The PPS assert/clear offset corrections are set by the PPS_SETPARAMS
ioctl in the pps_ktime structs, which also contain flags. The flags are
not initialized by applications (using the timepps.h header) and they
are not used by the kernel for anything except returning them back in
the PPS_GETPARAMS ioctl.
Set the flags to zero to make it clear they are unused and avoid leaking
uninitialized data of the PPS_SETPARAMS caller to other applications
that have a read access to the PPS device.
Link: http://lkml.kernel.org/r/20190702092251.24303-1-mlichvar@redhat.com Signed-off-by: Miroslav Lichvar <mlichvar@redhat.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Rodolfo Giometti <giometti@enneenne.com> Cc: Greg KH <greg@kroah.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This makes the setproctitle() special case very explicit indeed, and
handles it with a separate helper function entirely. In the process, it
re-instates the original semantics of simply stopping at the first NUL
character when the original last NUL character is no longer there.
[ The original semantics can still be seen in mm/util.c: get_cmdline()
that is limited to a fixed-size buffer ]
This makes the logic about when we use the string lengths etc much more
obvious, and makes it easier to see what we do and what the two very
different cases are.
Note that even when we allow walking past the end of the argument array
(because the setproctitle() might have overwritten and overflowed the
original argv[] strings), we only allow it when it overflows into the
environment region if it is immediately adjacent.
[ Fixed for missing 'count' checks noted by Alexey Izbyshev ]
Start off with a clean slate that only reads exactly from arg_start to
arg_end, without any oddities. This simplifies the code and in the
process removes the case that caused us to potentially leak an
uninitialized byte from the temporary kernel buffer.
Note that in order to start from scratch with an understandable base,
this simplifies things _too_ much, and removes all the legacy logic to
handle setproctitle() having changed the argument strings.
We'll add back those special cases very differently in the next commit.
When going through execve(), zero out the NUMA fault statistics instead of
freeing them.
During execve, the task is reachable through procfs and the scheduler. A
concurrent /proc/*/sched reader can read data from a freed ->numa_faults
allocation (confirmed by KASAN) and write it back to userspace.
I believe that it would also be possible for a use-after-free read to occur
through a race between a NUMA fault and execve(): task_numa_fault() can
lead to task_numa_compare(), which invokes task_weight() on the currently
running task of a different CPU.
Another way to fix this would be to make ->numa_faults RCU-managed or add
extra locking, but it seems easier to wipe the NUMA fault statistics on
execve.
Signed-off-by: Jann Horn <jannh@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Petr Mladek <pmladek@suse.com> Cc: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Will Deacon <will@kernel.org> Fixes: 82727018b0d3 ("sched/numa: Call task_numa_free() from do_execve()") Link: https://lkml.kernel.org/r/20190716152047.14424-1-jannh@google.com Signed-off-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Certain ttys operations (pty_unix98_ops) lack tiocmget() and tiocmset()
functions which are called by the certain HCI UART protocols (hci_ath,
hci_bcm, hci_intel, hci_mrvl, hci_qca) via hci_uart_set_flow_control()
or directly. This leads to an execution at NULL and can be triggered by
an unprivileged user. Fix this by adding a helper function and a check
for the missing tty operations in the protocols code.
This fixes CVE-2019-10207. The Fixes: lines list commits where calls to
tiocm[gs]et() or hci_uart_set_flow_control() were added to the HCI UART
protocols.
Link: https://syzkaller.appspot.com/bug?id=1b42faa2848963564a5b1b7f8c837ea7b55ffa50 Reported-by: syzbot+79337b501d6aa974d0f6@syzkaller.appspotmail.com Cc: stable@vger.kernel.org # v2.6.36+ Fixes: b3190df62861 ("Bluetooth: Support for Atheros AR300x serial chip") Fixes: 118612fb9165 ("Bluetooth: hci_bcm: Add suspend/resume PM functions") Fixes: ff2895592f0f ("Bluetooth: hci_intel: Add Intel baudrate configuration support") Fixes: 162f812f23ba ("Bluetooth: hci_uart: Add Marvell support") Fixes: fa9ad876b8e0 ("Bluetooth: hci_qca: Add support for Qualcomm Bluetooth chip wcn3990") Signed-off-by: Vladis Dronov <vdronov@redhat.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org> Reviewed-by: Yu-Chen, Cho <acho@suse.com> Tested-by: Yu-Chen, Cho <acho@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix a crash with multipath activated. It happends when ANA log
page is larger than MDTS and because of that ANA is disabled.
The driver then tries to access unallocated buffer when connecting
to a nvme target. The signature is as follows:
BUG: KASAN: use-after-free in __write_once_size include/linux/compiler.h:221
BUG: KASAN: use-after-free in hlist_del_rcu include/linux/rculist.h:455
BUG: KASAN: use-after-free in xfrm_hash_rebuild+0xa0d/0x1000 net/xfrm/xfrm_policy.c:1318
Write of size 8 at addr ffff888095e79c00 by task kworker/1:3/8066
Workqueue: events xfrm_hash_rebuild
Call Trace:
__write_once_size include/linux/compiler.h:221 [inline]
hlist_del_rcu include/linux/rculist.h:455 [inline]
xfrm_hash_rebuild+0xa0d/0x1000 net/xfrm/xfrm_policy.c:1318
process_one_work+0x814/0x1130 kernel/workqueue.c:2269
Allocated by task 8064:
__kmalloc+0x23c/0x310 mm/slab.c:3669
kzalloc include/linux/slab.h:742 [inline]
xfrm_hash_alloc+0x38/0xe0 net/xfrm/xfrm_hash.c:21
xfrm_policy_init net/xfrm/xfrm_policy.c:4036 [inline]
xfrm_net_init+0x269/0xd60 net/xfrm/xfrm_policy.c:4120
ops_init+0x336/0x420 net/core/net_namespace.c:130
setup_net+0x212/0x690 net/core/net_namespace.c:316
The faulting address is the address of the old chain head,
free'd by xfrm_hash_resize().
In xfrm_hash_rehash(), chain heads get re-initialized without
any hlist_del_rcu:
for (i = hmask; i >= 0; i--)
INIT_HLIST_HEAD(odst + i);
Then, hlist_del_rcu() gets called on the about to-be-reinserted policy
when iterating the per-net list of policies.
hlist_del_rcu() will then make chain->first be nonzero again:
static inline void __hlist_del(struct hlist_node *n)
{
struct hlist_node *next = n->next; // address of next element in list
struct hlist_node **pprev = n->pprev;// location of previous elem, this
// can point at chain->first
WRITE_ONCE(*pprev, next); // chain->first points to next elem
if (next)
next->pprev = pprev;
Then, when we walk chainlist to find insertion point, we may find a
non-empty list even though we're supposedly reinserting the first
policy to an empty chain.
To fix this first unlink all exact and inexact policies instead of
zeroing the list heads.
Add the commands equivalent to the syzbot reproducer to xfrm_policy.sh,
without fix KASAN catches the corruption as it happens, SLUB poisoning
detects it a bit later.
Reported-by: syzbot+0165480d4ef07360eeda@syzkaller.appspotmail.com Fixes: 1548bc4e0512 ("xfrm: policy: delete inexact policies from inexact list on hash rebuild") Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Change devm_k*alloc to k*alloc to manually allocate memory
The manual allocation and freeing of memory is necessary because when
the USB radio is disconnected, the memory associated with devm_k*alloc
is freed. Meaning if we still have unresolved references to the radio
device, then we get use-after-free errors.
This patch fixes this by manually allocating memory, and freeing it in
the v4l2.release callback that gets called when the last radio device
exits.
Reported-and-tested-by: syzbot+a4387f5b6b799f6becbf@syzkaller.appspotmail.com Signed-off-by: Luke Nowakowski-Krijger <lnowakow@eng.ucsd.edu> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
[hverkuil-cisco@xs4all.nl: cleaned up two small checkpatch.pl warnings]
[hverkuil-cisco@xs4all.nl: prefix subject with driver name] Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When the pvrusb2 driver detects that there's something wrong with the
device, it prints a warning message. Right now those message are
printed in two different formats:
1. ***WARNING*** message here
2. WARNING: message here
There's an issue with the second format. Syzkaller recognizes it as a
message produced by a WARN_ON(), which is used to indicate a bug in the
kernel. However pvrusb2 prints those warnings to indicate an issue with
the device, not the bug in the kernel.
This patch changes the pvrusb2 driver to consistently use the first
warning message format. This will unblock syzkaller testing of this
driver.
Kasan reported a use after free in cpia2_usb_disconnect()
It first freed everything and then woke up those waiting.
The reverse order is correct.
