Tom Herbert [Tue, 13 Mar 2018 19:01:43 +0000 (12:01 -0700)]
kcm: lock lower socket in kcm_attach
Need to lock lower socket in order to provide mutual exclusion
with kcm_unattach.
v2: Add Reported-by for syzbot
Fixes: ab7ac4eb9832e32a09f4e804 ("kcm: Kernel Connection Multiplexor module") Reported-by: syzbot+ea75c0ffcd353d32515f064aaebefc5279e6161e@syzkaller.appspotmail.com Signed-off-by: Tom Herbert <[email protected]> Signed-off-by: David S. Miller <[email protected]>
David S. Miller [Fri, 16 Mar 2018 14:03:48 +0000 (10:03 -0400)]
Merge branch 'vlan-untag-and-insert-fixes'
Toshiaki Makita says:
====================
Fix vlan untag and insertion for bridge and vlan with reorder_hdr off
As Brandon Carpenter reported[1], sending non-vlan-offloaded packets from
bridge devices ends up with corrupted packets. He narrowed down this problem
and found that the root cause is in skb_reorder_vlan_header().
While I was working on fixing this problem, I found that the function does
not work properly for double tagged packets with reorder_hdr off as well.
Patch 1 fixes these 2 problems in skb_reorder_vlan_header().
And it turned out that fixing skb_reorder_vlan_header() is not sufficient
to receive double tagged packets with reorder_hdr off while I was testing the
fix. Vlan tags got out of order when vlan devices with reorder_hdr disabled
were stacked. Patch 2 fixes this problem.
Toshiaki Makita [Tue, 13 Mar 2018 05:51:28 +0000 (14:51 +0900)]
vlan: Fix out of order vlan headers with reorder header off
With reorder header off, received packets are untagged in skb_vlan_untag()
called from within __netif_receive_skb_core(), and later the tag will be
inserted back in vlan_do_receive().
This caused out of order vlan headers when we create a vlan device on top
of another vlan device, because vlan_do_receive() inserts a tag as the
outermost vlan tag. E.g. the outer tag is first removed in skb_vlan_untag()
and inserted back in vlan_do_receive(), then the inner tag is next removed
and inserted back as the outermost tag.
This patch fixes the behaviour by inserting the inner tag at the right
position.
Toshiaki Makita [Tue, 13 Mar 2018 05:51:27 +0000 (14:51 +0900)]
net: Fix vlan untag for bridge and vlan_dev with reorder_hdr off
When we have a bridge with vlan_filtering on and a vlan device on top of
it, packets would be corrupted in skb_vlan_untag() called from
br_dev_xmit().
The problem sits in skb_reorder_vlan_header() used in skb_vlan_untag(),
which makes use of skb->mac_len. In this function mac_len is meant for
handling rx path with vlan devices with reorder_header disabled, but in
tx path mac_len is typically 0 and cannot be used, which is the problem
in this case.
The current code even does not properly handle rx path (skb_vlan_untag()
called from __netif_receive_skb_core()) with reorder_header off actually.
In rx path single tag case, it works as follows:
- Before skb_reorder_vlan_header()
mac_header data
v v
+-------------------+-------------+------+----
| ETH | VLAN | ETH |
| ADDRS | TPID | TCI | TYPE |
+-------------------+-------------+------+----
<-------- mac_len --------->
<------------->
to be removed
- After skb_reorder_vlan_header()
mac_header data
v v
+-------------------+------+----
| ETH | ETH |
| ADDRS | TYPE |
+-------------------+------+----
<-------- mac_len --------->
This is ok, but in rx double tag case, it corrupts packets:
- Before skb_reorder_vlan_header()
mac_header data
v v
+-------------------+-------------+-------------+------+----
| ETH | VLAN | VLAN | ETH |
| ADDRS | TPID | TCI | TPID | TCI | TYPE |
+-------------------+-------------+-------------+------+----
<--------------- mac_len ---------------->
<------------->
should be removed
<--------------------------->
actually will be removed
- After skb_reorder_vlan_header()
mac_header data
v v
+-------------------+------+----
| ETH | ETH |
| ADDRS | TYPE |
+-------------------+------+----
<--------------- mac_len ---------------->
So, two of vlan tags are both removed while only inner one should be
removed and mac_header (and mac_len) is broken.
skb_vlan_untag() is meant for removing the vlan header at (skb->data - 2),
so use skb->data and skb->mac_header to calculate the right offset.
