Darrick J. Wong [Thu, 26 Jan 2017 04:24:57 +0000 (20:24 -0800)]
xfs: clear _XBF_PAGES from buffers when readahead page
If we try to allocate memory pages to back an xfs_buf that we're trying
to read, it's possible that we'll be so short on memory that the page
allocation fails. For a blocking read we'll just wait, but for
readahead we simply dump all the pages we've collected so far.
Unfortunately, after dumping the pages we neglect to clear the
_XBF_PAGES state, which means that the subsequent call to xfs_buf_free
thinks that b_pages still points to pages we own. It then double-frees
the b_pages pages.
This results in screaming about negative page refcounts from the memory
manager, which xfs oughtn't be triggering. To reproduce this case,
mount a filesystem where the size of the inodes far outweighs the
availalble memory (a ~500M inode filesystem on a VM with 300MB memory
did the trick here) and run bulkstat in parallel with other memory
eating processes to put a huge load on the system. The "check summary"
phase of xfs_scrub also works for this purpose.
Jakub Kicinski [Wed, 25 Jan 2017 22:56:36 +0000 (14:56 -0800)]
virtio_net: reject XDP programs using header adjustment
commit 17bedab27231 ("bpf: xdp: Allow head adjustment in XDP prog")
added a new XDP helper to prepend and remove data from a frame.
Make virtio_net reject programs making use of this helper until
proper support is added.
John Fastabend [Thu, 26 Jan 2017 02:22:48 +0000 (18:22 -0800)]
virtio_net: use dev_kfree_skb for small buffer XDP receive
In the small buffer case during driver unload we currently use
put_page instead of dev_kfree_skb. Resolve this by adding a check
for virtnet mode when checking XDP queue type. Also name the
function so that the code reads correctly to match the additional
check.
David S. Miller [Thu, 26 Jan 2017 03:47:31 +0000 (22:47 -0500)]
Merge branch 'r8152-napi-fixes'
Hayes Wang says:
====================
r8152: fix scheduling napi
v3:
simply the argument for patch #3. Replace &tp->napi with napi.
v2:
Add smp_mb__after_atomic() for patch #1.
v1:
Scheduling the napi during the following periods would let it be ignored.
And the events wouldn't be handled until next napi_schedule() is called.
1. after napi_disable and before napi_enable().
2. after all actions of napi function is completed and before calling
napi_complete().
If no next napi_schedule() is called, tx or rx would stop working.
In order to avoid these situations, the followings solutions are applied.
1. prevent start_xmit() from calling napi_schedule() during runtime suspend
or after napi_disable().
2. re-schedule the napi for tx if it is necessary.
3. check if any rx is finished or not after napi_enable().
====================
hayeswang [Thu, 26 Jan 2017 01:38:33 +0000 (09:38 +0800)]
r8152: re-schedule napi for tx
Re-schedule napi after napi_complete() for tx, if it is necessay.
In r8152_poll(), if the tx is completed after tx_bottom() and before
napi_complete(), the scheduling of napi would be lost. Then, no
one handles the next tx until the next napi_schedule() is called.
Gao Pan [Tue, 17 Jan 2017 10:20:55 +0000 (18:20 +0800)]
i2c: imx-lpi2c: add VLLS mode support
When system enters VLLS mode, module power is turned off. As a result,
all registers are reset to HW default value. After exiting VLLS mode,
registers are still in default mode. As a result, the pinctrl settings
are incorrect, which will affect the module function.
The patch recovers the pinctrl setting when exit VLLS mode.
Mike Looijmans [Mon, 16 Jan 2017 14:49:38 +0000 (15:49 +0100)]
i2c: i2c-cadence: Initialize configuration before probing devices
The cadence I2C driver calls cdns_i2c_writereg(..) to setup a workaround
in the controller, but did so after calling i2c_add_adapter() which starts
probing devices on the bus. Change the order so that the configuration is
completely finished before using the adapter.
Florian Fainelli [Wed, 25 Jan 2017 17:10:41 +0000 (09:10 -0800)]
net: dsa: Bring back device detaching in dsa_slave_suspend()
Commit 448b4482c671 ("net: dsa: Add lockdep class to tx queues to avoid
lockdep splat") removed the netif_device_detach() call done in
dsa_slave_suspend() which is necessary, and paired with a corresponding
netif_device_attach(), bring it back.
Fixes: 448b4482c671 ("net: dsa: Add lockdep class to tx queues to avoid lockdep splat") Signed-off-by: Florian Fainelli <[email protected]> Reviewed-by: Andrew Lunn <[email protected]> Signed-off-by: David S. Miller <[email protected]>
David S. Miller [Wed, 25 Jan 2017 19:40:25 +0000 (14:40 -0500)]
Merge branch 'phy-truncated-led-names'
Geert Uytterhoeven says:
====================
net: phy: leds: Fix truncated LED trigger names and crashes
I started seeing crashes during s2ram and poweroff on all my ARM boards,
like:
Unable to handle kernel NULL pointer dereference at virtual address 00000000
...
[<c04116d4>] (__list_del_entry_valid) from [<c05e8948>] (led_trigger_unregister+0x34/0xcc)
[<c05e8948>] (led_trigger_unregister) from [<c05336c4>] (phy_led_triggers_unregister+0x28/0x34)
[<c05336c4>] (phy_led_triggers_unregister) from [<c0531d44>] (phy_detach+0x30/0x74)
[<c0531d44>] (phy_detach) from [<c0538bdc>] (sh_eth_close+0x64/0x9c)
[<c0538bdc>] (sh_eth_close) from [<c04d4ce0>] (dpm_run_callback+0x48/0xc8)
or:
list_del corruption. prev->next should be dede6540, but was 2e323931
------------[ cut here ]------------
kernel BUG at lib/list_debug.c:52!
...
[<c02f6d70>] (__list_del_entry_valid) from [<c0425168>] (led_trigger_unregister+0x34/0xcc)
[<c0425168>] (led_trigger_unregister) from [<c03a05a0>] (phy_led_triggers_unregister+0x28/0x34)
[<c03a05a0>] (phy_led_triggers_unregister) from [<c039ec04>] (phy_detach+0x30/0x74)
[<c039ec04>] (phy_detach) from [<c03a4fc0>] (sh_eth_close+0x6c/0xa4)
[<c03a4fc0>] (sh_eth_close) from [<c0483234>] (__dev_close_many+0xac/0xd0)
As the only clue was a kernel message like
sh-eth ee700000.ethernet eth0: No phy led trigger registered for speed(100)
I had to bisected this, leading to commit 4567d686f5c6d955 ("phy:
increase size of MII_BUS_ID_SIZE and bus_id"). Reverting that commit
fixed the issue.
More investigation revealed the crashes are due to the combination of
two things:
- Truncated LED trigger names, leading to duplicate names, and
registration failures,
- Bad error handling in case of registration failures.
Both are fixed by this patch series.