Fixes: 6c493f8b28c67 ("[media] cpia2: major overhaul to get it in a working state again") Signed-off-by: Oliver Neukum <oneukum@suse.com> Reported-by: syzbot+0c90fc937c84f97d0aa6@syzkaller.appspotmail.com Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 1dc92851849c ("bpf: kernel side support for BTF Var and DataSec")
added invocations of btf_type_is_resolve_source_only before
btf_type_nosize_or_null which checks for the NULL pointer.
Swap the order of btf_type_nosize_or_null and
btf_type_is_resolve_source_only to make sure the do the NULL pointer
check first.
Fixes: 1dc92851849c ("bpf: kernel side support for BTF Var and DataSec") Reported-by: syzbot <syzkaller@googlegroups.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Acked-by: Andrii Nakryiko <andriin@fb.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Phong Tran <tranmanphong@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The current vsock code for removal of socket from the list is both
subject to race and inefficient. It takes the lock, checks whether
the socket is in the list, drops the lock and if the socket was on the
list, deletes it from the list. This is subject to race because as soon
as the lock is dropped once it is checked for presence, that condition
cannot be relied upon for any decision. It is also inefficient because
if the socket is present in the list, it takes the lock twice.
Signed-off-by: Sunil Muthuswamy <sunilmut@microsoft.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Hrvoje reports that when a large fixed buffer is registered and IO is
being done to the latter pages of said buffer, the IO submission time
is much worse:
reading to the start of the buffer: 11238 ns
reading to the end of the buffer: 1039879 ns
In fact, it's worse by two orders of magnitude. The reason for that is
how io_uring figures out how to setup the iov_iter. We point the iter
at the first bvec, and then use iov_iter_advance() to fast-forward to
the offset within that buffer we need.
However, that is abysmally slow, as it entails iterating the bvecs
that we setup as part of buffer registration. There's really no need
to use this generic helper, as we know it's a BVEC type iterator, and
we also know that each bvec is PAGE_SIZE in size, apart from possibly
the first and last. Hence we can just use a shift on the offset to
find the right index, and then adjust the iov_iter appropriately.
After this fix, the timings are:
reading to the start of the buffer: 10135 ns
reading to the end of the buffer: 1377 ns
We could queue a work for each req in defer and link list without
increasing async_list->cnt, so we shouldn't decrease it while exiting
from workqueue as well if we didn't process the req in async list.
Thanks to Jens Axboe <axboe@kernel.dk> for his guidance.
Daniel reports that when testing an http server that uses io_uring
to poll for incoming connections, sometimes it hard crashes. This is
due to an uninitialized list member for the io_uring request. Normally
this doesn't trigger and none of the test cases caught it.
Reported-by: Daniel Kozak <kozzi11@gmail.com> Tested-by: Daniel Kozak <kozzi11@gmail.com> Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is a hang issue while using fio to do some basic test. The issue
can be easily reproduced using the below script:
while true
do
fio --ioengine=io_uring -rw=write -bs=4k -numjobs=1 \
-size=1G -iodepth=64 -name=uring --filename=/dev/zero
done
After several minutes (or more), fio would block at
io_uring_enter->io_cqring_wait in order to waiting for previously
committed sqes to be completed and can't return to user anymore until
we send a SIGTERM to fio. After receiving SIGTERM, fio hangs at
io_ring_ctx_wait_and_kill with a backtrace like this:
As atomic_ops.rst states, atomic_read does not guarantee that the
runtime modification by any other thread is visible yet, so we must take
care of that with a proper implicit or explicit memory barrier.
This issue was detected with the help of Jackie's <liuyun01@kylinos.cn>
It turns out that 'access()' (and 'faccessat()') can cause a lot of RCU
work because it installs a temporary credential that gets allocated and
freed for each system call.
The allocation and freeing overhead is mostly benign, but because
credentials can be accessed under the RCU read lock, the freeing
involves a RCU grace period.
Which is not a huge deal normally, but if you have a lot of access()
calls, this causes a fair amount of seconday damage: instead of having a
nice alloc/free patterns that hits in hot per-CPU slab caches, you have
all those delayed free's, and on big machines with hundreds of cores,
the RCU overhead can end up being enormous.
But it turns out that all of this is entirely unnecessary. Exactly
because access() only installs the credential as the thread-local
subjective credential, the temporary cred pointer doesn't actually need
to be RCU free'd at all. Once we're done using it, we can just free it
synchronously and avoid all the RCU overhead.
So add a 'non_rcu' flag to 'struct cred', which can be set by users that
know they only use it in non-RCU context (there are other potential
users for this). We can make it a union with the rcu freeing list head
that we need for the RCU case, so this doesn't need any extra storage.
Note that this also makes 'get_current_cred()' clear the new non_rcu
flag, in case we have filesystems that take a long-term reference to the
cred and then expect the RCU delayed freeing afterwards. It's not
entirely clear that this is required, but it makes for clear semantics:
the subjective cred remains non-RCU as long as you only access it
synchronously using the thread-local accessors, but you _can_ use it as
a generic cred if you want to.
It is possible that we should just remove the whole RCU markings for
->cred entirely. Only ->real_cred is really supposed to be accessed
through RCU, and the long-term cred copies that nfs uses might want to
explicitly re-enable RCU freeing if required, rather than have
get_current_cred() do it implicitly.
But this is a "minimal semantic changes" change for the immediate
problem.
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Acked-by: Eric Dumazet <edumazet@google.com> Acked-by: Paul E. McKenney <paulmck@linux.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Jan Glauber <jglauber@marvell.com> Cc: Jiri Kosina <jikos@kernel.org> Cc: Jayachandran Chandrasekharan Nair <jnair@marvell.com> Cc: Greg KH <greg@kroah.com> Cc: Kees Cook <keescook@chromium.org> Cc: David Howells <dhowells@redhat.com> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The idea behind keeping the saturation mask local to a context backfired
spectacularly. The premise with the local mask was that we would be more
proactive in attempting to use semaphores after each time the context
idled, and that all new contexts would attempt to use semaphores
ignoring the current state of the system. This turns out to be horribly
optimistic. If the system state is still oversaturated and the existing
workloads have all stopped using semaphores, the new workloads would
attempt to use semaphores and be deprioritised behind real work. The
new contexts would not switch off using semaphores until their initial
batch of low priority work had completed. Given sufficient backload load
of equal user priority, this would completely starve the new work of any
GPU time.
To compensate, remove the local tracking in favour of keeping it as
global state on the engine -- once the system is saturated and
semaphores are disabled, everyone stops attempting to use semaphores
until the system is idle again. One of the reason for preferring local
context tracking was that it worked with virtual engines, so for
switching to global state we could either do a complete check of all the
virtual siblings or simply disable semaphores for those requests. This
takes the simpler approach of disabling semaphores on virtual engines.
The downside is that the decision that the engine is saturated is a
local measure -- we are only checking whether or not this context was
scheduled in a timely fashion, it may be legitimately delayed due to user
priorities. We still have the same dilemma though, that we do not want
to employ the semaphore poll unless it will be used.
v2: Explain why we need to assume the worst wrt virtual engines.
The combination of KASAN_STACK and GCC_PLUGIN_STRUCTLEAK_BYREF
leads to much larger kernel stack usage, as seen from the warnings
about functions that now exceed the 2048 byte limit:
drivers/media/i2c/tvp5150.c:253:1: error: the frame size of 3936 bytes is larger than 2048 bytes
drivers/media/tuners/r820t.c:1327:1: error: the frame size of 2816 bytes is larger than 2048 bytes
drivers/net/wireless/broadcom/brcm80211/brcmsmac/phy/phy_n.c:16552:1: error: the frame size of 3144 bytes is larger than 2048 bytes [-Werror=frame-larger-than=]
fs/ocfs2/aops.c:1892:1: error: the frame size of 2088 bytes is larger than 2048 bytes
fs/ocfs2/dlm/dlmrecovery.c:737:1: error: the frame size of 2088 bytes is larger than 2048 bytes
fs/ocfs2/namei.c:1677:1: error: the frame size of 2584 bytes is larger than 2048 bytes
fs/ocfs2/super.c:1186:1: error: the frame size of 2640 bytes is larger than 2048 bytes
fs/ocfs2/xattr.c:3678:1: error: the frame size of 2176 bytes is larger than 2048 bytes
net/bluetooth/l2cap_core.c:7056:1: error: the frame size of 2144 bytes is larger than 2048 bytes [-Werror=frame-larger-than=]
net/bluetooth/l2cap_core.c: In function 'l2cap_recv_frame':
net/bridge/br_netlink.c:1505:1: error: the frame size of 2448 bytes is larger than 2048 bytes
net/ieee802154/nl802154.c:548:1: error: the frame size of 2232 bytes is larger than 2048 bytes
net/wireless/nl80211.c:1726:1: error: the frame size of 2224 bytes is larger than 2048 bytes
net/wireless/nl80211.c:2357:1: error: the frame size of 4584 bytes is larger than 2048 bytes
net/wireless/nl80211.c:5108:1: error: the frame size of 2760 bytes is larger than 2048 bytes
net/wireless/nl80211.c:6472:1: error: the frame size of 2112 bytes is larger than 2048 bytes
The structleak plugin was previously disabled for CONFIG_COMPILE_TEST,
but meant we missed some bugs, so this time we should address them.