Reported-by: Brandon Carpenter <[email protected]> Fixes: a6e18ff11170 ("vlan: Fix untag operations of stacked vlans with REORDER_HEADER off") Signed-off-by: Toshiaki Makita <[email protected]> Signed-off-by: David S. Miller <[email protected]>
The offending patch was merged in 4.16-rc4 and was promptly applied to
stable kernels 4.14.25 and 4.15.8.
The patch causes a corruption in several superblock items on big-endian
machines because of messed up endianity conversions. The damage is
manually repairable. A filesystem cannot be mounted again after it has
been unmounted once.
We do a full revert and not a fixup so stable can pick that patch ASAP.
Tom Lendacky [Thu, 8 Mar 2018 23:17:31 +0000 (17:17 -0600)]
KVM: x86: Fix device passthrough when SME is active
When using device passthrough with SME active, the MMIO range that is
mapped for the device should not be mapped encrypted. Add a check in
set_spte() to insure that a page is not mapped encrypted if that page
is a device MMIO page as indicated by kvm_is_mmio_pfn().
Michal Simek [Thu, 22 Feb 2018 14:19:37 +0000 (15:19 +0100)]
microblaze: Setup dependencies for ASM optimized lib functions
The patch:
"microblaze: Setup proper dependency for optimized lib functions"
(sha1: 7b6ce52be3f86520524711a6f33f3866f9339694)
didn't setup all dependencies properly.
Optimized lib functions in C are also present for little endian
and optimized library functions in assembler are implemented only for
big endian version.
x86/speculation: Remove Skylake C2 from Speculation Control microcode blacklist
In accordance with Intel's microcode revision guidance from March 6 MCU
rev 0xc2 is cleared on both Skylake H/S and Skylake Xeon E3 processors
that share CPUID 506E3.
Evgeniy Didin [Wed, 14 Mar 2018 19:30:51 +0000 (22:30 +0300)]
mmc: dw_mmc: fix falling from idmac to PIO mode when dw_mci_reset occurs
It was found that in IDMAC mode after soft-reset driver switches
to PIO mode.
That's what happens in case of DTO timeout overflow calculation failure:
1. soft-reset is called
2. driver restarts dma
3. descriptors states are checked, one of descriptor is owned by the IDMAC.
4. driver can't use DMA and then switches to PIO mode.
Failure was already fixed in:
https://www.spinics.net/lists/linux-mmc/msg48125.html.
Behaviour while soft-reset is not something we except or
even want to happen. So we switch from dw_mci_idmac_reset
to dw_mci_idmac_init, so descriptors are cleaned before starting dma.
And while at it explicitly zero des0 which otherwise might
contain garbage as being allocated by dmam_alloc_coherent().
Dave Airlie [Fri, 16 Mar 2018 02:51:35 +0000 (12:51 +1000)]
Merge tag 'drm-intel-fixes-2018-03-15' of git://anongit.freedesktop.org/drm/drm-intel into drm-fixes
Only GVT fixes:
- Two warnings fix for runtime pm and usr copy (Xiong, Zhenyu)
- OA context fix for vGPU profiling (Min)
- privilege batch buffer reloc fix (Fred)
* tag 'drm-intel-fixes-2018-03-15' of git://anongit.freedesktop.org/drm/drm-intel:
drm/i915/gvt: fix user copy warning by whitelist workload rb_tail field
drm/i915/gvt: Correct the privilege shadow batch buffer address
drm/i915/gvt: keep oa config in shadow ctx
drm/i915/gvt: Add runtime_pm_get/put into gvt_switch_mmio
Dan Williams [Fri, 16 Mar 2018 02:49:14 +0000 (19:49 -0700)]
acpi, numa: fix pxm to online numa node associations
Commit 99759869faf1 "acpi: Add acpi_map_pxm_to_online_node()" added
support for mapping a given proximity to its nearest, by SLIT distance,
online node. However, it sometimes returns unexpected results due to the
fact that it switches from comparing the PXM node to the last node that
was closer than the current max.
for_each_online_node(n) {
dist = node_distance(node, n);
if (dist < min_dist) {
min_dist = dist;
node = n; <---- from this point we're using the
wrong node for node_distance()
Leon Romanovsky [Thu, 15 Mar 2018 13:33:02 +0000 (15:33 +0200)]
RDMA/ucma: Check AF family prior resolving address
Garbage supplied by user will cause to UCMA module provide zero
memory size for memcpy(), because it wasn't checked, it will
produce unpredictable results in rdma_resolve_addr().