Changes compared to v1:
- Add Reviewed-by,
- New patch "net: phy: leds: Break dependency of phy.h on
phy_led_triggers.h",
- Drop moving the include of <linux/phy_led_triggers.h>, as
<linux/phy.h> no longer includes it,
- #include <linux/phy.h> from <linux/phy_led_triggers.h>.
====================
Commit 4567d686f5c6d955 ("phy: increase size of MII_BUS_ID_SIZE and
bus_id") increased the size of MII bus IDs, but forgot to update the
private definition in <linux/phy_led_triggers.h>.
This may cause:
1. Truncation of LED trigger names,
2. Duplicate LED trigger names,
3. Failures registering LED triggers,
4. Crashes due to bad error handling in the LED trigger failure path.
To fix this, and prevent the definitions going out of sync again in the
future, let the PHY LED trigger code use the existing MII_BUS_ID_SIZE
definition.
Example:
- Before I had triggers "ee700000.etherne:01:100Mbps" and
"ee700000.etherne:01:10Mbps",
- After the increase of MII_BUS_ID_SIZE, both became
"ee700000.ethernet-ffffffff:01:" => FAIL,
- Now, the triggers are "ee700000.ethernet-ffffffff:01:100Mbps" and
"ee700000.ethernet-ffffffff:01:10Mbps", which are unique again.
Fixes: 4567d686f5c6d955 ("phy: increase size of MII_BUS_ID_SIZE and bus_id") Fixes: 2e0bc452f4721520 ("net: phy: leds: add support for led triggers on phy link state change") Signed-off-by: Geert Uytterhoeven <[email protected]> Reviewed-by: Andrew Lunn <[email protected]> Signed-off-by: David S. Miller <[email protected]>
net: phy: leds: Break dependency of phy.h on phy_led_triggers.h
<linux/phy.h> includes <linux/phy_led_triggers.h>, which is not really
needed. Drop the include from <linux/phy.h>, and add it to all users
that didn't include it explicitly.
net: phy: leds: Clear phy_num_led_triggers on failure to avoid crash
phy_attach_direct() ignores errors returned by
phy_led_triggers_register(). I think that's OK, as LED triggers can be
considered a non-critical feature.
However, this causes problems later:
- phy_led_trigger_change_speed() will access the array
phy_device.phy_led_triggers, which has been freed in the error path
of phy_led_triggers_register(), which may lead to a crash.
- phy_led_triggers_unregister() will access the same array, leading to
crashes during s2ram or poweroff, like:
Unable to handle kernel NULL pointer dereference at virtual address 00000000
...
[<c04116d4>] (__list_del_entry_valid) from [<c05e8948>] (led_trigger_unregister+0x34/0xcc)
[<c05e8948>] (led_trigger_unregister) from [<c05336c4>] (phy_led_triggers_unregister+0x28/0x34)
[<c05336c4>] (phy_led_triggers_unregister) from [<c0531d44>] (phy_detach+0x30/0x74)
[<c0531d44>] (phy_detach) from [<c0538bdc>] (sh_eth_close+0x64/0x9c)
[<c0538bdc>] (sh_eth_close) from [<c04d4ce0>] (dpm_run_callback+0x48/0xc8)
or:
list_del corruption. prev->next should be dede6540, but was 2e323931
------------[ cut here ]------------
kernel BUG at lib/list_debug.c:52!
...
[<c02f6d70>] (__list_del_entry_valid) from [<c0425168>] (led_trigger_unregister+0x34/0xcc)
[<c0425168>] (led_trigger_unregister) from [<c03a05a0>] (phy_led_triggers_unregister+0x28/0x34)
[<c03a05a0>] (phy_led_triggers_unregister) from [<c039ec04>] (phy_detach+0x30/0x74)
[<c039ec04>] (phy_detach) from [<c03a4fc0>] (sh_eth_close+0x6c/0xa4)
[<c03a4fc0>] (sh_eth_close) from [<c0483234>] (__dev_close_many+0xac/0xd0)
To fix this, clear phy_device.phy_num_led_triggers in the error path of
phy_led_triggers_register() fails.
Note that the "No phy led trigger registered for speed" message will
still be printed on link speed changes, which is a good cue that
something went wrong with the LED triggers.
Fixes: 2e0bc452f4721520 ("net: phy: leds: add support for led triggers on phy link state change") Signed-off-by: Geert Uytterhoeven <[email protected]> Reviewed-by: Florian Fainelli <[email protected]> Signed-off-by: David S. Miller <[email protected]>
John Crispin [Wed, 25 Jan 2017 08:20:55 +0000 (09:20 +0100)]
net-next: ethernet: mediatek: change the compatible string
When the binding was defined, I was not aware that mt2701 was an earlier
version of the SoC. For sake of consistency, the ethernet driver should
use mt2701 inside the compat string as this is the earliest SoC with the
ethernet core.
The ethernet driver is currently of no real use until we finish and
upstream the DSA driver. There are no users of this binding yet. It should
be safe to fix this now before it is too late and we need to provide
backward compatibility for the mt7623-eth compat string.
John Crispin [Wed, 25 Jan 2017 08:20:54 +0000 (09:20 +0100)]
Documentation: devicetree: change the mediatek ethernet compatible string
When the binding was defined, I was not aware that mt2701 was an earlier
version of the SoC. For sake of consistency, the ethernet driver should
use mt2701 inside the compat string as this is the earliest SoC with the
ethernet core.
The ethernet driver is currently of no real use until we finish and
upstream the DSA driver. There are no users of this binding yet. It should
be safe to fix this now before it is too late and we need to provide
backward compatibility for the mt7623-eth compat string.
David S. Miller [Wed, 25 Jan 2017 18:27:14 +0000 (13:27 -0500)]
Merge branch 'bnxt_en-rtnl-fixes'
Michael Chan says:
====================
bnxt_en: Fix RTNL lock usage in bnxt_sp_task().
There are 2 function calls from bnxt_sp_task() that have buggy RTNL
usage. These 2 functions take RTNL lock under some conditions, but
some callers (such as open, ethtool) have already taken RTNL. These
3 patches fix the issue by making it clear that callers must take
RTNL. If the caller is bnxt_sp_task() which does not automatically
take RTNL, we add a common scheme for bnxt_sp_task() to call these
functions properly under RTNL.
====================
Michael Chan [Wed, 25 Jan 2017 07:55:09 +0000 (02:55 -0500)]
bnxt_en: Fix RTNL lock usage on bnxt_get_port_module_status().
bnxt_get_port_module_status() calls bnxt_update_link() which expects
RTNL to be held. In bnxt_sp_task() that does not hold RTNL, we need to
call it with a prior call to bnxt_rtnl_lock_sp() and the call needs to
be moved to the end of bnxt_sp_task().
Michael Chan [Wed, 25 Jan 2017 07:55:08 +0000 (02:55 -0500)]
bnxt_en: Fix RTNL lock usage on bnxt_update_link().
bnxt_update_link() is called from multiple code paths. Most callers,
such as open, ethtool, already hold RTNL. Only the caller bnxt_sp_task()
does not. So it is a bug to take RTNL inside bnxt_update_link().
