meson8m2 the PERIPH clock is CPUCLK
divided by 4. This also matched for all three test-cases from above (in
all cases the TWD timer clock rate was approx. one fourth of the CPU
clock rate).
A small note regarding the "fixes" tag: the original issue seems to
exist virtually since forever. Even commit 28b9fcd016126e ("clk:
meson8b: Add support for Meson8b clocks") seems to handle this wrong. I
still decided to use commit 251b6fd38bcb9c ("clk: meson: rework meson8b
cpu clock") because this is the first commit which gets the CPU hiearchy
correct and thus it's the first commit where the cpu_scale_div register
is used correctly (apart from the bug in the cpu_scale_table).
Consider the following scenario:
1. nonblocking enable crtc
2. wait for the event
3. nonblocking disable crtc
On i915 this can lead to a spurious -EBUSY from step 3 on
account of non-enabled planes getting the fake_commit in step 1
and we don't complete the fake_commit-> flip_done until
drm_atomic_helper_commit_hw_done() which can happen a long
time after the flip event was sent out.
This will become somewhat easy to hit on SKL+ once we start
to add all the planes for the crtc to every modeset commit
for the purposes of forcing a watermark register programming
[1].
To make the race a little less pronounced let's complete
fake_commit->flip_done after drm_atomic_helper_wait_for_flip_done().
For the single crtc case this should make the race quite
theoretical, assuming drm_atomic_helper_wait_for_flip_done()
actually has to wait for the real commit flip_done. In case
the real commit flip_done gets completed singificantly before
drm_atomic_helper_wait_for_flip_done(), or we are dealing with
multiple crtcs whose vblanks don't line up nicely the race still
exists.
The armv8_pmuv3 driver doesn't have a remove function, and when the test
'CONFIG_DEBUG_TEST_DRIVER_REMOVE=y' is enabled, the following Call trace
can be seen.
Rework to set suppress_bind_attrs flag to avoid removing the device when
CONFIG_DEBUG_TEST_DRIVER_REMOVE=y, since there's no real reason to
remove the armv8_pmuv3 driver.
Cc: Arnd Bergmann <arnd@arndb.de> Co-developed-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Anders Roxell <anders.roxell@linaro.org> Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
ecc_point_mult is supposed to be used with a regularized scalar,
otherwise, it's possible to deduce the position of the top bit of the
scalar with timing attack. This is important when the scalar is a
private key.
ecc_point_mult is already using a regular algorithm (i.e. having an
operation flow independent of the input scalar) but regularization step
is not implemented.
Arrange scalar to always have fixed top bit by adding a multiple of the
curve order (n).
References:
The constant time regularization step is based on micro-ecc by Kenneth
MacKay and also referenced in the literature (Bernstein, D. J., & Lange,
T. (2017). Montgomery curves and the Montgomery ladder. (Cryptology
ePrint Archive; Vol. 2017/293). s.l.: IACR. Chapter 4.6.2.)
The Broadcom SiByte BCM1250, BCM1125, and BCM1125H SOCs have an onchip
DRAM controller that supports memory amounts of up to 16GiB, and due to
how the address decoder has been wired in the SOC any memory beyond 1GiB
is actually mapped starting from 4GiB physical up, that is beyond the
32-bit addressable limit[1]. Consequently if the maximum amount of
memory has been installed, then it will span up to 19GiB.
Many of the evaluation boards we support that are based on one of these
SOCs have their memory soldered and the amount present fits in the
32-bit address range. The BCM91250A SWARM board however has actual DIMM
slots and accepts, depending on the peripherals revision of the SOC, up
to 4GiB or 8GiB of memory in commercially available JEDEC modules[2].
I believe this is also the case with the BCM91250C2 LittleSur board.
This means that up to either 3GiB or 7GiB of memory requires 64-bit
addressing to access.
I believe the BCM91480B BigSur board, which has the BCM1480 SOC instead,
accepts at least as much memory, although I have no documentation or
actual hardware available to verify that.
Both systems have PCI slots installed for use by any PCI option boards,
including ones that only support 32-bit addressing (additionally the
32-bit PCI host bridge of the BCM1250, BCM1125, and BCM1125H SOCs limits
addressing to 32-bits), and there is no IOMMU available. Therefore for
PCI DMA to work in the presence of memory beyond enable swiotlb for the
affected systems.
All the other SOC onchip DMA devices use 40-bit addressing and therefore
can address the whole memory, so only enable swiotlb if PCI support and
support for DMA beyond 4GiB have been both enabled in the configuration
of the kernel.