Linus Torvalds [Fri, 16 Mar 2018 01:57:14 +0000 (18:57 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Pull vfs fixes from Al Viro:
- backport-friendly part of lock_parent() race fix
- a fix for an assumption in the heurisic used by path_connected() that
is not true on NFS
- livelock fixes for d_alloc_parallel()
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
fs: Teach path_connected to handle nfs filesystems with multiple roots.
fs: dcache: Use READ_ONCE when accessing i_dir_seq
fs: dcache: Avoid livelock between d_alloc_parallel and __d_add
lock_parent() needs to recheck if dentry got __dentry_kill'ed under it
Lukas Wunner [Sat, 17 Feb 2018 12:40:23 +0000 (13:40 +0100)]
drm/nouveau/bl: Fix oops on driver unbind
Unbinding nouveau on a dual GPU MacBook Pro oopses because we iterate
over the bl_connectors list in nouveau_backlight_exit() but skipped
initializing it in nouveau_backlight_init(). Stacktrace for posterity:
Māris Nartišs [Fri, 16 Mar 2018 01:38:43 +0000 (11:38 +1000)]
drm/nouveau/mmu: ALIGN_DOWN correct variable
Commit 7110c89bb8852ff8b0f88ce05b332b3fe22bd11e ("mmu: swap out round
for ALIGN") replaced two calls to round/rounddown with ALIGN/ALIGN_DOWN,
but erroneously applied ALIGN_DOWN to a different variable (addr) and left
intended variable (tail) not rounded/ALIGNed.
As a result screen corruption, X lockups are observable. An example of kernel
log of affected system with NV98 card where it was bisected:
nouveau 0000:01:00.0: gr: TRAP_M2MF 00000002 [IN]
nouveau 0000:01:00.0: gr: TRAP_M2MF 00320951400007c00000000004000000
nouveau 0000:01:00.0: gr: 00200000 [] ch 1 [000fbbe000 DRM] subc 4 class 5039
mthd 0100 data 00000000
nouveau 0000:01:00.0: fb: trapped read at 0040000000 on channel 1
[0fbbe000 DRM]
engine 00 [PGRAPH] client 03 [DISPATCH] subclient 04 [M2M_IN] reason 00000006
[NULL_DMAOBJ]
Fixes bug 105173 ("[MCP79][Regression] Unhandled NULL pointer dereference in
nvkm_object_unmap since kernel 4.15")
https://bugs.freedesktop.org/show_bug.cgi?id=105173
fs: Teach path_connected to handle nfs filesystems with multiple roots.
On nfsv2 and nfsv3 the nfs server can export subsets of the same
filesystem and report the same filesystem identifier, so that the nfs
client can know they are the same filesystem. The subsets can be from
disjoint directory trees. The nfsv2 and nfsv3 filesystems provides no
way to find the common root of all directory trees exported form the
server with the same filesystem identifier.
The practical result is that in struct super s_root for nfs s_root is
not necessarily the root of the filesystem. The nfs mount code sets
s_root to the root of the first subset of the nfs filesystem that the
kernel mounts.
This effects the dcache invalidation code in generic_shutdown_super
currently called shrunk_dcache_for_umount and that code for years
has gone through an additional list of dentries that might be dentry
trees that need to be freed to accomodate nfs.
When I wrote path_connected I did not realize nfs was so special, and
it's hueristic for avoiding calling is_subdir can fail.
The practical case where this fails is when there is a move of a
directory from the subtree exposed by one nfs mount to the subtree
exposed by another nfs mount. This move can happen either locally or
remotely. With the remote case requiring that the move directory be cached
before the move and that after the move someone walks the path
to where the move directory now exists and in so doing causes the
already cached directory to be moved in the dcache through the magic
of d_splice_alias.
If someone whose working directory is in the move directory or a
subdirectory and now starts calling .. from the initial mount of nfs
(where s_root == mnt_root), then path_connected as a heuristic will
not bother with the is_subdir check. As s_root really is not the root
of the nfs filesystem this heuristic is wrong, and the path may
actually not be connected and path_connected can fail.