Fix it by removing the RTNL inside bnxt_update_link(). The function
now expects the caller to always hold RTNL.
In bnxt_sp_task(), call bnxt_rtnl_lock_sp() before calling
bnxt_update_link(). We also need to move the call to the end of
bnxt_sp_task() since it will be clearing the BNXT_STATE_IN_SP_TASK bit.
Michael Chan [Wed, 25 Jan 2017 07:55:07 +0000 (02:55 -0500)]
bnxt_en: Fix bnxt_reset() in the slow path task.
In bnxt_sp_task(), we set a bit BNXT_STATE_IN_SP_TASK so that bnxt_close()
will synchronize and wait for bnxt_sp_task() to finish. Some functions
in bnxt_sp_task() require us to clear BNXT_STATE_IN_SP_TASK and then
acquire rtnl_lock() to prevent race conditions.
There are some bugs related to this logic. This patch refactors the code
to have common bnxt_rtnl_lock_sp() and bnxt_rtnl_unlock_sp() to handle
the RTNL and the clearing/setting of the bit. Multiple functions will
need the same logic. We also need to move bnxt_reset() to the end of
bnxt_sp_task(). Functions that clear BNXT_STATE_IN_SP_TASK must be the
last functions to be called in bnxt_sp_task(). The common scheme will
handle the condition properly.
Linus Torvalds [Wed, 25 Jan 2017 18:25:36 +0000 (10:25 -0800)]
Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Pull virtio/vhost fixes from Michael Tsirkin:
- ARM DMA fixes
- vhost vsock bugfix
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
vring: Force use of DMA API for ARM-based systems with legacy devices
virtio_mmio: Set DMA masks appropriately
vhost/vsock: handle vhost_vq_init_access() error
Xin Long [Tue, 24 Jan 2017 06:05:16 +0000 (14:05 +0800)]
sctp: sctp gso should set feature with NETIF_F_SG when calling skb_segment
Now sctp gso puts segments into skb's frag_list, then processes these
segments in skb_segment. But skb_segment handles them only when gs is
enabled, as it's in the same branch with skb's frags.
Although almost all the NICs support sg other than some old ones, but
since commit 1e16aa3ddf86 ("net: gso: use feature flag argument in all
protocol gso handlers"), features &= skb->dev->hw_enc_features, and
xfrm_output_gso call skb_segment with features = 0, which means sctp
gso would call skb_segment with sg = 0, and skb_segment would not work
as expected.
This patch is to fix it by setting features param with NETIF_F_SG when
calling skb_segment so that it can go the right branch to process the
skb's frag_list.
Xin Long [Tue, 24 Jan 2017 06:01:53 +0000 (14:01 +0800)]
sctp: sctp_addr_id2transport should verify the addr before looking up assoc
sctp_addr_id2transport is a function for sockopt to look up assoc by
address. As the address is from userspace, it can be a v4-mapped v6
address. But in sctp protocol stack, it always handles a v4-mapped
v6 address as a v4 address. So it's necessary to convert it to a v4
address before looking up assoc by address.
This patch is to fix it by calling sctp_verify_addr in which it can do
this conversion before calling sctp_endpoint_lookup_assoc, just like
what sctp_sendmsg and __sctp_connect do for the address from users.
xfs: extsize hints are not unlikely in xfs_bmap_btalloc
With COW files they are the hotpath, just like for files with the
extent size hint attribute. We really shouldn't micro-manage anything
but failure cases with unlikely.
Additionally Arnd Bergmann recently reported that one of these two
unlikely annotations causes link failures together with an upcoming
kernel instrumentation patch, so let's get rid of it ASAP.
Brian Foster [Wed, 25 Jan 2017 15:53:43 +0000 (07:53 -0800)]
xfs: remove racy hasattr check from attr ops
xfs_attr_[get|remove]() have unlocked attribute fork checks to optimize
away a lock cycle in cases where the fork does not exist or is otherwise
empty. This check is not safe, however, because an attribute fork short
form to extent format conversion includes a transient state that causes
the xfs_inode_hasattr() check to fail. Specifically,
xfs_attr_shortform_to_leaf() creates an empty extent format attribute
fork and then adds the existing shortform attributes to it.
This means that lookup of an existing xattr can spuriously return
-ENOATTR when racing against a setxattr that causes the associated
format conversion. This was originally reproduced by an untar on a
particularly configured glusterfs volume, but can also be reproduced on
demand with properly crafted xattr requests.
The format conversion occurs under the exclusive ilock. xfs_attr_get()
and xfs_attr_remove() already have the proper locking and checks further
down in the functions to handle this situation correctly. Drop the
unlocked checks to avoid the spurious failure and rely on the existing
logic.
Currently we try to rely on the global reserved block pool for block
allocations for the free inode btree, but I have customer reports
(fairly complex workload, need to find an easier reproducer) where that
is not enough as the AG where we free an inode that requires a new
finobt block is entirely full. This causes us to cancel a dirty
transaction and thus a file system shutdown.
I think the right way to guard against this is to treat the finot the same
way as the refcount btree and have a per-AG reservations for the possible
worst case size of it, and the patch below implements that.
Note that this could increase mount times with large finobt trees. In
an ideal world we would have added a field for the number of finobt
fields to the AGI, similar to what we did for the refcount blocks.
We should do add it next time we rev the AGI or AGF format by adding
new fields.
xfs: only update mount/resv fields on success in __xfs_ag_resv_init
Try to reserve the blocks first and only then update the fields in
or hanging off the mount structure. This way we can call __xfs_ag_resv_init
again after a previous failure.
Hans de Goede [Sun, 22 Jan 2017 12:24:05 +0000 (13:24 +0100)]
Revert "ACPI / video: Add force_native quirk for HP Pavilion dv6"
Revert commit 6276e53fa8c0 (ACPI / video: Add force_native quirk for
HP Pavilion dv6).
In the commit message for the quirk this revert removes I wrote:
"Note that there are quite a few HP Pavilion dv6 variants, some
woth ATI and some with NVIDIA hybrid gfx, both seem to need this
quirk to have working backlight control. There are also some versions
with only Intel integrated gfx, these may not need this quirk, but it
should not hurt there."
Unfortunately that seems wrong, I've already received 2 reports of
this commit causing regressions on some dv6 variants (at least one
of which actually has a nvidia GPU). So it seems that HP has made a
mess here by using the same model-name both in marketing and in the
DMI data for many different variants. Some of which need
acpi_backlight=native for functional backlight control (as the
quirk this commit reverts was doing), where as others are broken by
it. So lets get back to the old sitation so as to avoid regressing
on models which used to work without any kernel cmdline arguments
before.
Fixes: 6276e53fa8c0 (ACPI / video: Add force_native quirk for HP Pavilion dv6) Signed-off-by: Hans de Goede <[email protected]> Signed-off-by: Rafael J. Wysocki <[email protected]>
drm/i915: Split insertion/binding of an object into the VM
It is useful to have this trace as it pairs nicely with the vma_unbind
one to track vma activity.