Add the proper includes and make smca_get_name() static.
Fix an actual bug too which the warning triggered:
arch/x86/kernel/cpu/mcheck/therm_throt.c:395:39: error: conflicting \
types for ‘smp_thermal_interrupt’
asmlinkage __visible void __irq_entry smp_thermal_interrupt(struct pt_regs *r)
^~~~~~~~~~~~~~~~~~~~~
In file included from arch/x86/kernel/cpu/mcheck/therm_throt.c:29:
./arch/x86/include/asm/traps.h:107:17: note: previous declaration of \
‘smp_thermal_interrupt’ was here
asmlinkage void smp_thermal_interrupt(void);
This commit adds support for APOGEE duet FireWire, launched 2007, already
discontinued. This model uses Oxford Semiconductor FW971 as its
communication engine. Below is information on Configuration ROM of this
unit. The unit supports some AV/C commands defined by Audio subunit
specification and vendor dependent commands.
Currently BPF verifier allows narrow loads for a context field only with
offset zero. E.g. if there is a __u32 field then only the following
loads are permitted:
* off=0, size=1 (narrow);
* off=0, size=2 (narrow);
* off=0, size=4 (full).
On the other hand LLVM can generate a load with offset different than
zero that make sense from program logic point of view, but verifier
doesn't accept it.
E.g. tools/testing/selftests/bpf/sendmsg4_prog.c has code:
where `*(u8 *)(r1 + 7)` means narrow load for ctx->user_ip4 with size=1
and offset=3 (7 - sizeof(ctx->user_family) = 3). This load is currently
rejected by verifier.
Verifier code that rejects such loads is in bpf_ctx_narrow_access_ok()
what means any is_valid_access implementation, that uses the function,
works this way, e.g. bpf_skb_is_valid_access() for __sk_buff or
sock_addr_is_valid_access() for bpf_sock_addr.
The patch makes such loads supported. Offset can be in [0; size_default)
but has to be multiple of load size. E.g. for __u32 field the following
loads are supported now:
* off=0, size=1 (narrow);
* off=1, size=1 (narrow);
* off=2, size=1 (narrow);
* off=3, size=1 (narrow);
* off=0, size=2 (narrow);
* off=2, size=2 (narrow);
* off=0, size=4 (full).
When the test 'CONFIG_DEBUG_TEST_DRIVER_REMOVE=y' is enabled,
arch_initcall(pl011_init) came before subsys_initcall(default_bdi_init).
devtmpfs gets killed because we try to remove a file and decrement the
wb reference count before the noop_backing_device_info gets initialized.
Rework to set suppress_bind_attrs flag to avoid removing the device when
CONFIG_DEBUG_TEST_DRIVER_REMOVE=y. This applies for pic32_uart and
xilinx_uartps as well.
This happened while running in qemu-system-aarch64, the AMBA PL011 UART
driver when enabling CONFIG_DEBUG_TEST_DRIVER_REMOVE.
arch_initcall(pl011_init) came before subsys_initcall(default_bdi_init),
devtmpfs' handle_remove() crashes because the reference count is a NULL
pointer only because wb->bdi hasn't been initialized yet.
Rework so that wb_put have an extra check if wb->bdi before decrement
wb->refcnt and also add a WARN_ON_ONCE to get a warning if it happens again
in other drivers.
The refcount of a newly added overlay node decrements to one
(instead of zero) when the overlay changeset is destroyed. This
change will cause the final decrement be to zero.
After applying this patch, new validation warnings will be
reported from the devicetree unittest during boot due to
a pre-existing devicetree bug. The warnings will be similar to:
OF: ERROR: memory leak before free overlay changeset, /testcase-data/overlay-node/test-bus/test-unittest4
This pre-existing devicetree bug will also trigger a WARN_ONCE() from
refcount_sub_and_test_checked() when an overlay changeset is
destroyed without having first been applied. This scenario occurs
when an error in the overlay is detected during the overlay changeset
creation:
(unwind_backtrace) from (show_stack+0x10/0x14)
(show_stack) from (dump_stack+0x6c/0x8c)
(dump_stack) from (__warn+0xdc/0x104)
(__warn) from (warn_slowpath_fmt+0x44/0x6c)
(warn_slowpath_fmt) from (refcount_sub_and_test_checked+0xa8/0xbc)
(refcount_sub_and_test_checked) from (kobject_put+0x24/0x208)
(kobject_put) from (of_changeset_destroy+0x2c/0xb4)
(of_changeset_destroy) from (free_overlay_changeset+0x1c/0x9c)
(free_overlay_changeset) from (of_overlay_remove+0x284/0x2cc)
(of_overlay_remove) from (of_unittest_apply_revert_overlay_check.constprop.4+0xf8/0x1e8)
(of_unittest_apply_revert_overlay_check.constprop.4) from (of_unittest_overlay+0x960/0xed8)
(of_unittest_overlay) from (of_unittest+0x1cc4/0x2138)
(of_unittest) from (do_one_initcall+0x4c/0x28c)
(do_one_initcall) from (kernel_init_freeable+0x29c/0x378)
(kernel_init_freeable) from (kernel_init+0x8/0x110)
(kernel_init) from (ret_from_fork+0x14/0x2c)
Tested-by: Alan Tull <atull@kernel.org> Signed-off-by: Frank Rowand <frank.rowand@sony.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
libbpf is now able to load successfully test_l4lb_noinline.o and
samples/bpf/tracex3_kern.o.