The is_subdir function might be cheap enough that we can call it
unconditionally. Verifying that will take some benchmarking and
the result may not be the same on all kernels this fix needs
to be backported to. So I am avoiding that for now.
Filesystems with snapshots such as nilfs and btrfs do something
similar. But as the directory tree of the snapshots are disjoint
from one another and from the main directory tree rename won't move
things between them and this problem will not occur.
David S. Miller [Thu, 15 Mar 2018 21:18:00 +0000 (14:18 -0700)]
sparc64: Fix regression in pmdp_invalidate().
pmdp_invalidate() was changed to update the pmd atomically
(to not lose dirty/access bits) and return the original pmd
value.
However, in doing so, we lost a lot of the essential work that
set_pmd_at() does, namely to update hugepage mapping counts and
queuing up the batched TLB flush entry.
Thus we were not flushing entries out of the TLB when making
such PMD changes.
Fix this by abstracting the accounting work of set_pmd_at() out into a
separate function, and call it from pmdp_establish().
Fixes: a8e654f01cb7 ("sparc64: update pmdp_invalidate() to return old pmd value") Signed-off-by: David S. Miller <[email protected]>
Paolo Bonzini [Thu, 15 Mar 2018 20:45:37 +0000 (21:45 +0100)]
Merge tag 'kvm-arm-fixes-for-v4.16-2' of git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm into kvm-master
kvm/arm fixes for 4.16, take 2
- Peace of mind locking fix in vgic_mmio_read_pending
- Allow hw-mapped interrupts to be reset when the VM resets
- Fix GICv2 multi-source SGI injection
- Fix MMIO synchronization for GICv2 on v3 emulation
- Remove excess verbosity on the console
Roman Mashak [Mon, 12 Mar 2018 20:20:58 +0000 (16:20 -0400)]
net sched actions: return explicit error when tunnel_key mode is not specified
If set/unset mode of the tunnel_key action is not provided, ->init() still
returns 0, and the caller proceeds with bogus 'struct tc_action *' object,
this results in crash:
% tc actions add action tunnel_key src_ip 1.1.1.1 dst_ip 2.2.2.1 id 7 index 1
Hans de Goede [Wed, 14 Mar 2018 22:06:03 +0000 (23:06 +0100)]
Bluetooth: hci_bcm: Set pulsed_host_wake flag in sleep parameters
The IRQ output of the bcm bt-device is really a level IRQ signal, which
signals a logical high as long as the device's buffer contains data. Since
the draining in the buffer is done in the tty driver, we cannot (easily)
wait in a threaded interrupt handler for the draining, after which the
IRQ should go low again.
So instead we treat the IRQ as an edge interrupt. This opens the window
for a theoretical race where we wakeup, read some data and then autosuspend
*before* the IRQ has gone (logical) low, followed by the device just at
that moment receiving more data, causing the IRQ to stay high and we never
see an edge.
Since we call pm_runtime_mark_last_busy() on every received byte, there
should be plenty time for the IRQ to go (logical) low before we ever
suspend, so this should never happen, but after commit 43fff7683468
("Bluetooth: hci_bcm: Streamline runtime PM code"), which has been reverted
since, this was actually happening causing the device to get stuck in
runtime suspend.
The bcm bt-device actually has a workaround for this, if we set the
pulsed_host_wake flag in the sleep parameters, then the device monitors
if the host is draining the buffer and if not then after a timeout the
device will pulse the IRQ line, causing us to see an edge, fixing the
stuck in suspend condition.
This commit sets the pulsed_host_wake flag to fix the (mostly theoretical)
race caused by us treating the IRQ as an edge IRQ.
This reverts commit 43fff7683468 ("Bluetooth: hci_bcm: Streamline runtime
PM code"). The commit msg for this commit states "No functional change
intended.", but replacing:
Does result in a functional change, pm_request_resume() only calls
pm_runtime_mark_last_busy() if the device was suspended before the call.