Added inside the i915_vma_bind function (was outside before) to keep a
similar placement as trace_i915_vma_unbind.
Chris Wilson [Mon, 23 Jan 2017 21:29:39 +0000 (21:29 +0000)]
drm/i915: Move atomic state free from out of fence release
Fences are required to support being released from under an atomic context.
The drm_atomic_state struct may take a mutex when being released and so
we cannot drop a reference to the drm_atomic_state from the fence release
path directly, and so we need to defer that unreference to a worker.
drm/i915: Fix calculation of rotated x and y offsets for planar formats
Parameters tile_size, tile_width and tile_height were passed in the
wrong order to _intel_adjust_tile_offset() when calculating the rotated
offsets.
This doesn't fix any user visible bug, since for packed formats new
and old offset are the same and the rotated offsets are within a tile
before they are fed to _intel_adjust_tile_offset(). In that case, the
offsets are unchanged. That is not true for planar formats, but those
are currently not supported.
drm/i915: Don't init hpd polling for vlv and chv from runtime_suspend()
An error in the condition for avoiding the call to intel_hpd_poll_init()
for valleyview and cherryview from intel_runtime_suspend() caused it to
be called unconditionally. Fix it.
Chris Wilson [Sun, 15 Jan 2017 12:58:25 +0000 (12:58 +0000)]
drm/i915: Avoid drm_atomic_state_put(NULL) in intel_display_resume
intel_display_resume() may be called without an atomic state to restore,
i.e. dev_priv->modeset_reset_restore state is NULL. One such case is
following a lid open/close event and the forced modeset in
intel_lid_notify().
Alex Williamson [Tue, 24 Jan 2017 20:15:43 +0000 (13:15 -0700)]
drm/i915/gvt: Fix kmem_cache_create() name
According to kmem_cache_sanity_check(), spaces are not allowed in the
name of a cache and results in a kernel oops with CONFIG_DEBUG_VM.
Convert to underscores.
Linus Torvalds [Wed, 25 Jan 2017 00:54:39 +0000 (16:54 -0800)]
Merge branch 'akpm' (patches from Andrew)
Merge fixes from Andrew Morton:
"26 fixes"
* emailed patches from Andrew Morton <[email protected]>: (26 commits)
MAINTAINERS: add Dan Streetman to zbud maintainers
MAINTAINERS: add Dan Streetman to zswap maintainers
mm: do not export ioremap_page_range symbol for external module
mn10300: fix build error of missing fpu_save()
romfs: use different way to generate fsid for BLOCK or MTD
frv: add missing atomic64 operations
mm, page_alloc: fix premature OOM when racing with cpuset mems update
mm, page_alloc: move cpuset seqcount checking to slowpath
mm, page_alloc: fix fast-path race with cpuset update or removal
mm, page_alloc: fix check for NULL preferred_zone
kernel/panic.c: add missing \n
fbdev: color map copying bounds checking
frv: add atomic64_add_unless()
mm/mempolicy.c: do not put mempolicy before using its nodemask
radix-tree: fix private list warnings
Documentation/filesystems/proc.txt: add VmPin
mm, memcg: do not retry precharge charges
proc: add a schedule point in proc_pid_readdir()
mm: alloc_contig: re-allow CMA to compact FS pages
mm/slub.c: trace free objects at KERN_INFO
...
zhong jiang [Tue, 24 Jan 2017 23:18:52 +0000 (15:18 -0800)]
mm: do not export ioremap_page_range symbol for external module
Recently, I've found cases in which ioremap_page_range was used
incorrectly, in external modules, leading to crashes. This can be
partly attributed to the fact that ioremap_page_range is lower-level,
with fewer protections, as compared to the other functions that an
external module would typically call. Those include:
Coly Li [Tue, 24 Jan 2017 23:18:46 +0000 (15:18 -0800)]
romfs: use different way to generate fsid for BLOCK or MTD
Commit 8a59f5d25265 ("fs/romfs: return f_fsid for statfs(2)") generates
a 64bit id from sb->s_bdev->bd_dev. This is only correct when romfs is
defined with CONFIG_ROMFS_ON_BLOCK. If romfs is only defined with
CONFIG_ROMFS_ON_MTD, sb->s_bdev is NULL, referencing sb->s_bdev->bd_dev
will triger an oops.
Richard Weinberger points out that when CONFIG_ROMFS_BACKED_BY_BOTH=y,
both CONFIG_ROMFS_ON_BLOCK and CONFIG_ROMFS_ON_MTD are defined.
Therefore when calling huge_encode_dev() to generate a 64bit id, I use
the follow order to choose parameter,
- CONFIG_ROMFS_ON_BLOCK defined
use sb->s_bdev->bd_dev
- CONFIG_ROMFS_ON_BLOCK undefined and CONFIG_ROMFS_ON_MTD defined
use sb->s_dev when,
- both CONFIG_ROMFS_ON_BLOCK and CONFIG_ROMFS_ON_MTD undefined
leave id as 0
When CONFIG_ROMFS_ON_MTD is defined and sb->s_mtd is not NULL, sb->s_dev
is set to a device ID generated by MTD_BLOCK_MAJOR and mtd index,
otherwise sb->s_dev is 0.
This is a try-best effort to generate a uniq file system ID, if all the
above conditions are not meet, f_fsid of this romfs instance will be 0.
Generally only one romfs can be built on single MTD block device, this
method is enough to identify multiple romfs instances in a computer.
Vlastimil Babka [Tue, 24 Jan 2017 23:18:41 +0000 (15:18 -0800)]
mm, page_alloc: fix premature OOM when racing with cpuset mems update
Ganapatrao Kulkarni reported that the LTP test cpuset01 in stress mode
triggers OOM killer in few seconds, despite lots of free memory. The
test attempts to repeatedly fault in memory in one process in a cpuset,
while changing allowed nodes of the cpuset between 0 and 1 in another
process.
The problem comes from insufficient protection against cpuset changes,
which can cause get_page_from_freelist() to consider all zones as
non-eligible due to nodemask and/or current->mems_allowed. This was
masked in the past by sufficient retries, but since commit 682a3385e773
("mm, page_alloc: inline the fast path of the zonelist iterator") we fix
the preferred_zoneref once, and don't iterate over the whole zonelist in
further attempts, thus the only eligible zones might be placed in the
zonelist before our starting point and we always miss them.
A previous patch fixed this problem for current->mems_allowed. However,
cpuset changes also update the task's mempolicy nodemask. The fix has
two parts. We have to repeat the preferred_zoneref search when we
detect cpuset update by way of seqcount, and we have to check the
seqcount before considering OOM.
Vlastimil Babka [Tue, 24 Jan 2017 23:18:38 +0000 (15:18 -0800)]
mm, page_alloc: move cpuset seqcount checking to slowpath
This is a preparation for the following patch to make review simpler.