The frame size warnings are distracting, and risking a kernel stack
overflow is generally not beneficial to performance, so it may be best
to disallow that particular combination. This can be done by turning
off either one. I picked the dependency in GCC_PLUGIN_STRUCTLEAK_BYREF
and GCC_PLUGIN_STRUCTLEAK_BYREF_ALL, as this option is designed to
make uninitialized stack usage less harmful when enabled on its own,
but it also prevents KASAN from detecting those cases in which it was
in fact needed.
KASAN_STACK is currently implied by KASAN on gcc, but could be made a
user selectable option if we want to allow combining (non-stack) KASAN
with GCC_PLUGIN_STRUCTLEAK_BYREF.
Note that it would be possible to specifically address the files that
print the warning, but presumably the overall stack usage is still
significantly higher than in other configurations, so this would not
address the full problem.
I could not test this with CONFIG_INIT_STACK_ALL, which may or may not
suffer from a similar problem.
In preparation for fixing a deadlock between wait_for_bus_probe_idle()
and the nvdimm_bus_list_mutex arrange for __nd_ioctl() without
nvdimm_bus_list_mutex held. This also unifies the 'dimm' and 'bus' level
ioctls into a common nd_ioctl() preamble implementation.
Marked for -stable as it is a pre-requisite for a follow-on fix.
Cc: <stable@vger.kernel.org> Fixes: bf9bccc14c05 ("libnvdimm: pmem label sets and namespace instantiation") Cc: Vishal Verma <vishal.l.verma@intel.com> Tested-by: Jane Chu <jane.chu@oracle.com> Link: https://lore.kernel.org/r/156341209518.292348.7183897251740665198.stgit@dwillia2-desk3.amr.corp.intel.com Signed-off-by: Dan Williams <dan.j.williams@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Namespace activation expects to be able to reference region badblocks.
The following warning sometimes triggers when asynchronous namespace
activation races in front of the completion of namespace probing. Move
all possible namespace probing after region badblocks initialization.
Otherwise, lockdep sometimes catches the uninitialized state of the
badblocks seqlock with stack trace signatures like:
INFO: trying to register non-static key.
pmem2: detected capacity change from 0 to 136365211648
the code is fine but needs lockdep annotation.
turning off the locking correctness validator.
CPU: 9 PID: 358 Comm: kworker/u80:5 Tainted: G OE 5.2.0-rc4+ #3382
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 0.0.0 02/06/2015
Workqueue: events_unbound async_run_entry_fn
Call Trace:
dump_stack+0x85/0xc0
pmem1.12: detected capacity change from 0 to 8589934592
register_lock_class+0x56a/0x570
? check_object+0x140/0x270
__lock_acquire+0x80/0x1710
? __mutex_lock+0x39d/0x910
lock_acquire+0x9e/0x180
? nd_pfn_validate+0x28f/0x440 [libnvdimm]
badblocks_check+0x93/0x1f0
? nd_pfn_validate+0x28f/0x440 [libnvdimm]
nd_pfn_validate+0x28f/0x440 [libnvdimm]
? lockdep_hardirqs_on+0xf0/0x180
nd_dax_probe+0x9a/0x120 [libnvdimm]
nd_pmem_probe+0x6d/0x180 [nd_pmem]
nvdimm_bus_probe+0x90/0x2c0 [libnvdimm]
Fixes: 48af2f7e52f4 ("libnvdimm, pfn: during init, clear errors...") Cc: <stable@vger.kernel.org> Cc: Vishal Verma <vishal.l.verma@intel.com> Reviewed-by: Vishal Verma <vishal.l.verma@intel.com> Link: https://lore.kernel.org/r/156341208365.292348.1547528796026249120.stgit@dwillia2-desk3.amr.corp.intel.com Signed-off-by: Dan Williams <dan.j.williams@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A multithreaded namespace creation/destruction stress test currently
fails with signatures like the following:
sysfs group 'power' not found for kobject 'dax1.1'
RIP: 0010:sysfs_remove_group+0x76/0x80
Call Trace:
device_del+0x73/0x370
device_unregister+0x16/0x50
nd_async_device_unregister+0x1e/0x30 [libnvdimm]
async_run_entry_fn+0x39/0x160
process_one_work+0x23c/0x5e0
worker_thread+0x3c/0x390
The libnvdimm subsystem arranges for devices to be destroyed as a result
of a sysfs operation. Since device_unregister() cannot be called from
an actively running sysfs attribute of the same device libnvdimm
arranges for device_unregister() to be performed in an out-of-line async
context.
The driver core maintains a 'dead' state for coordinating its own racing
async registration / de-registration requests. Rather than add local
'dead' state tracking infrastructure to libnvdimm device objects, export
the existing state tracking via a new kill_device() helper.
The kill_device() helper simply marks the device as dead, i.e. that it
is on its way to device_del(), or returns that the device was already
dead. This can be used in advance of calling device_unregister() for
subsystems like libnvdimm that might need to handle multiple user
threads racing to delete a device.
This refactoring does not change any behavior, but it is a pre-requisite
for follow-on fixes and therefore marked for -stable.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Fixes: 4d88a97aa9e8 ("libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver...") Cc: <stable@vger.kernel.org> Tested-by: Jane Chu <jane.chu@oracle.com> Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Link: https://lore.kernel.org/r/156341207332.292348.14959761496009347574.stgit@dwillia2-desk3.amr.corp.intel.com Signed-off-by: Dan Williams <dan.j.williams@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The stub function for !CONFIG_IOMMU_IOVA needs to be
'static inline'.
Fixes: effa467870c76 ('iommu/vt-d: Don't queue_iova() if there is no flush queue') Signed-off-by: Joerg Roedel <jroedel@suse.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since the cached32_node is allowed to be advanced above dma_32bit_pfn
(to provide a shortcut into the limited range), we need to be careful to
remove the to be freed node if it is the cached32_node.