For the test_l4lb_noinline, uncomment related tests from test_libbpf.c
and remove the associated "TODO".
For tracex3_kern.o, instead of loading a program from samples/bpf/ that
might not have been compiled at this stage, try loading a program from
BPF selftests. Since this test case is about loading a program compiled
without the "-target bpf" flag, change the Makefile to compile one
program accordingly (instead of passing the flag for compiling all
programs).
Regarding test_xdp_noinline.o: in its current shape the program fails to
load because it provides no version section, but the loader needs one.
The test was added to make sure that libbpf could load XDP programs even
if they do not provide a version number in a dedicated section. But
libbpf is already capable of doing that: in our case loading fails
because the loader does not know that this is an XDP program (it does
not need to, since it does not attach the program). So trying to load
test_xdp_noinline.o does not bring much here: just delete this subtest.
For the record, the error message obtained with tracex3_kern.o was
fixed by commit e3d91b0ca523 ("tools/libbpf: handle issues with bpf ELF
objects containing .eh_frames")
I have not been abled to reproduce the "libbpf: incorrect bpf_call
opcode" error for test_l4lb_noinline.o, even with the version of libbpf
present at the time when test_libbpf.sh and test_libbpf_open.c were
created.
RFC -> v1:
- Compile test_xdp without the "-target bpf" flag, and try to load it
instead of ../../samples/bpf/tracex3_kern.o.
- Delete test_xdp_noinline.o subtest.
During HARD_RESET the data link is disconnected.
For self powered device, the spec is advising against doing that.
>From USB_PD_R3_0
7.1.5 Response to Hard Resets
Device operation during and after a Hard Reset is defined as follows:
Self-powered devices Should Not disconnect from USB during a Hard Reset
(see Section 9.1.2).
Bus powered devices will disconnect from USB during a Hard Reset due to the
loss of their power source.
Tackle this by letting TCPM know whether the device is self or bus powered.
This overcomes unnecessary port disconnections from hard reset.
Also, speeds up the enumeration time when connected to Type-A ports.
Signed-off-by: Badhri Jagan Sridharan <badhri@google.com> Reviewed-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
---------
Version history:
V3:
Rebase on top of usb-next
V2:
Based on feedback from heikki.krogerus@linux.intel.com
- self_powered added to the struct tcpm_port which is populated from
a. "connector" node of the device tree in tcpm_fw_get_caps()
b. "self_powered" node of the tcpc_config in tcpm_copy_caps
Based on feedbase from linux@roeck-us.net
- Code was refactored
- SRC_HARD_RESET_VBUS_OFF sets the link state to false based
on self_powered flag
It seems with some NICs supported by the e1000e driver a SYSTIM reading
may occasionally be few microseconds before the previous reading and if
enabled also pass e1000e_sanitize_systim() without reaching the maximum
number of rereads, even if the function is modified to check three
consecutive readings (i.e. it doesn't look like a double read error).
This causes an underflow in the timecounter and the PHC time jumps hours
ahead.
This was observed on 82574, I217 and I219. The fastest way to reproduce
it is to run a program that continuously calls the PTP_SYS_OFFSET ioctl
on the PHC.
Modify e1000e_phc_gettime() to use timecounter_cyc2time() instead of
timecounter_read() in order to allow non-monotonic SYSTIM readings and
prevent the PHC from jumping.