While the primary motivation is a bug fix, this also simplifies the fast
path, although the moved code is only enabled when cpusets are in use.
Vlastimil Babka [Tue, 24 Jan 2017 23:18:35 +0000 (15:18 -0800)]
mm, page_alloc: fix fast-path race with cpuset update or removal
Ganapatrao Kulkarni reported that the LTP test cpuset01 in stress mode
triggers OOM killer in few seconds, despite lots of free memory. The
test attempts to repeatedly fault in memory in one process in a cpuset,
while changing allowed nodes of the cpuset between 0 and 1 in another
process.
One possible cause is that in the fast path we find the preferred
zoneref according to current mems_allowed, so that it points to the
middle of the zonelist, skipping e.g. zones of node 1 completely. If
the mems_allowed is updated to contain only node 1, we never reach it in
the zonelist, and trigger OOM before checking the cpuset_mems_cookie.
This patch fixes the particular case by redoing the preferred zoneref
search if we switch back to the original nodemask. The condition is
also slightly changed so that when the last non-root cpuset is removed,
we don't miss it.
Note that this is not a full fix, and more patches will follow.
Vlastimil Babka [Tue, 24 Jan 2017 23:18:32 +0000 (15:18 -0800)]
mm, page_alloc: fix check for NULL preferred_zone
Patch series "fix premature OOM regression in 4.7+ due to cpuset races".
This is v2 of my attempt to fix the recent report based on LTP cpuset
stress test [1]. The intention is to go to stable 4.9 LTSS with this,
as triggering repeated OOMs is not nice. That's why the patches try to
be not too intrusive.
Unfortunately why investigating I found that modifying the testcase to
use per-VMA policies instead of per-task policies will bring the OOM's
back, but that seems to be much older and harder to fix problem. I have
posted a RFC [2] but I believe that fixing the recent regressions has a
higher priority.
Longer-term we might try to think how to fix the cpuset mess in a better
and less error prone way. I was for example very surprised to learn,
that cpuset updates change not only task->mems_allowed, but also
nodemask of mempolicies. Until now I expected the parameter to
alloc_pages_nodemask() to be stable. I wonder why do we then treat
cpusets specially in get_page_from_freelist() and distinguish HARDWALL
etc, when there's unconditional intersection between mempolicy and
cpuset. I would expect the nodemask adjustment for saving overhead in
g_p_f(), but that clearly doesn't happen in the current form. So we
have both crazy complexity and overhead, AFAICS.
Since commit c33d6c06f60f ("mm, page_alloc: avoid looking up the first
zone in a zonelist twice") we have a wrong check for NULL preferred_zone,
which can theoretically happen due to concurrent cpuset modification. We
check the zoneref pointer which is never NULL and we should check the zone
pointer. Also document this in first_zones_zonelist() comment per Michal
Hocko.
Vlastimil Babka [Tue, 24 Jan 2017 23:18:18 +0000 (15:18 -0800)]
mm/mempolicy.c: do not put mempolicy before using its nodemask
Since commit be97a41b291e ("mm/mempolicy.c: merge alloc_hugepage_vma to
alloc_pages_vma") alloc_pages_vma() can potentially free a mempolicy by
mpol_cond_put() before accessing the embedded nodemask by
__alloc_pages_nodemask(). The commit log says it's so "we can use a
single exit path within the function" but that's clearly wrong. We can
still do that when doing mpol_cond_put() after the allocation attempt.
Make sure the mempolicy is not freed prematurely, otherwise
__alloc_pages_nodemask() can end up using a bogus nodemask, which could
lead e.g. to premature OOM.
Fabian Frederick [Tue, 24 Jan 2017 23:18:13 +0000 (15:18 -0800)]
Documentation/filesystems/proc.txt: add VmPin
Commit bc3e53f682d9 ("mm: distinguish between mlocked and pinned pages")
added VmPin in /proc/<pid>/status. Report that in
Documentation/filesystems/proc.txt
David Rientjes [Tue, 24 Jan 2017 23:18:10 +0000 (15:18 -0800)]
mm, memcg: do not retry precharge charges
When memory.move_charge_at_immigrate is enabled and precharges are
depleted during move, mem_cgroup_move_charge_pte_range() will attempt to
increase the size of the precharge.
Prevent precharges from ever looping by setting __GFP_NORETRY. This was
probably the intention of the GFP_KERNEL & ~__GFP_NORETRY, which is
pointless as written.
Lucas Stach [Tue, 24 Jan 2017 23:18:05 +0000 (15:18 -0800)]
mm: alloc_contig: re-allow CMA to compact FS pages
Commit 73e64c51afc5 ("mm, compaction: allow compaction for GFP_NOFS
requests") changed compation to skip FS pages if not explicitly allowed
to touch them, but missed to update the CMA compact_control.
This leads to a very high isolation failure rate, crippling performance
of CMA even on a lightly loaded system. Re-allow CMA to compact FS
pages by setting the correct GFP flags, restoring CMA behavior and
performance to the kernel 4.9 level.
Daniel Thompson [Tue, 24 Jan 2017 23:18:02 +0000 (15:18 -0800)]
mm/slub.c: trace free objects at KERN_INFO
Currently when trace is enabled (e.g. slub_debug=T,kmalloc-128 ) the
trace messages are mostly output at KERN_INFO. However the trace code
also calls print_section() to hexdump the head of a free object. This
is hard coded to use KERN_ERR, meaning the console is deluged with trace
messages even if we've asked for quiet.
Fix this the obvious way but adding a level parameter to
print_section(), allowing calls from the trace code to use the same
trace level as other trace messages.
Andrea Arcangeli [Tue, 24 Jan 2017 23:17:59 +0000 (15:17 -0800)]
userfaultfd: fix SIGBUS resulting from false rwsem wakeups
With >=32 CPUs the userfaultfd selftest triggered a graceful but
unexpected SIGBUS because VM_FAULT_RETRY was returned by
handle_userfault() despite the UFFDIO_COPY wasn't completed.
This seems caused by rwsem waking the thread blocked in
handle_userfault() and we can't run up_read() before the wait_event
sequence is complete.
Keeping the wait_even sequence identical to the first one, would require
running userfaultfd_must_wait() again to know if the loop should be
repeated, and it would also require retaking the rwsem and revalidating
the whole vma status.
It seems simpler to wait the targeted wakeup so that if false wakeups
materialize we still wait for our specific wakeup event, unless of
course there are signals or the uffd was released.
Debug code collecting the stack trace of the wakeup showed this:
This always happens when the main userfault selftest thread is running
clone() while glibc runs either mprotect or mmap (both taking mmap_sem
down_write()) to allocate the thread stack of the background threads,
while locking/userfault threads already run at full throttle and are
susceptible to false wakeups that may cause handle_userfault() to return
before than expected (which results in graceful SIGBUS at the next
attempt).
This was reproduced only with >=32 CPUs because the loop to start the
thread where clone() is too quick with fewer CPUs, while with 32 CPUs
there's already significant activity on ~32 locking and userfault
threads when the last background threads are started with clone().