[ 48.477773] BUG: KASAN: use-after-free in __cached_rbnode_delete_update+0x68/0x110
[ 48.477812] Read of size 8 at addr ffff88870fc19020 by task kworker/u8:1/37
[ 48.477843]
[ 48.477879] CPU: 1 PID: 37 Comm: kworker/u8:1 Tainted: G U 5.2.0+ #735
[ 48.477915] Hardware name: Intel Corporation NUC7i5BNK/NUC7i5BNB, BIOS BNKBL357.86A.0052.2017.0918.1346 09/18/2017
[ 48.478047] Workqueue: i915 __i915_gem_free_work [i915]
[ 48.478075] Call Trace:
[ 48.478111] dump_stack+0x5b/0x90
[ 48.478137] print_address_description+0x67/0x237
[ 48.478178] ? __cached_rbnode_delete_update+0x68/0x110
[ 48.478212] __kasan_report.cold.3+0x1c/0x38
[ 48.478240] ? __cached_rbnode_delete_update+0x68/0x110
[ 48.478280] ? __cached_rbnode_delete_update+0x68/0x110
[ 48.478308] __cached_rbnode_delete_update+0x68/0x110
[ 48.478344] private_free_iova+0x2b/0x60
[ 48.478378] iova_magazine_free_pfns+0x46/0xa0
[ 48.478403] free_iova_fast+0x277/0x340
[ 48.478443] fq_ring_free+0x15a/0x1a0
[ 48.478473] queue_iova+0x19c/0x1f0
[ 48.478597] cleanup_page_dma.isra.64+0x62/0xb0 [i915]
[ 48.478712] __gen8_ppgtt_cleanup+0x63/0x80 [i915]
[ 48.478826] __gen8_ppgtt_cleanup+0x42/0x80 [i915]
[ 48.478940] __gen8_ppgtt_clear+0x433/0x4b0 [i915]
[ 48.479053] __gen8_ppgtt_clear+0x462/0x4b0 [i915]
[ 48.479081] ? __sg_free_table+0x9e/0xf0
[ 48.479116] ? kfree+0x7f/0x150
[ 48.479234] i915_vma_unbind+0x1e2/0x240 [i915]
[ 48.479352] i915_vma_destroy+0x3a/0x280 [i915]
[ 48.479465] __i915_gem_free_objects+0xf0/0x2d0 [i915]
[ 48.479579] __i915_gem_free_work+0x41/0xa0 [i915]
[ 48.479607] process_one_work+0x495/0x710
[ 48.479642] worker_thread+0x4c7/0x6f0
[ 48.479687] ? process_one_work+0x710/0x710
[ 48.479724] kthread+0x1b2/0x1d0
[ 48.479774] ? kthread_create_worker_on_cpu+0xa0/0xa0
[ 48.479820] ret_from_fork+0x1f/0x30
[ 48.479864]
[ 48.479907] Allocated by task 631:
[ 48.479944] save_stack+0x19/0x80
[ 48.479994] __kasan_kmalloc.constprop.6+0xc1/0xd0
[ 48.480038] kmem_cache_alloc+0x91/0xf0
[ 48.480082] alloc_iova+0x2b/0x1e0
[ 48.480125] alloc_iova_fast+0x58/0x376
[ 48.480166] intel_alloc_iova+0x90/0xc0
[ 48.480214] intel_map_sg+0xde/0x1f0
[ 48.480343] i915_gem_gtt_prepare_pages+0xb8/0x170 [i915]
[ 48.480465] huge_get_pages+0x232/0x2b0 [i915]
[ 48.480590] ____i915_gem_object_get_pages+0x40/0xb0 [i915]
[ 48.480712] __i915_gem_object_get_pages+0x90/0xa0 [i915]
[ 48.480834] i915_gem_object_prepare_write+0x2d6/0x330 [i915]
[ 48.480955] create_test_object.isra.54+0x1a9/0x3e0 [i915]
[ 48.481075] igt_shared_ctx_exec+0x365/0x3c0 [i915]
[ 48.481210] __i915_subtests.cold.4+0x30/0x92 [i915]
[ 48.481341] __run_selftests.cold.3+0xa9/0x119 [i915]
[ 48.481466] i915_live_selftests+0x3c/0x70 [i915]
[ 48.481583] i915_pci_probe+0xe7/0x220 [i915]
[ 48.481620] pci_device_probe+0xe0/0x180
[ 48.481665] really_probe+0x163/0x4e0
[ 48.481710] device_driver_attach+0x85/0x90
[ 48.481750] __driver_attach+0xa5/0x180
[ 48.481796] bus_for_each_dev+0xda/0x130
[ 48.481831] bus_add_driver+0x205/0x2e0
[ 48.481882] driver_register+0xca/0x140
[ 48.481927] do_one_initcall+0x6c/0x1af
[ 48.481970] do_init_module+0x106/0x350
[ 48.482010] load_module+0x3d2c/0x3ea0
[ 48.482058] __do_sys_finit_module+0x110/0x180
[ 48.482102] do_syscall_64+0x62/0x1f0
[ 48.482147] entry_SYSCALL_64_after_hwframe+0x44/0xa9
[ 48.482190]
[ 48.482224] Freed by task 37:
[ 48.482273] save_stack+0x19/0x80
[ 48.482318] __kasan_slab_free+0x12e/0x180
[ 48.482363] kmem_cache_free+0x70/0x140
[ 48.482406] __free_iova+0x1d/0x30
[ 48.482445] fq_ring_free+0x15a/0x1a0
[ 48.482490] queue_iova+0x19c/0x1f0
[ 48.482624] cleanup_page_dma.isra.64+0x62/0xb0 [i915]
[ 48.482749] __gen8_ppgtt_cleanup+0x63/0x80 [i915]
[ 48.482873] __gen8_ppgtt_cleanup+0x42/0x80 [i915]
[ 48.482999] __gen8_ppgtt_clear+0x433/0x4b0 [i915]
[ 48.483123] __gen8_ppgtt_clear+0x462/0x4b0 [i915]
[ 48.483250] i915_vma_unbind+0x1e2/0x240 [i915]
[ 48.483378] i915_vma_destroy+0x3a/0x280 [i915]
[ 48.483500] __i915_gem_free_objects+0xf0/0x2d0 [i915]
[ 48.483622] __i915_gem_free_work+0x41/0xa0 [i915]
[ 48.483659] process_one_work+0x495/0x710
[ 48.483704] worker_thread+0x4c7/0x6f0
[ 48.483748] kthread+0x1b2/0x1d0
[ 48.483787] ret_from_fork+0x1f/0x30
[ 48.483831]
[ 48.483868] The buggy address belongs to the object at ffff88870fc19000
[ 48.483868] which belongs to the cache iommu_iova of size 40
[ 48.483920] The buggy address is located 32 bytes inside of
[ 48.483920] 40-byte region [ffff88870fc19000, ffff88870fc19028)
[ 48.483964] The buggy address belongs to the page:
[ 48.484006] page:ffffea001c3f0600 refcount:1 mapcount:0 mapping:ffff8888181a91c0 index:0x0 compound_mapcount: 0
[ 48.484045] flags: 0x8000000000010200(slab|head)
[ 48.484096] raw: 8000000000010200ffffea001c421a08ffffea001c447e88ffff8888181a91c0
[ 48.484141] raw: 0000000000000000000000000012001200000001ffffffff0000000000000000
[ 48.484188] page dumped because: kasan: bad access detected
[ 48.484230]
[ 48.484265] Memory state around the buggy address:
[ 48.484314] ffff88870fc18f00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[ 48.484361] ffff88870fc18f80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[ 48.484406] >ffff88870fc19000: fb fb fb fb fb fc fc fc fc fc fc fc fc fc fc fc
[ 48.484451] ^
[ 48.484494] ffff88870fc19080: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[ 48.484530] ffff88870fc19100: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
Intel VT-d driver was reworked to use common deferred flushing
implementation. Previously there was one global per-cpu flush queue,
afterwards - one per domain.
Before deferring a flush, the queue should be allocated and initialized.
Currently only domains with IOMMU_DOMAIN_DMA type initialize their flush
queue. It's probably worth to init it for static or unmanaged domains
too, but it may be arguable - I'm leaving it to iommu folks.
Prevent queuing an iova flush if the domain doesn't have a queue.
The defensive check seems to be worth to keep even if queue would be
initialized for all kinds of domains. And is easy backportable.
On 4.19.43 stable kernel it has a user-visible effect: previously for
devices in si domain there were crashes, on sata devices:
sq->cached_sq_head and cq->cached_cq_tail are both unsigned int. If
cached_sq_head overflows before cached_cq_tail, then we may miss a
barrier req. As cached_cq_tail always follows cached_sq_head, the NQ
should be enough.
Cc: stable@vger.kernel.org Fixes: de0617e46717 ("io_uring: add support for marking commands as draining") Signed-off-by: Zhengyuan Liu <liuzhengyuan@kylinos.cn> Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The ability to run nested guests under KVM means that a guest can also
act as a hypervisor for it's own nested guest. Currently
ppc_set_pmu_inuse() assumes that either FW_FEATURE_LPAR is set,
indicating a guest environment, and so sets the pmcregs_in_use flag in
the lppaca, or that it isn't set, indicating a hypervisor environment,
and so sets the pmcregs_in_use flag in the paca.
The pmcregs_in_use flag in the lppaca is used to communicate this
information to a hypervisor and so must be set in a guest environment.
The pmcregs_in_use flag in the paca is used by KVM code to determine
whether the host state of the performance monitoring unit (PMU) must
be saved and restored when running a guest.
Thus when a guest also acts as a hypervisor it must set this bit in
both places since it needs to ensure both that the real hypervisor
saves it's PMU registers when it runs (requires pmcregs_in_use flag in
lppaca), and that it saves it's own PMU registers when running a
nested guest (requires pmcregs_in_use flag in paca).
Modify ppc_set_pmu_inuse() so that the pmcregs_in_use bit is set in
both the lppaca and the paca when a guest (LPAR) is running with the
capability of running it's own guests (CONFIG_KVM_BOOK3S_HV_POSSIBLE).
On systems like P9 powernv where we have no TM (or P8 booted with
ppc_tm=off), userspace can construct a signal context which still has
the MSR TS bits set. The kernel tries to restore this context which
results in the following crash:
The problem is the signal code assumes TM is enabled when
CONFIG_PPC_TRANSACTIONAL_MEM is enabled. This may not be the case as
with P9 powernv or if `ppc_tm=off` is used on P8.
This means any local user can crash the system.
Fix the problem by returning a bad stack frame to the user if they try
to set the MSR TS bits with sigreturn() on systems where TM is not
supported.
Found with sigfuz kernel selftest on P9.
This fixes CVE-2019-13648.
Fixes: 2b0a576d15e0 ("powerpc: Add new transactional memory state to the signal context") Cc: stable@vger.kernel.org # v3.9 Reported-by: Praveen Pandey <Praveen.Pandey@in.ibm.com> Signed-off-by: Michael Neuling <mikey@neuling.org> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20190719050502.405-1-mikey@neuling.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The virtual real mode addressing (VRMA) mechanism is used when a
partition is using HPT (Hash Page Table) translation and performs real
mode accesses (MSR[IR|DR] = 0) in non-hypervisor mode. In this mode
effective address bits 0:23 are treated as zero (i.e. the access is
aliased to 0) and the access is performed using an implicit 1TB SLB
entry.