This results in the following happening:
1) Device is runtime suspended
2) Device drives host_wake IRQ logically high as it starts receiving data
3) bcm_host_wake() gets called, causes the device to runtime-resume,
current time gets marked as last_busy time
4) After 5 seconds the autosuspend timer expires and the dev autosuspends
as no one has been calling pm_runtime_mark_last_busy(), the device was
resumed during those 5 seconds, so all the pm_request_resume() calls
while receiving data and/or bcm_host_wake() calls were nops
5) If 4) happens while the device has (just received) data in its buffer to
be read by the host the IRQ line is *already* / still logically high
when we autosuspend and since we use an edge triggered IRQ, the IRQ
will never trigger, causing the device to get stuck in suspend
Therefor this commit has to be reverted, so that we avoid the device
getting stuck in suspend.
Takashi Iwai [Thu, 15 Mar 2018 16:02:34 +0000 (17:02 +0100)]
Bluetooth: btusb: Fix quirk for Atheros 1525/QCA6174
The Atheros 1525/QCA6174 BT doesn't seem working properly on the
recent kernels, as it tries to load a wrong firmware
ar3k/AthrBT_0x00000200.dfu and it fails.
This seems to have been a problem for some time, and the known
workaround is to apply BTUSB_QCA_ROM quirk instead of BTUSB_ATH3012.
Eddie James [Thu, 8 Mar 2018 20:57:20 +0000 (14:57 -0600)]
clk: aspeed: Prevent reset if clock is enabled
According to the Aspeed specification, the reset and enable sequence
should be done when the clock is stopped. The specification doesn't
define behavior if the reset is done while the clock is enabled.
From testing on the AST2500, the LPC Controller has problems if the
clock is reset while enabled.
Therefore, check whether the clock is enabled or not before performing
the reset and enable sequence in the Aspeed clock driver.
Eddie James [Thu, 8 Mar 2018 20:57:19 +0000 (14:57 -0600)]
clk: aspeed: Fix is_enabled for certain clocks
Some of the Aspeed clocks are disabled by setting the relevant bit in
the "clock stop control" register to one, while others are disabled by
setting their bit to zero. The driver already uses a flag per gate to
identify this behavior, but doesn't apply it in the clock is_enabled
function.
Use the existing gate flag to correctly return whether or not a clock
is enabled in the aspeed_clk_is_enabled function.
Signed-off-by: Eddie James <[email protected]> Fixes: 6671507f0fbd ("clk: aspeed: Handle inverse polarity of USB port 1 clock gate") Reviewed-by: Joel Stanley <[email protected]> Signed-off-by: Stephen Boyd <[email protected]>
Linus Torvalds [Thu, 15 Mar 2018 18:04:46 +0000 (11:04 -0700)]
Merge tag 'for-4.16/dm-fixes-3' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm
Pull device mapper fixes from Mike Snitzer:
- a stable DM multipath fix to restore ability to pass integrity data
- two DM multipath fixes for a fix that was merged into 4.16-rc5
* tag 'for-4.16/dm-fixes-3' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
dm mpath: fix passing integrity data
dm mpath: eliminate need to use scsi_device_from_queue
dm mpath: fix uninitialized 'pg_init_wait' waitqueue_head NULL pointer
Lucas Stach [Thu, 15 Mar 2018 09:11:59 +0000 (10:11 +0100)]
drm/imx: move arming of the vblank event to atomic_flush
Right now the vblank event completion is racing with the atomic update,
which is especially bad when the PRE is in use, as one of the hardware
issue workaround might extend the atomic commit for quite some time.
If the vblank IRQ happens to trigger during that time, we will prematurely
signal the atomic commit completion to userspace, which causes tearing
when userspace re-uses a framebuffer we haven't managed to flip away from
yet.
Fabio Estevam [Tue, 13 Feb 2018 19:11:34 +0000 (17:11 -0200)]
drm/imx: ipuv3-plane: Make functions static when possible
ipu_plane_state_reset(), ipu_plane_duplicate_state() and
ipu_plane_destroy_state() are only used in this file, so make them static.
This fixes the following sparse warnings:
drivers/gpu/drm/imx/ipuv3-plane.c:275:6: warning: symbol 'ipu_plane_state_reset' was not declared. Should it be static?
drivers/gpu/drm/imx/ipuv3-plane.c:295:24: warning: symbol 'ipu_plane_duplicate_state' was not declared. Should it be static?
drivers/gpu/drm/imx/ipuv3-plane.c:309:6: warning: symbol 'ipu_plane_destroy_state' was not declared. Should it be static?