This >=32 CPUs SMP race condition is likely reproducible only with the
selftest because of the much heavier userfault load it generates if
compared to real apps.
We'll have to allow "one more" VM_FAULT_RETRY for the WP support and a
patch floating around that provides it also hidden this problem but in
reality only is successfully at hiding the problem.
False wakeups could still happen again the second time
handle_userfault() is invoked, even if it's a so rare race condition
that getting false wakeups twice in a row is impossible to reproduce.
This full fix is needed for correctness, the only alternative would be
to allow VM_FAULT_RETRY to be returned infinitely. With this fix the WP
support can stick to a strict "one more" VM_FAULT_RETRY logic (no need
of returning it infinite times to avoid the SIGBUS).
gcc-7 produces a harmless false-postive warning about a possible NULL
pointer access:
drivers/memstick/core/memstick.c: In function 'h_memstick_read_dev_id':
drivers/memstick/core/memstick.c:309:3: error: argument 2 null where non-null expected [-Werror=nonnull]
memcpy(mrq->data, buf, mrq->data_len);
This can't happen because the caller sets the command to 'MS_TPC_READ_REG',
which causes the data direction to be 'READ' and the NULL pointer not
accessed.
As a simple workaround for the warning, we can pass a pointer to the
data that we actually want to read into. This is not needed here, but
also harmless, and lets the compiler know that the access is ok.
Don Zickus [Tue, 24 Jan 2017 23:17:53 +0000 (15:17 -0800)]
kernel/watchdog: prevent false hardlockup on overloaded system
On an overloaded system, it is possible that a change in the watchdog
threshold can be delayed long enough to trigger a false positive.
This can easily be achieved by having a cpu spinning indefinitely on a
task, while another cpu updates watchdog threshold.
What happens is while trying to park the watchdog threads, the hrtimers
on the other cpus trigger and reprogram themselves with the new slower
watchdog threshold. Meanwhile, the nmi watchdog is still programmed
with the old faster threshold.
Because the one cpu is blocked, it prevents the thread parking on the
other cpus from completing, which is needed to shutdown the nmi watchdog
and reprogram it correctly. As a result, a false positive from the nmi
watchdog is reported.
Fix this by setting a park_in_progress flag to block all lockups until
the parking is complete.
Ross Zwisler [Tue, 24 Jan 2017 23:17:51 +0000 (15:17 -0800)]
dax: fix build warnings with FS_DAX and !FS_IOMAP
As reported by Arnd:
https://lkml.org/lkml/2017/1/10/756
Compiling with the following configuration:
# CONFIG_EXT2_FS is not set
# CONFIG_EXT4_FS is not set
# CONFIG_XFS_FS is not set
# CONFIG_FS_IOMAP depends on the above filesystems, as is not set
CONFIG_FS_DAX=y
generates build warnings about unused functions in fs/dax.c:
fs/dax.c:878:12: warning: `dax_insert_mapping' defined but not used [-Wunused-function]
static int dax_insert_mapping(struct address_space *mapping,
^~~~~~~~~~~~~~~~~~
fs/dax.c:572:12: warning: `copy_user_dax' defined but not used [-Wunused-function]
static int copy_user_dax(struct block_device *bdev, sector_t sector, size_t size,
^~~~~~~~~~~~~
fs/dax.c:542:12: warning: `dax_load_hole' defined but not used [-Wunused-function]
static int dax_load_hole(struct address_space *mapping, void **entry,
^~~~~~~~~~~~~
fs/dax.c:312:14: warning: `grab_mapping_entry' defined but not used [-Wunused-function]
static void *grab_mapping_entry(struct address_space *mapping, pgoff_t index,
^~~~~~~~~~~~~~~~~~
Now that the struct buffer_head based DAX fault paths and I/O path have
been removed we really depend on iomap support being present for DAX.
Make this explicit by selecting FS_IOMAP if we compile in DAX support.
This allows us to remove conditional selections of FS_IOMAP when FS_DAX
was present for ext2 and ext4, and to remove an #ifdef in fs/dax.c.
Keno Fischer [Tue, 24 Jan 2017 23:17:48 +0000 (15:17 -0800)]
mm/huge_memory.c: respect FOLL_FORCE/FOLL_COW for thp
In commit 19be0eaffa3a ("mm: remove gup_flags FOLL_WRITE games from
__get_user_pages()"), the mm code was changed from unsetting FOLL_WRITE
after a COW was resolved to setting the (newly introduced) FOLL_COW
instead. Simultaneously, the check in gup.c was updated to still allow
writes with FOLL_FORCE set if FOLL_COW had also been set.
However, a similar check in huge_memory.c was forgotten. As a result,
remote memory writes to ro regions of memory backed by transparent huge
pages cause an infinite loop in the kernel (handle_mm_fault sets
FOLL_COW and returns 0 causing a retry, but follow_trans_huge_pmd bails
out immidiately because `(flags & FOLL_WRITE) && !pmd_write(*pmd)` is
true.
While in this state the process is stil SIGKILLable, but little else
works (e.g. no ptrace attach, no other signals). This is easily
reproduced with the following code (assuming thp are set to always):
Fix this by updating follow_trans_huge_pmd in huge_memory.c analogously
to the update in gup.c in the original commit. The same pattern exists
in follow_devmap_pmd. However, we should not be able to reach that
check with FOLL_COW set, so add WARN_ONCE to make sure we notice if we
ever do.
memory_hotplug: make zone_can_shift() return a boolean value
online_{kernel|movable} is used to change the memory zone to
ZONE_{NORMAL|MOVABLE} and online the memory.
To check that memory zone can be changed, zone_can_shift() is used.
Currently the function returns minus integer value, plus integer
value and 0. When the function returns minus or plus integer value,
it means that the memory zone can be changed to ZONE_{NORNAL|MOVABLE}.
But when the function returns 0, there are two meanings.
One of the meanings is that the memory zone does not need to be changed.
For example, when memory is in ZONE_NORMAL and onlined by online_kernel
the memory zone does not need to be changed.
Another meaning is that the memory zone cannot be changed. When memory
is in ZONE_NORMAL and onlined by online_movable, the memory zone may
not be changed to ZONE_MOVALBE due to memory online limitation(see
Documentation/memory-hotplug.txt). In this case, memory must not be
onlined.
The patch changes the return type of zone_can_shift() so that memory
online operation fails when memory zone cannot be changed as follows:
Before applying patch:
# grep -A 35 "Node 2" /proc/zoneinfo
Node 2, zone Normal
<snip>
node_scanned 0
spanned 8388608
present 7864320
managed 7864320
# echo online_movable > memory4097/state
# grep -A 35 "Node 2" /proc/zoneinfo
Node 2, zone Normal
<snip>
node_scanned 0
spanned 8388608
present 8388608
managed 8388608
online_movable operation succeeded. But memory is onlined as
ZONE_NORMAL, not ZONE_MOVABLE.