The size of the RMA (Real Memory Area) is communicated to the guest as
the size of the first memory region in the device tree. And because of
the mechanism described above can be expected to not exceed 1TB. In
the event that the host erroneously represents the RMA as being larger
than 1TB, guest accesses in real mode to memory addresses above 1TB
will be aliased down to below 1TB. This means that a memory access
performed in real mode may differ to one performed in virtual mode for
the same memory address, which would likely have unintended
consequences.
To avoid this outcome have the guest explicitly limit the size of the
RMA to the current maximum, which is 1TB. This means that even if the
first memory block is larger than 1TB, only the first 1TB should be
accessed in real mode.
xive_find_target_in_mask() has the following for(;;) loop which has a
bug when @first == cpumask_first(@mask) and condition 1 fails to hold
for every CPU in @mask. In this case we loop forever in the for-loop.
first = cpu;
for (;;) {
if (cpu_online(cpu) && xive_try_pick_target(cpu)) // condition 1
return cpu;
cpu = cpumask_next(cpu, mask);
if (cpu == first) // condition 2
break;
if (cpu >= nr_cpu_ids) // condition 3
cpu = cpumask_first(mask);
}
This is because, when @first == cpumask_first(@mask), we never hit the
condition 2 (cpu == first) since prior to this check, we would have
executed "cpu = cpumask_next(cpu, mask)" which will set the value of
@cpu to a value greater than @first or to nr_cpus_ids. When this is
coupled with the fact that condition 1 is not met, we will never exit
this loop.
This was discovered by the hard-lockup detector while running LTP test
concurrently with SMT switch tests.
To fix this, move the check for condition 2 after the check for
condition 3, so that we are able to break out of the loop soon after
iterating through all the CPUs in the @mask in the problem case. Use
do..while() to achieve this.
Fixes: 243e25112d06 ("powerpc/xive: Native exploitation of the XIVE interrupt controller") Cc: stable@vger.kernel.org # v4.12+ Reported-by: Indira P. Joga <indira.priya@in.ibm.com> Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/1563359724-13931-1-git-send-email-ego@linux.vnet.ibm.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The refactor of powerpc DMA functions in commit 6666cc17d780
("powerpc/dma: remove dma_nommu_mmap_coherent") incorrectly
changes the way DMA mappings are handled on powerpc.
Since this change, all mapped pages are marked as cache-inhibited
through the default implementation of arch_dma_mmap_pgprot.
This differs from the previous behavior of only marking pages
in noncoherent mappings as cache-inhibited and has resulted in
sporadic system crashes in certain hardware configurations and
workloads (see Bugzilla).
This commit restores the previous correct behavior by providing
an implementation of arch_dma_mmap_pgprot that only marks
pages in noncoherent mappings as cache-inhibited. As this behavior
should be universal for all powerpc platforms a new file,
dma-generic.c, was created to store it.
Fixes: 6666cc17d780 ("powerpc/dma: remove dma_nommu_mmap_coherent")
# NOTE: fixes commit 6666cc17d780 released in v5.1.
# Consider a stable tag:
# Cc: stable@vger.kernel.org # v5.1+
# NOTE: fixes commit 6666cc17d780 released in v5.1.
# Consider a stable tag:
# Cc: stable@vger.kernel.org # v5.1+ Cc: stable@vger.kernel.org # v5.1+ Signed-off-by: Shawn Anastasio <shawn@anastas.io> Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20190717235437.12908-1-shawn@anastas.io Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This conexant codec isn't in the supported codec list yet, the hda
generic driver can drive this codec well, but on a Lenovo machine
with mute/mic-mute leds, we need to apply CXT_FIXUP_THINKPAD_ACPI
to make the leds work. After adding this codec to the list, the
driver patch_conexant.c will apply THINKPAD_ACPI to this machine.
It turned out that the recent Intel HD-audio controller chips show a
significant stall during the system PM resume intermittently. It
doesn't happen so often and usually it may read back successfully
after one or more seconds, but in some rare worst cases the driver
went into fallback mode.
After trial-and-error, we found out that the communication stall seems
covered by issuing the sync after each verb write, as already done for
AMD and other chipsets. So this patch enables the write-sync flag for
the recent Intel chips, Skylake and onward, as a workaround.
Also, since Broxton and co have the very same driver flags as Skylake,
refer to the Skylake driver flags instead of defining the same
contents again for simplification.
The recent rewrite of PCM link lock management introduced the refcount
in snd_pcm_group object, managed by the kernel refcount_t API. This
caused unexpected kernel warnings when the kernel is built with
CONFIG_REFCOUNT_FULL=y. As the warning line indicates, the problem is
obviously that we start with refcount=0 and do refcount_inc() for
adding each PCM link, while refcount_t API doesn't like refcount_inc()
performed on zero.
For adapting the proper refcount_t usage, this patch changes the logic
slightly:
- The initial refcount is 1, assuming the single list entry
- The refcount is incremented / decremented at each PCM link addition
and deletion
- ... which allows us concentrating only on the refcount as a release
condition
This patch adds support for the Armadeus ST0700 Adapt. It comes with a
Santek ST0700I5Y-RBSLW 7.0" WVGA (800x480) TFT and an adapter board so
that it can be connected on the TFT header of Armadeus Dev boards.
misc/eeprom/{at24,at25,eeprom_93xx46} drivers all register their
corresponding devices in the nvmem framework in compat mode which requires
nvmem sysfs interface to be present. The latter, however, has been split
out from nvmem under a separate Kconfig in commit ae0c2d725512 ("nvmem:
core: add NVMEM_SYSFS Kconfig"). As a result, probing certain I2C-attached
EEPROMs now fails with
at24: probe of 0-0050 failed with error -38
because of a stub implementation of nvmem_sysfs_setup_compat()
in drivers/nvmem/nvmem.h. Update the nvmem dependency for these drivers
so they could load again:
Currently, a transaction to context manager from its own process
is prevented by checking if its binder_proc struct is the same as
that of the sender. However, this would not catch cases where the
process opens the binder device again and uses the new fd to send
a transaction to the context manager.
In case the target node requests a security context, the
extra_buffers_size is increased with the size of the security context.
But, that size is not available for use by regular scatter-gather
buffers; make sure the ending of that buffer is marked correctly.
When arch_stack_walk_user() is called from atomic contexts, access_ok() can
trigger the following warning if compiled with CONFIG_DEBUG_ATOMIC_SLEEP=y.
Fix it by calling __range_not_ok() directly instead of access_ok() as
copy_from_user_nmi() does. This is fine here because the actual copy is
inside a pagefault disabled region.
Reported-by: Juri Lelli <juri.lelli@gmail.com> Signed-off-by: Eiichi Tsukata <devel@etsukata.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Link: https://lkml.kernel.org/r/20190722083216.16192-2-devel@etsukata.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
X86_HYPER_NATIVE isn't accurate for checking if running on native platform,
e.g. CONFIG_HYPERVISOR_GUEST isn't set or "nopv" is enabled.
Checking the CPU feature bit X86_FEATURE_HYPERVISOR to determine if it's
running on native platform is more accurate.
This still doesn't cover the platforms on which X86_FEATURE_HYPERVISOR is
unsupported, e.g. VMware, but there is nothing which can be done about this
scenario.
Some Lenovo 2-in-1s with a detachable keyboard have a portrait screen but
advertise a landscape resolution and pitch, resulting in a messed up
display if the kernel tries to show anything on the efifb (because of the
wrong pitch).
Fix this by adding a new DMI match table for devices which need to have
their width and height swapped.
At first it was tried to use the existing table for overriding some of the
efifb parameters, but some of the affected devices have variants with
different LCD resolutions which will not work with hardcoded override
values.
Reference: https://bugzilla.redhat.com/show_bug.cgi?id=1730783 Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20190721152418.11644-1-hdegoede@redhat.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Note
If nodatacow or nodatasum are enabled, compression is disabled.
If NODATASUM or NODATACOW set, we should not compress the extent.
Normally NODATACOW is detected properly in run_delalloc_range() so
compression won't happen for NODATACOW.
However for NODATASUM we don't have any check, and it can cause
compressed extent without csum pretty easily, just by:
mkfs.btrfs -f $dev
mount $dev $mnt -o nodatasum
touch $mnt/foobar
mount -o remount,datasum,compress $mnt
xfs_io -f -c "pwrite 0 128K" $mnt/foobar
And in fact, we have a bug report about corrupted compressed extent
without proper data checksum so even RAID1 can't recover the corruption.
(https://bugzilla.kernel.org/show_bug.cgi?id=199707)
Running compression without proper checksum could cause more damage when
corruption happens, as compressed data could make the whole extent
unreadable, so there is no need to allow compression for
NODATACSUM.