Arnd Bergmann [Thu, 15 Mar 2018 16:19:57 +0000 (17:19 +0100)]
gpu: ipu-v3: prg: avoid possible array underflow
gcc-8 reports that we access an array with a negative index
in an error case:
drivers/gpu/ipu-v3/ipu-prg.c: In function 'ipu_prg_channel_disable':
drivers/gpu/ipu-v3/ipu-prg.c:252:43: error: array subscript -22 is below array bounds of 'struct ipu_prg_channel[3]' [-Werror=array-bounds]
This moves the range check in front of the first time that
variable gets used.
Ursula Braun [Tue, 13 Mar 2018 09:41:54 +0000 (10:41 +0100)]
net/smc: simplify wait when closing listen socket
Closing of a listen socket wakes up kernel_accept() of
smc_tcp_listen_worker(), and then has to wait till smc_tcp_listen_worker()
gives up the internal clcsock. The wait logic introduced with
commit 127f49705823 ("net/smc: release clcsock from tcp_listen_worker")
might wait longer than necessary. This patch implements the idea to
implement the wait just with flush_work(), and gets rid of the extra
smc_close_wait_listen_clcsock() function.
Fixes: 127f49705823 ("net/smc: release clcsock from tcp_listen_worker") Reported-by: Hans Wippel <[email protected]> Signed-off-by: Ursula Braun <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Thierry Reding [Thu, 15 Mar 2018 10:09:35 +0000 (11:09 +0100)]
drm/tegra: plane: Fix RGB565 format on older Tegra
The opaque/alpha format conversion code is currently only looking at
XRGB formats because they have an equivalent ARGB format. The opaque
format for RGB565 is RGB565 itself, much like the YUV formats map to
themselves.
Bastian Stender [Thu, 8 Mar 2018 14:08:11 +0000 (15:08 +0100)]
mmc: block: fix updating ext_csd caches on ioctl call
PARTITION_CONFIG is cached in mmc_card->ext_csd.part_config and the
currently active partition in mmc_blk_data->part_curr. These caches do
not always reflect changes if the ioctl call modifies the
PARTITION_CONFIG registers, e.g. by changing BOOT_PARTITION_ENABLE.
Write the PARTITION_CONFIG value extracted from the ioctl call to the
cache and update the currently active partition accordingly. This
ensures that the user space cannot change the values behind the
kernel's back. The next call to mmc_blk_part_switch() will operate on
the data set by the ioctl and reflect the changes appropriately.
fred gao [Thu, 15 Mar 2018 05:21:10 +0000 (13:21 +0800)]
drm/i915/gvt: Correct the privilege shadow batch buffer address
Once the ring buffer is copied to ring_scan_buffer and scanned,
the shadow batch buffer start address is only updated into
ring_scan_buffer, not the real ring address allocated through
intel_ring_begin in later copy_workload_to_ring_buffer.
This patch is only to set the right shadow batch buffer address
from Ring buffer, not include the shadow_wa_ctx.
v2:
- refine some comments. (Zhenyu)
v3:
- fix typo in title. (Zhenyu)
v4:
- remove the unnecessary comments. (Zhenyu)
- add comments in bb_start_cmd_va update. (Zhenyu)
Linus Torvalds [Thu, 15 Mar 2018 00:02:49 +0000 (17:02 -0700)]
Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull SCSI fixes from James Bottomley:
"This is four patches, consisting of one regression from the merge
window (qla2xxx), one long-standing memory leak (sd_zbc), one event
queue mislabelling which we want to eliminate to discourage the
pattern (mpt3sas), and one behaviour change because re-reading the
partition table shouldn't clear the ro flag"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: sd: Keep disk read-only when re-reading partition
scsi: qla2xxx: Fix crashes in qla2x00_probe_one on probe failure
scsi: sd_zbc: Fix potential memory leak
scsi: mpt3sas: Do not mark fw_event workqueue as WQ_MEM_RECLAIM
Joern Engel [Tue, 13 Mar 2018 18:36:49 +0000 (11:36 -0700)]
btree: avoid variable-length allocations
geo->keylen cannot be larger than 4. So we might as well make
fixed-size allocations.
Given the one remaining user, geo->keylen cannot even be larger than 1.