After applying patch:
# grep -A 35 "Node 2" /proc/zoneinfo
Node 2, zone Normal
<snip>
node_scanned 0
spanned 8388608
present 7864320
managed 7864320
# echo online_movable > memory4097/state
bash: echo: write error: Invalid argument
# grep -A 35 "Node 2" /proc/zoneinfo
Node 2, zone Normal
<snip>
node_scanned 0
spanned 8388608
present 7864320
managed 7864320
online_movable operation failed because of failure of changing
the memory zone from ZONE_NORMAL to ZONE_MOVABLE
This is because the legacy virtio-blk device is behind an SMMU, so we
have consequently swizzled its DMA ops and configured the SMMU to
translate accesses. This then requires the vring code to use the DMA API
to establish translations, otherwise all transactions will result in
fatal faults and termination.
Given that ARM-based systems only see an SMMU if one is really present
(the topology is all described by firmware tables such as device-tree or
IORT), then we can safely use the DMA API for all legacy virtio devices.
Modern devices can advertise the prescense of an IOMMU using the
VIRTIO_F_IOMMU_PLATFORM feature flag.
Robin Murphy [Tue, 10 Jan 2017 17:51:17 +0000 (17:51 +0000)]
virtio_mmio: Set DMA masks appropriately
Once DMA API usage is enabled, it becomes apparent that virtio-mmio is
inadvertently relying on the default 32-bit DMA mask, which leads to
problems like rapidly exhausting SWIOTLB bounce buffers.
Ensure that we set the appropriate 64-bit DMA mask whenever possible,
with the coherent mask suitably limited for the legacy vring as per a0be1db4304f ("virtio_pci: Limit DMA mask to 44 bits for legacy virtio
devices").
Vineet Gupta [Tue, 21 Jun 2016 08:54:33 +0000 (14:24 +0530)]
ARCv2: smp-boot: wake_flag polling by non-Masters needs to be uncached
This is needed on HS38 cores, for setting up IO-Coherency aperture properly
The polling could perturb the caches and coherecy fabric which could be
wrong in the small window when Master is setting up IOC aperture etc
in arc_cache_init()
We do it only for ARCv2 based builds to not affect EZChip ARCompact
based platform.
David S. Miller [Tue, 24 Jan 2017 21:21:37 +0000 (16:21 -0500)]
Merge branch 'lwt-module-unload'
Robert Shearman says:
====================
net: Fix oops on state free after lwt module unload
An oops is seen in lwtstate_free after an lwt ops module has been
unloaded. This patchset fixes this by preventing modules implementing
lwtunnel ops from being unloaded whilst there's state alive using
those ops. The first patch adds fills in a new owner field in all lwt
ops and the second patch makes use of this to reference count the
modules as state is built and destroyed using them.
Changes in v3:
- don't put module reference if try_module_get fails on building state
Changes in v2:
- specify module owner for all modules as suggested by DaveM
- reference count all modules building lwt state, not just those ops
implementing destroy_state, as also suggested by DaveM.
- rebased on top of David Ahern's lwtunnel changes
====================
The problem is after the module for the encap can be unloaded the
corresponding ops is removed and is thus NULL here.
Modules implementing lwtunnel ops should not be allowed to unload
while there is state alive using those ops, so grab the module
reference for the ops on creating lwtunnel state and of course release
the reference when freeing the state.
Fixes: 1104d9ba443a ("lwtunnel: Add destroy state operation") Signed-off-by: Robert Shearman <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Robert Shearman [Tue, 24 Jan 2017 16:26:47 +0000 (16:26 +0000)]
net: Specify the owning module for lwtunnel ops
Modules implementing lwtunnel ops should not be allowed to unload
while there is state alive using those ops, so specify the owning
module for all lwtunnel ops.
Maor Gottlieb [Thu, 19 Jan 2017 13:25:58 +0000 (15:25 +0200)]
IB/rxe: Fix rxe dev insertion to rxe_dev_list
The first argument of list_add_tail is the new item and the second
is the head of the list. Fix the code to pass arguments in the
right order, otherwise not all the rxe devices will be removed
during teardown.
David S. Miller [Tue, 24 Jan 2017 21:14:59 +0000 (16:14 -0500)]
Merge branch 'tipc-topology-fixes'
Parthasarathy Bhuvaragan says:
====================
tipc: topology server fixes for nametable soft lockup
In this series, we revert the commit 333f796235a527 ("tipc: fix a
race condition leading to subscriber refcnt bug") and provide an
alternate solution to fix the race conditions in commits 2-4.
We have to do this as the above commit introduced a nametbl soft
lockup at module exit as described by patch#4.
====================
In tipc_server_stop(), we iterate over the connections with limiting
factor as server's idr_in_use. We ignore the fact that this variable
is decremented in tipc_close_conn(), leading to premature exit.
In this commit, we iterate until the we have no connections left.
tipc: ignore requests when the connection state is not CONNECTED
In tipc_conn_sendmsg(), we first queue the request to the outqueue
followed by the connection state check. If the connection is not
connected, we should not queue this message.
In this commit, we reject the messages if the connection state is
not CF_CONNECTED.
Commit 333f796235a527 ("tipc: fix a race condition leading to
subscriber refcnt bug") reveals a soft lockup while acquiring
nametbl_lock.
Before commit 333f796235a527, we call tipc_conn_shutdown() from
tipc_close_conn() in the context of tipc_topsrv_stop(). In that
context, we are allowed to grab the nametbl_lock.
Commit 333f796235a527, moved tipc_conn_release (renamed from
tipc_conn_shutdown) to the connection refcount cleanup. This allows
either tipc_nametbl_withdraw() or tipc_topsrv_stop() to the cleanup.
Since tipc_exit_net() first calls tipc_topsrv_stop() and then
tipc_nametble_withdraw() increases the chances for the later to
perform the connection cleanup.
The soft lockup occurs in the call chain of tipc_nametbl_withdraw(),
when it performs the tipc_conn_kref_release() as it tries to grab
nametbl_lock again while holding it already.
tipc_nametbl_withdraw() grabs nametbl_lock
tipc_nametbl_remove_publ()
tipc_subscrp_report_overlap()
tipc_subscrp_send_event()
tipc_conn_sendmsg()
<< if (con->flags != CF_CONNECTED) we do conn_put(),
triggering the cleanup as refcount=0. >>
tipc_conn_kref_release
tipc_sock_release
tipc_conn_release
tipc_subscrb_delete
tipc_subscrp_delete
tipc_nametbl_unsubscribe << Soft Lockup >>
The previous changes in this series fixes the race conditions fixed
by commit 333f796235a527. Hence we can now revert the commit.
Until now, the generic server framework maintains the connection
id's per subscriber in server's conn_idr. At tipc_close_conn, we
remove the connection id from the server list, but the connection is
valid until we call the refcount cleanup. Hence we have a window
where the server allocates the same connection to an new subscriber
leading to inconsistent reference count. We have another refcount
warning we grab the refcount in tipc_conn_lookup() for connections
with flag with CF_CONNECTED not set. This usually occurs at shutdown
when the we stop the topology server and withdraw TIPC_CFG_SRV
publication thereby triggering a withdraw message to subscribers.