The fix will refactor the inode compression check into two parts:
- inode_can_compress()
As the hard requirement, checked at btrfs_run_delalloc_range(), so no
compression will happen for NODATASUM inode at all.
- inode_need_compress()
As the soft requirement, checked at btrfs_run_delalloc_range() and
compress_file_range().
Reported-by: James Harvey <jamespharvey20@gmail.com> CC: stable@vger.kernel.org # 4.4+ Signed-off-by: Qu Wenruo <wqu@suse.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The V4L2_PIX_FMT_BGRA444 define clashed with the pre-existing V4L2_PIX_FMT_SGRBG12
which strangely enough used the same fourcc, even though that fourcc made no sense
for a Bayer format. In any case, you can't have duplicates, so change the fourcc of
V4L2_PIX_FMT_BGRA444.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Cc: <stable@vger.kernel.org> # for v5.2 and up Fixes: 6c84f9b1d2900 ("media: v4l: Add definitions for missing 16-bit RGB4444 formats") Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The XIVE device structure is now allocated in kvmppc_xive_get_device()
and kfree'd in kvmppc_core_destroy_vm(). In case of an OPAL error when
allocating the XIVE VPs, the kfree() call in kvmppc_xive_*create()
will result in a double free and corrupt the host memory.
Fixes: 5422e95103cf ("KVM: PPC: Book3S HV: XIVE: Replace the 'destroy' method by a 'release' method") Cc: stable@vger.kernel.org # v5.2+ Signed-off-by: Cédric Le Goater <clg@kaod.org> Tested-by: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/6ea6998b-a890-2511-01d1-747d7621eb19@kaod.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The Performance Stop Status and Control Register (PSSCR) is used to
control the power saving facilities of the processor. This register
has various fields, some of which can be modified only in hypervisor
state, and others which can be modified in both hypervisor and
privileged non-hypervisor state. The bits which can be modified in
privileged non-hypervisor state are referred to as guest visible.
Currently the L0 hypervisor saves and restores both it's own host
value as well as the guest value of the PSSCR when context switching
between the hypervisor and guest. However a nested hypervisor running
it's own nested guests (as indicated by kvmhv_on_pseries()) doesn't
context switch the PSSCR register. That means if a nested (L2) guest
modifies the PSSCR then the L1 guest hypervisor will run with that
modified value, and if the L1 guest hypervisor modifies the PSSCR and
then goes to run the nested (L2) guest again then the L2 PSSCR value
will be lost.
Fix this by having the (L1) nested hypervisor save and restore both
its host and the guest PSSCR value when entering and exiting a
nested (L2) guest. Note that only the guest visible parts of the PSSCR
are context switched since this is all the L1 nested hypervisor can
access, this is fine however as these are the only fields the L0
hypervisor provides guest control of anyway and so all other fields
are ignored.
This could also have been implemented by adding the PSSCR register to
the hv_regs passed to the L0 hypervisor as input to the H_ENTER_NESTED
hcall, however this would have meant updating the structure layout and
thus required modifications to both the L0 and L1 kernels. Whereas the
approach used doesn't require L0 kernel modifications while achieving
the same result.
The performance monitoring unit (PMU) registers are saved on guest
exit when the guest has set the pmcregs_in_use flag in its lppaca, if
it exists, or unconditionally if it doesn't. If a nested guest is
being run then the hypervisor doesn't, and in most cases can't, know
if the PMU registers are in use since it doesn't know the location of
the lppaca for the nested guest, although it may have one for its
immediate guest. This results in the values of these registers being
lost across nested guest entry and exit in the case where the nested
guest was making use of the performance monitoring facility while it's
nested guest hypervisor wasn't.
Further more the hypervisor could interrupt a guest hypervisor between
when it has loaded up the PMU registers and it calling H_ENTER_NESTED
or between returning from the nested guest to the guest hypervisor and
the guest hypervisor reading the PMU registers, in
kvmhv_p9_guest_entry(). This means that it isn't sufficient to just
save the PMU registers when entering or exiting a nested guest, but
that it is necessary to always save the PMU registers whenever a guest
is capable of running nested guests to ensure the register values
aren't lost in the context switch.
Ensure the PMU register values are preserved by always saving their
value into the vcpu struct when a guest is capable of running nested
guests.
This should have minimal performance impact however any impact can be
avoided by booting a guest with "-machine pseries,cap-nested-hv=false"
on the qemu commandline.
The idea before commit 240c35a37 (which has just been reverted)
was that we have the following FPU states:
userspace (QEMU) guest
---------------------------------------------------------------------------
processor vcpu->arch.guest_fpu
>>> KVM_RUN: kvm_load_guest_fpu
vcpu->arch.user_fpu processor
>>> preempt out
vcpu->arch.user_fpu current->thread.fpu
>>> preempt in
vcpu->arch.user_fpu processor
>>> back to userspace
>>> kvm_put_guest_fpu
processor vcpu->arch.guest_fpu
---------------------------------------------------------------------------
With the new lazy model we want to get the state back to the processor
when schedule in from current->thread.fpu.
Reported-by: Thomas Lambertz <mail@thomaslambertz.de> Reported-by: anthony <antdev66@gmail.com> Tested-by: anthony <antdev66@gmail.com> Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Radim Krčmář <rkrcmar@redhat.com> Cc: Thomas Lambertz <mail@thomaslambertz.de> Cc: anthony <antdev66@gmail.com> Cc: stable@vger.kernel.org Fixes: 5f409e20b (x86/fpu: Defer FPU state load until return to userspace) Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
[Add a comment in front of the warning. - Paolo] Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is a partial revert of 73d31def1aab "usb: usb251xb: Create a ports
field collector method", which broke a existing devicetree
(arch/arm64/boot/dts/freescale/imx8mq.dtsi).
There is no reason why the swap-dx-lanes property should not apply to
the upstream port. The reason given in the breaking commit was that it's
inconsitent with respect to other port properties, but in fact it is not.
All other properties which only apply to the downstream ports explicitly
reject port 0, so there is pretty strong precedence that the driver
referred to the upstream port as port 0. So there is no inconsistency in
this property at all, other than the swapping being also applicable to
the upstream port.
This reverts commit 3342ce35a1, as there is no need for this separate
property and it breaks compatibility with existing devicetree files
(arch/arm64/boot/dts/freescale/imx8mq.dtsi).
The AMD PLL USB quirk is incorrectly enabled on newer Ryzen
chipsets. The logic in usb_amd_find_chipset_info currently checks
for unaffected chipsets rather than affected ones. This broke
once a new chipset was added in e788787ef. It makes more sense
to reverse the logic so it won't need to be updated as new
chipsets are added. Note that the core of the workaround in
usb_amd_quirk_pll does correctly check the chipset.
Signed-off-by: Ryan Kennedy <ryan5544@gmail.com> Fixes: e788787ef4f9 ("usb:xhci:Add quirk for Certain failing HP keyboard on reset after resume") Cc: stable <stable@vger.kernel.org> Acked-by: Alan Stern <stern@rowland.harvard.edu> Link: https://lore.kernel.org/r/20190704153529.9429-2-ryan5544@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
syzboot reported that
https://syzkaller.appspot.com/bug?extid=fd2bd7df88c606eea4ef
There is not consitency parameter in cluste_id_get/put calling.
In case of getting the id with result is failure, the wusbhc->cluster_id
will not be updated and this can not be used for wusb_cluster_id_put().
This patch fixes an issue that the following error happens on
swiotlb environment:
xhci-hcd ee000000.usb: swiotlb buffer is full (sz: 524288 bytes), total 32768 (slots), used 1338 (slots)
On the kernel v5.1, block settings of a usb-storage with SuperSpeed
were the following so that the block layer will allocate buffers
up to 64 KiB, and then the issue didn't happen.
max_segment_size = 65536
max_hw_sectors_kb = 1024
After the commit 09324d32d2a0 ("block: force an unlimited segment
size on queues with a virt boundary") is applied, the block settings
are the following. So, the block layer will allocate buffers up to
1024 KiB, and then the issue happens:
To fix the issue, the usb-storage driver checks the maximum size of
a mapping for the device and then adjusts the max_hw_sectors_kb
if required. After this patch is applied, the block settings will
be the following, and then the issue doesn't happen.
Fixes: 09324d32d2a0 ("block: force an unlimited segment size on queues with a virt boundary") Cc: stable <stable@vger.kernel.org> Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com> Acked-by: Alan Stern <stern@rowland.harvard.edu> Reviewed-by: Christoph Hellwig <hch@lst.de> Link: https://lore.kernel.org/r/1563793105-20597-1-git-send-email-yoshihiro.shimoda.uh@renesas.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A second regression was found in the immediate data transfer (IDT)
support which was added to 5.2 kernel
IDT is used to transfer small amounts of data (up to 8 bytes) in the
field normally used for data dma address, thus avoiding dma mapping.