Cc: Richard Cochran <richardcochran@gmail.com> Signed-off-by: Miroslav Lichvar <mlichvar@redhat.com> Acked-by: Jacob Keller <jacob.e.keller@intel.com> Tested-by: Aaron Brown <aaron.f.brown@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In the past, Asus firmwares would change the panel backlight directly
through the EC when the display off hotkey (Fn+F7) was pressed, and
only notify the OS of such change, with 0x33 when the LCD was ON and
0x34 when the LCD was OFF. These are currently mapped to
KEY_DISPLAYTOGGLE and KEY_DISPLAY_OFF, respectively.
Most recently the EC on Asus most machines lost ability to toggle the
LCD backlight directly, but unless the OS informs the firmware it is
going to handle the display toggle hotkey events, the firmware still
tries change the brightness through the EC, to no effect. The end result
is a long list (at Endless we counted 11) of Asus laptop models where
the display toggle hotkey does not perform any action. Our firmware
engineers contacts at Asus were surprised that there were still machines
out there with the old behavior.
Calling WMNB(ASUS_WMI_DEVID_BACKLIGHT==0x00050011, 2) on the _WDG device
tells the firmware that it should let the OS handle the display toggle
event, in which case it will simply notify the OS of a key press with
0x35, as shown by the DSDT excerpts bellow.
Signed-off-by: João Paulo Rechi Vita <jprvita@endlessm.com> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When it's possible that the PF might end up trying to send a
packet to one of its own VFs, we have to forbid IPsec offload
because the device drops the packets into a black hole.
See commit 47b6f50077e6 ("ixgbe: disallow IPsec Tx offload
when in SR-IOV mode") for more info.
This really is only necessary when the device is in the default
VEB mode. If instead the device is running in VEPA mode,
the packets will go through the encryption engine and out the
MAC/PHY as normal, and get "hairpinned" as needed by the switch.
So let's not block IPsec offload when in VEPA mode. To get
there with the ixgbe device, use the handy 'bridge' command:
bridge link set dev eth1 hwmode vepa
Signed-off-by: Shannon Nelson <shannon.nelson@oracle.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Christian König <christian.koenig@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The igt@kms_plane@pixel-format-pipe tests can create a sequence where
stream_state is NULL during amdgpu_dm_crtc_set_crc_source which results
in a null pointer dereference.
[How]
Guard against stream_state being NULL before accessing its fields. This
doesn't fix the root cause of the issue so a DRM_ERROR is generated
to still fail the tests.
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Reviewed-by: David Francis <David.Francis@amd.com> Acked-by: Leo Li <sunpeng.li@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Keeping the irq_chip definition static will make it shared with multiple
giochips in the system. This practice is considered to be bad and now we
will get the below warning from gpiolib core:
"detected irqchip that is shared with multiple gpiochips: please fix the
driver."
Hence, move the irq_chip definition from static to `struct pl061` for
using a unique irq_chip for each gpiochip.
There doesn't seem to be any reason to restrict MAC address
matching to source MAC addresses in set types bitmap:ipmac,
hash:ipmac and hash:mac. With this patch, and this setup:
ip netns add A
ip link add veth1 type veth peer name veth2 netns A
ip addr add 192.0.2.1/24 dev veth1
ip -net A addr add 192.0.2.2/24 dev veth2
ip link set veth1 up
ip -net A link set veth2 up
ip netns exec A ipset create test hash:mac
dst=$(ip netns exec A cat /sys/class/net/veth2/address)
ip netns exec A ipset add test ${dst}
ip netns exec A iptables -P INPUT DROP
ip netns exec A iptables -I INPUT -m set --match-set test dst -j ACCEPT
ipset will match packets based on destination MAC address:
Matteo reported forwarding issues inside the linux bridge,
if the enslaved interfaces use the fq qdisc.
Similar to commit 8203e2d844d3 ("net: clear skb->tstamp in
forwarding paths"), we need to clear the tstamp field in
the bridge forwarding path.
Fixes: 80b14dee2bea ("net: Add a new socket option for a future transmit time.") Fixes: fb420d5d91c1 ("tcp/fq: move back to CLOCK_MONOTONIC") Reported-and-tested-by: Matteo Croce <mcroce@redhat.com> Signed-off-by: Paolo Abeni <pabeni@redhat.com> Acked-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Acked-by: Roopa Prabhu <roopa@cumulusnetworks.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
I realized the last patch calls dev_get_by_index_rcu in a branch not
holding the rcu lock. Add the calls to rcu_read_lock and rcu_read_unlock.
Fixes: ec90ad334986 ("ipv6: Consider sk_bound_dev_if when binding a socket to a v4 mapped address") Signed-off-by: David Ahern <dsahern@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>