In this commit, we:
1. remove the connection from the server list at recount cleanup.
2. grab the refcount for a connection only if CF_CONNECTED is set.
tipc: add subscription refcount to avoid invalid delete
Until now, the subscribers keep track of the subscriptions using
reference count at subscriber level. At subscription cancel or
subscriber delete, we delete the subscription only if the timer
was pending for the subscription. This approach is incorrect as:
1. del_timer() is not SMP safe, if on CPU0 the check for pending
timer returns true but CPU1 might schedule the timer callback
thereby deleting the subscription. Thus when CPU0 is scheduled,
it deletes an invalid subscription.
2. We export tipc_subscrp_report_overlap(), which accesses the
subscription pointer multiple times. Meanwhile the subscription
timer can expire thereby freeing the subscription and we might
continue to access the subscription pointer leading to memory
violations.
In this commit, we introduce subscription refcount to avoid deleting
an invalid subscription.
tipc: fix nametbl_lock soft lockup at node/link events
We trigger a soft lockup as we grab nametbl_lock twice if the node
has a pending node up/down or link up/down event while:
- we process an incoming named message in tipc_named_rcv() and
perform an tipc_update_nametbl().
- we have pending backlog items in the name distributor queue
during a nametable update using tipc_nametbl_publish() or
tipc_nametbl_withdraw().
The following are the call chain associated:
tipc_named_rcv() Grabs nametbl_lock
tipc_update_nametbl() (publish/withdraw)
tipc_node_subscribe()/unsubscribe()
tipc_node_write_unlock()
<< lockup occurs if an outstanding node/link event
exits, as we grabs nametbl_lock again >>
tipc_nametbl_withdraw() Grab nametbl_lock
tipc_named_process_backlog()
tipc_update_nametbl()
<< rest as above >>
The function tipc_node_write_unlock(), in addition to releasing the
lock processes the outstanding node/link up/down events. To do this,
we need to grab the nametbl_lock again leading to the lockup.
In this commit we fix the soft lockup by introducing a fast variant of
node_unlock(), where we just release the lock. We adapt the
node_subscribe()/node_unsubscribe() to use the fast variants.
netfilter: nf_tables: bump set->ndeact on set flush
Add missing set->ndeact update on each deactivated element from the set
flush path. Otherwise, sets with fixed size break after flush since
accounting breaks.
# nft add set x y { type ipv4_addr\; size 2\; }
# nft add element x y { 1.1.1.1 }
# nft add element x y { 1.1.1.2 }
# nft flush set x y
# nft add element x y { 1.1.1.1 }
<cmdline>:1:1-28: Error: Could not process rule: Too many open files in system
Fixes: 8411b6442e59 ("netfilter: nf_tables: support for set flushing") Reported-by: Elise Lennion <[email protected]> Signed-off-by: Pablo Neira Ayuso <[email protected]>
netfilter: nf_tables: fix set->nelems counting with no NLM_F_EXCL
If the element exists and no NLM_F_EXCL is specified, do not bump
set->nelems, otherwise we leak one set element slot. This problem
amplifies if the set is full since the abort path always decrements the
counter for the -ENFILE case too, giving one spare extra slot.
Fix this by moving set->nelems update to nft_add_set_elem() after
successful element insertion. Moreover, remove the element if the set is
full so there is no need to rely on the abort path to undo things
anymore.
Fixes: c016c7e45ddf ("netfilter: nf_tables: honor NLM_F_EXCL flag in set element insertion") Signed-off-by: Pablo Neira Ayuso <[email protected]>
Liping Zhang [Sun, 22 Jan 2017 14:10:32 +0000 (22:10 +0800)]
netfilter: nft_log: restrict the log prefix length to 127
First, log prefix will be truncated to NF_LOG_PREFIXLEN-1, i.e. 127,
at nf_log_packet(), so the extra part is useless.
Second, after adding a log rule with a very very long prefix, we will
fail to dump the nft rules after this _special_ one, but acctually,
they do exist. For example:
# name_65000=$(printf "%0.sQ" {1..65000})
# nft add rule filter output log prefix "$name_65000"
# nft add rule filter output counter
# nft add rule filter output counter
# nft list chain filter output
table ip filter {
chain output {
type filter hook output priority 0; policy accept;
}
}
So now, restrict the log prefix length to NF_LOG_PREFIXLEN-1.
Kenneth Lee [Thu, 5 Jan 2017 07:00:05 +0000 (15:00 +0800)]
IB/umem: Release pid in error and ODP flow
1. Release pid before enter odp flow
2. Release pid when fail to allocate memory
Fixes: 87773dd56d54 ("IB: ib_umem_release() should decrement mm->pinned_vm from ib_umem_get") Fixes: 8ada2c1c0c1d ("IB/core: Add support for on demand paging regions") Signed-off-by: Kenneth Lee <[email protected]> Reviewed-by: Haggai Eran <[email protected]> Reviewed-by: Leon Romanovsky <[email protected]> Signed-off-by: Doug Ledford <[email protected]>
Linus Torvalds [Tue, 24 Jan 2017 20:38:43 +0000 (12:38 -0800)]
Merge tag 'platform-drivers-x86-v4.10-4' of git://git.infradead.org/linux-platform-drivers-x86
Pull x86 platform-driver fixes from Andy Shevchenko:
"This is my first pull request since I become a co-maintainer of
Platform Drivers x86 subsystem. It's a bit bigger than usual due to
material collected for almost two weeks in a row.
MAINTAINERS:
- Add myself to X86 PLATFORM DRIVERS as a co-maintainer
ideapad-laptop:
- handle ACPI event 1
intel_mid_powerbtn:
- Set IRQ_ONESHOT
surface3-wmi:
- fix uninitialized symbol
- Shut up unused-function warning
mlx-platform:
- free first dev on error"
* tag 'platform-drivers-x86-v4.10-4' of git://git.infradead.org/linux-platform-drivers-x86:
MAINTAINERS: Add myself to X86 PLATFORM DRIVERS as a co-maintainer
platform/x86: ideapad-laptop: handle ACPI event 1
platform/x86: intel_mid_powerbtn: Set IRQ_ONESHOT
platform/x86: surface3-wmi: fix uninitialized symbol
platform/x86: surface3-wmi: Shut up unused-function warning
platform/x86: mlx-platform: free first dev on error
Ram Amrani [Tue, 24 Jan 2017 11:51:43 +0000 (13:51 +0200)]
RDMA/qedr: Dispatch port active event from qedr_add
Relying on qede to trigger qedr on startup is problematic. When probing
both if qedr loads slowly then qede can assume qedr is missing and not
trigger it. This patch adds a triggering from qedr and protects against
a race via an atomic bit.
projects / linux.git / log