If the data was not already dma mapped, then IDT support assumed data was
in urb->transfer_buffer, and did not take into accound that even
small amounts of data (8 bytes) can be in a scatterlist instead.
This caused a NULL pointer dereference when sg_dma_len() was used
with non-dma mapped data.
Solve this by not using IDT if scatter gather buffer list is used.
When swapin is performed, after getting the swap entry information from
the page table, system will swap in the swap entry, without any lock held
to prevent the swap device from being swapoff. This may cause the race
like below,
Because swapoff is usually done when system shutdown only, the race may
not hit many people in practice. But it is still a race need to be fixed.
To fix the race, get_swap_device() is added to check whether the specified
swap entry is valid in its swap device. If so, it will keep the swap
entry valid via preventing the swap device from being swapoff, until
put_swap_device() is called.
Because swapoff() is very rare code path, to make the normal path runs as
fast as possible, rcu_read_lock/unlock() and synchronize_rcu() instead of
reference count is used to implement get/put_swap_device(). >From
get_swap_device() to put_swap_device(), RCU reader side is locked, so
synchronize_rcu() in swapoff() will wait until put_swap_device() is
called.
In addition to swap_map, cluster_info, etc. data structure in the struct
swap_info_struct, the swap cache radix tree will be freed after swapoff,
so this patch fixes the race between swap cache looking up and swapoff
too.
Races between some other swap cache usages and swapoff are fixed too via
calling synchronize_rcu() between clearing PageSwapCache() and freeing
swap cache data structure.
Another possible method to fix this is to use preempt_off() +
stop_machine() to prevent the swap device from being swapoff when its data
structure is being accessed. The overhead in hot-path of both methods is
similar. The advantages of RCU based method are,
1. stop_machine() may disturb the normal execution code path on other
CPUs.
2. File cache uses RCU to protect its radix tree. If the similar
mechanism is used for swap cache too, it is easier to share code
between them.
3. RCU is used to protect swap cache in total_swapcache_pages() and
exit_swap_address_space() already. The two mechanisms can be
merged to simplify the logic.
Link: http://lkml.kernel.org/r/20190522015423.14418-1-ying.huang@intel.com Fixes: 235b62176712 ("mm/swap: add cluster lock") Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Reviewed-by: Andrea Parri <andrea.parri@amarulasolutions.com> Not-nacked-by: Hugh Dickins <hughd@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Daniel Jordan <daniel.m.jordan@oracle.com> Cc: Michal Hocko <mhocko@suse.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tim Chen <tim.c.chen@linux.intel.com> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Jérôme Glisse <jglisse@redhat.com> Cc: Yang Shi <yang.shi@linux.alibaba.com> Cc: David Rientjes <rientjes@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Dave Jiang <dave.jiang@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>
This function is used by ptrace and proc files like /proc/pid/cmdline and
/proc/pid/environ.
Access_remote_vm never returns error codes, all errors are ignored and
only size of successfully read data is returned. So, if current task was
killed we'll simply return 0 (bytes read).
Mmap_sem could be locked for a long time or forever if something goes
wrong. Using a killable lock permits cleanup of stuck tasks and
simplifies investigation.
Link: http://lkml.kernel.org/r/156007494202.3335.16782303099589302087.stgit@buzz Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru> Reviewed-by: Michal Koutný <mkoutny@suse.com> Acked-by: Oleg Nesterov <oleg@redhat.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Alexey Dobriyan <adobriyan@gmail.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Cyrill Gorcunov <gorcunov@gmail.com> Cc: Kirill Tkhai <ktkhai@virtuozzo.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Roman Gushchin <guro@fb.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>
The stats variable nr_unused_locks is incremented every time a new lock
class is register and decremented when the lock is first used in
__lock_acquire(). And after all, it is shown and checked in lockdep_stats.
However, under configurations that either CONFIG_TRACE_IRQFLAGS or
CONFIG_PROVE_LOCKING is not defined:
The commit:
091806515124b20 ("locking/lockdep: Consolidate lock usage bit initialization")
missed marking the LOCK_USED flag at IRQ usage initialization because
as mark_usage() is not called. And the commit:
The cudbg_collect_mem_region() and cudbg_read_fw_mem() both use several
hundred kilobytes of kernel stack space. One gets inlined into the other,
which causes the stack usage to be combined beyond the warning limit
when building with clang:
drivers/net/ethernet/chelsio/cxgb4/cudbg_lib.c:1057:12: error: stack frame size of 1244 bytes in function 'cudbg_collect_mem_region' [-Werror,-Wframe-larger-than=]
Restructuring cudbg_collect_mem_region() lets clang do the same
optimization that gcc does and reuse the stack slots as it can
see that the large variables are never used together.
A better fix might be to avoid using cudbg_meminfo on the stack
altogether, but that requires a larger rewrite.
Fixes: a1c69520f785 ("cxgb4: collect MC memory dump") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
Make mmu_notifier_register() safer by issuing a memory barrier before
registering a new notifier. This fixes a theoretical bug on weakly
ordered CPUs. For example, take this simplified use of notifiers by a
driver:
Once mmu_notifier_register() releases the mm locks, another thread can
invalidate a range:
mmu_notifier_invalidate_range()
...
hlist_for_each_entry_rcu(mn, &mm->mmu_notifiers, hlist) {
if (mn->ops->invalidate_range)
The read side relies on the data dependency between mn and ops to ensure
that the pointer is properly initialized. But the write side doesn't have
any dependency between (1) and (2), so they could be reordered and the
readers could dereference an invalid mn->ops. mmu_notifier_register()
does take all the mm locks before adding to the hlist, but those have
acquire semantics which isn't sufficient.
By calling hlist_add_head_rcu() instead of hlist_add_head() we update the
hlist using a store-release, ensuring that readers see prior
initialization of my_struct. This situation is better illustated by
litmus test MP+onceassign+derefonce.
Commit d46eb14b735b ("fs: fsnotify: account fsnotify metadata to
kmemcg") added remote memcg charging for fanotify and inotify event
objects. The aim was to charge the memory to the listener who is
interested in the events but without triggering the OOM killer.
Otherwise there would be security concerns for the listener.
At the time, oom-kill trigger was not in the charging path. A parallel
work added the oom-kill back to charging path i.e. commit 29ef680ae7c2
("memcg, oom: move out_of_memory back to the charge path"). So to not
trigger oom-killer in the remote memcg, explicitly add
__GFP_RETRY_MAYFAIL to the fanotigy and inotify event allocations.
Link: http://lkml.kernel.org/r/20190514212259.156585-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Acked-by: Jan Kara <jack@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Cc: Amir Goldstein <amir73il@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>
If we end up without a PGD or PUD entry backing the gate area, don't BUG
-- just fail gracefully.
It's not entirely implausible that this could happen some day on x86. It
doesn't right now even with an execute-only emulated vsyscall page because
the fixmap shares the PUD, but the core mm code shouldn't rely on that
particular detail to avoid OOPSing.
Via commit 4b3ef9daa4fc ("mm/swap: split swap cache into 64MB trunks"),
after swapoff, the address_space associated with the swap device will be
freed. So swap_address_space() users which touch the address_space need
some kind of mechanism to prevent the address_space from being freed
during accessing.
When mincore processes an unmapped range for swapped shmem pages, it
doesn't hold the lock to prevent swap device from being swapped off. So
the following race is possible:
The address space may be accessed after being freed.
To fix the race, get_swap_device()/put_swap_device() is used to enclose
find_get_page() to check whether the swap entry is valid and prevent the
swap device from being swapoff during accessing.
Link: http://lkml.kernel.org/r/20190611020510.28251-1-ying.huang@intel.com Fixes: 4b3ef9daa4fc ("mm/swap: split swap cache into 64MB trunks") Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Hugh Dickins <hughd@google.com> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tim Chen <tim.c.chen@linux.intel.com> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Jérôme Glisse <jglisse@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Yang Shi <yang.shi@linux.alibaba.com> Cc: David Rientjes <rientjes@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Dave Jiang <dave.jiang@intel.com> Cc: Daniel Jordan <daniel.m.jordan@oracle.com> Cc: Andrea Parri <andrea.parri@amarulasolutions.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>
Fix the callback 9p passes to read_cache_page to actually have the
proper type expected. Casting around function pointers can easily
hide typing bugs, and defeats control flow protection.
Link: http://lkml.kernel.org/r/20190520055731.24538-5-hch@lst.de Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Kees Cook <keescook@chromium.org> Cc: Sami Tolvanen <samitolvanen@google.com> Cc: Nick Desaulniers <ndesaulniers@google.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>
in_softirq() is a wrong predicate to check if we are in a softirq
context. It also returns true if we have BH disabled, so objects are
falsely stamped with "softirq" comm. The correct predicate is
in_serving_softirq().
If user does cat from /sys/kernel/debug/kmemleak previously they would
see this, which is clearly wrong, this is system call context (see the
comm):
Fixes: 958f2a0f8121 ("nvme-tcp: set the STABLE_WRITES flag when data digests are enabled") Cc: Christoph Hellwig <hch@lst.de> Cc: Keith Busch <kbusch@kernel.org> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Minwoo Im <minwoo.im.dev@gmail.com> Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
In bio_integrity_prep(), a kernel buffer is allocated through kmalloc() to
hold integrity metadata. Later on, the buffer will be attached to the bio
structure through bio_integrity_add_page(), which returns the number of
bytes of integrity metadata attached. Due to unexpected situations,
bio_integrity_add_page() may return 0. As a result, bio_integrity_prep()
needs to be terminated with 'false' returned to indicate this error.
However, the allocated kernel buffer is not freed on this execution path,
leading to a memory leak.
To fix this issue, free the allocated buffer before returning from
bio_integrity_prep().
Reviewed-by: Ming Lei <ming.lei@redhat.com> Acked-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Wenwen Wang <wenwen@cs.uga.edu> Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
In commit 4a7b06c157a2 ("powerpc/eeh: Handle hugepages in ioremap
space") support for using hugepages in the vmalloc and ioremap areas was
enabled for radix. Unfortunately this broke EEH MMIO error checking.
Detection works by inserting a hook which checks the results of the
ioreadXX() set of functions. When a read returns a 0xFFs response we
need to check for an error which we do by mapping the (virtual) MMIO
address back to a physical address, then mapping physical address to a
PCI device via an interval tree.
When translating virt -> phys we currently assume the ioremap space is
only populated by PAGE_SIZE mappings. If a hugepage mapping is found we
emit a WARN_ON(), but otherwise handles the check as though a normal
page was found. In pathalogical cases such as copying a buffer
containing a lot of 0xFFs from BAR memory this can result in the system
not booting because it's too busy printing WARN_ON()s.
There's no real reason to assume huge pages can't be present and we're
prefectly capable of handling them, so do that.
Fixes: 4a7b06c157a2 ("powerpc/eeh: Handle hugepages in ioremap space") Reported-by: Sachin Sant <sachinp@linux.vnet.ibm.com> Signed-off-by: Oliver O'Halloran <oohall@gmail.com> Tested-by: Sachin Sant <sachinp@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20190710150517.27114-1-oohall@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org>
If the DLM lowcomms stack is shut down before any DLM
traffic can be generated, flush_workqueue() and
destroy_workqueue() can be called on empty send and/or recv
workqueues.
Insert guard conditionals to only call flush_workqueue()
and destroy_workqueue() on workqueues that are not NULL.
Signed-off-by: David Windsor <dwindsor@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Previously, if mbox_request_channel_byname was used with a name
which did not exist in the "mbox-names" property of a mailbox
client, the mailbox corresponding to the last entry in the
"mbox-names" list would be incorrectly selected.
With this patch, -EINVAL is returned if the named mailbox is
not found.
In umount, we give an constand time to handle pending discard, previously,
in __issue_discard_cmd() we missed to check timeout condition in loop,
result in delaying long time, fix it.
Prior to
commit d021fabf525ff ("rds: rdma: add consumer reject")
function "rds_rdma_cm_event_handler_cmn" would always honor a rejected
connection attempt by issuing a "rds_conn_drop".
The commit mentioned above added a "break", eliminating
the "fallthrough" case and made the "rds_conn_drop" rather conditional:
Now it only happens if a "consumer defined" reject (i.e. "rdma_reject")
carries an integer-value of "1" inside "private_data":
if (!conn)
break;
err = (int *)rdma_consumer_reject_data(cm_id, event, &len);
if (!err || (err && ((*err) == RDS_RDMA_REJ_INCOMPAT))) {
pr_warn("RDS/RDMA: conn <%pI6c, %pI6c> rejected, dropping connection\n",
&conn->c_laddr, &conn->c_faddr);
conn->c_proposed_version = RDS_PROTOCOL_COMPAT_VERSION;
rds_conn_drop(conn);
}
rdsdebug("Connection rejected: %s\n",
rdma_reject_msg(cm_id, event->status));
break;
/* FALLTHROUGH */
A number of issues are worth mentioning here:
#1) Previous versions of the RDS code simply rejected a connection
by calling "rdma_reject(cm_id, NULL, 0);"
So the value of the payload in "private_data" will not be "1",
but "0".
#2) Now the code has become dependent on host byte order and sizing.
If one peer is big-endian, the other is little-endian,
or there's a difference in sizeof(int) (e.g. ILP64 vs LP64),
the *err check does not work as intended.
#3) There is no check for "len" to see if the data behind *err is even valid.
Luckily, it appears that the "rdma_reject(cm_id, NULL, 0)" will always
carry 148 bytes of zeroized payload.
But that should probably not be relied upon here.
#4) With the added "break;",
we might as well drop the misleading "/* FALLTHROUGH */" comment.
This commit does _not_ address issue #2, as the sender would have to
agree on a byte order as well.
Here is the sequence of messages in this observed error-scenario:
Host-A is pre-QoS changes (excluding the commit mentioned above)
Host-B is post-QoS changes (including the commit mentioned above)
#1 Host-B
issues a connection request via function "rds_conn_path_transition"
connection state transitions to "RDS_CONN_CONNECTING"
#2 Host-A
rejects the incompatible connection request (from #1)
It does so by calling "rdma_reject(cm_id, NULL, 0);"
#3 Host-B
receives an "RDMA_CM_EVENT_REJECTED" event (from #2)
But since the code is changed in the way described above,
it won't drop the connection here, simply because "*err == 0".
#4 Host-A
issues a connection request
#5 Host-B
receives an "RDMA_CM_EVENT_CONNECT_REQUEST" event
and ends up calling "rds_ib_cm_handle_connect".
But since the state is already in "RDS_CONN_CONNECTING"
(as of #1) it will end up issuing a "rdma_reject" without
dropping the connection:
if (rds_conn_state(conn) == RDS_CONN_CONNECTING) {
/* Wait and see - our connect may still be succeeding */
rds_ib_stats_inc(s_ib_connect_raced);
}
goto out;
#6 Host-A
receives an "RDMA_CM_EVENT_REJECTED" event (from #5),
drops the connection and tries again (goto #4) until it gives up.
We discovered a problem in newer kernels where a disconnect of a NBD
device while the flush request was pending would result in a hang. This
is because the blk mq timeout handler does
if (!refcount_inc_not_zero(&rq->ref))
return true;
to determine if it's ok to run the timeout handler for the request.
Flush_rq's don't have a ref count set, so we'd skip running the timeout
handler for this request and it would just sit there in limbo forever.
Fix this by always setting the refcount of any request going through
blk_init_rq() to 1. I tested this with a nbd-server that dropped flush
requests to verify that it hung, and then tested with this patch to
verify I got the timeout as expected and the error handling kicked in.
Thanks,
The next commit will make the way of passing CONFIG options more robust.
Unfortunately, it would uncover another hidden issue; without this
commit, skiroot_defconfig would be broken like this:
| WRAP arch/powerpc/boot/zImage.pseries
| arch/powerpc/boot/wrapper.a(decompress.o): In function `bcj_powerpc.isra.10':
| decompress.c:(.text+0x720): undefined reference to `get_unaligned_be32'
| decompress.c:(.text+0x7a8): undefined reference to `put_unaligned_be32'
| make[1]: *** [arch/powerpc/boot/Makefile;383: arch/powerpc/boot/zImage.pseries] Error 1
| make: *** [arch/powerpc/Makefile;295: zImage] Error 2
skiroot_defconfig is the only defconfig that enables CONFIG_KERNEL_XZ
for ppc, which has never been correctly built before.
I figured out the root cause in lib/decompress_unxz.c:
When CONFIG_PPC_IRQ_SOFT_MASK_DEBUG is enabled (uncommon), we have a
series of WARN_ON's in arch_local_irq_restore().
These are "should never happen" conditions, but if they do happen they
can flood the console and render the system unusable. So switch them
to WARN_ON_ONCE().
Fixes: e2b36d591720 ("powerpc/64: Don't trace code that runs with the soft irq mask unreconciled") Fixes: 9b81c0211c24 ("powerpc/64s: make PACA_IRQ_HARD_DIS track MSR[EE] closely") Fixes: 7c0482e3d055 ("powerpc/irq: Fix another case of lazy IRQ state getting out of sync") Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20190708061046.7075-1-mpe@ellerman.id.au Signed-off-by: Sasha Levin <sashal@kernel.org>