Logfs used to have 64bit and 128bit keys, tcm_qla2xxx only has 32bit
keys. But let's not break the code if we don't have to.
Linus Torvalds [Wed, 14 Mar 2018 23:51:33 +0000 (16:51 -0700)]
Merge branch 'percpu_ref-rcu-audit-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/misc
Pull percpu_ref rcu fixes from Tejun Heo:
"Jann Horn found that aio was depending on the internal RCU grace
periods of percpu-ref and that it's broken because aio uses regular
RCU while percpu_ref uses sched-RCU.
Depending on percpu_ref's internal grace periods isn't a good idea
because
- The RCU type might not match.
- percpu_ref's grace periods are used to switch to atomic mode. They
aren't between the last put and the invocation of the last release.
This is easy to get confused about and can lead to subtle bugs.
- percpu_ref might not have grace periods at all depending on its
current operation mode.
This patchset audits and fixes percpu_ref users for their RCU usages"
[ There's a continuation of this series that clarifies percpu_ref
documentation that the internal grace periods must not be depended
upon, and introduces rcu_work to simplify bouncing to a workqueue
after an RCU grace period.
That will go in for 4.17 - this is just the minimal set with the fixes
that are tagged for -stable ]
* 'percpu_ref-rcu-audit-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/misc:
RDMAVT: Fix synchronization around percpu_ref
fs/aio: Use RCU accessors for kioctx_table->table[]
fs/aio: Add explicit RCU grace period when freeing kioctx
Commit 864b75f9d6b0 ("mm/page_alloc: fix memmap_init_zone pageblock
alignment") modified the logic in memmap_init_zone() to initialize
struct pages associated with invalid PFNs, to appease a VM_BUG_ON()
in move_freepages(), which is redundant by its own admission, and
dereferences struct page fields to obtain the zone without checking
whether the struct pages in question are valid to begin with.
Commit 864b75f9d6b0 only makes it worse, since the rounding it does
may cause pfn assume the same value it had in a prior iteration of
the loop, resulting in an infinite loop and a hang very early in the
boot. Also, since it doesn't perform the same rounding on start_pfn
itself but only on intermediate values following an invalid PFN, we
may still hit the same VM_BUG_ON() as before.
So instead, let's fix this at the core, and ensure that the BUG
check doesn't dereference struct page fields of invalid pages.
Dave Airlie [Wed, 14 Mar 2018 23:26:11 +0000 (09:26 +1000)]
Merge tag 'drm-intel-fixes-2018-03-14' of git://anongit.freedesktop.org/drm/drm-intel into drm-fixes
- 1 display fix for bxt
- 1 gem fix for fences
- 1 gem/pm fix for rps freq
* tag 'drm-intel-fixes-2018-03-14' of git://anongit.freedesktop.org/drm/drm-intel:
drm/i915: Kick the rps worker when changing the boost frequency
drm/i915: Only prune fences after wait-for-all
drm/i915: Enable VBT based BL control for DP
Dave Airlie [Wed, 14 Mar 2018 23:23:54 +0000 (09:23 +1000)]
Merge branch 'drm-fixes-4.16' of git://people.freedesktop.org/~agd5f/linux into drm-fixes
A few fixes for 4.16:
- Fix a backlight S/R regression on amdgpu
- Fix prime teardown on radeon and amdgpu
- DP fix for amdgpu
* 'drm-fixes-4.16' of git://people.freedesktop.org/~agd5f/linux:
drm/amdgpu/dce: Don't turn off DP sink when disconnected
drm/amdgpu: save/restore backlight level in legacy dce code
drm/radeon: fix prime teardown order
drm/amdgpu: fix prime teardown order
Arnd Bergmann [Tue, 20 Feb 2018 20:56:27 +0000 (21:56 +0100)]
infiniband: bnxt_re: use BIT_ULL() for 64-bit bit masks
On 32-bit targets, we otherwise get a warning about an impossible constant
integer expression:
In file included from include/linux/kernel.h:11,
from include/linux/interrupt.h:6,
from drivers/infiniband/hw/bnxt_re/ib_verbs.c:39:
drivers/infiniband/hw/bnxt_re/ib_verbs.c: In function 'bnxt_re_query_device':
include/linux/bitops.h:7:24: error: left shift count >= width of type [-Werror=shift-count-overflow]
#define BIT(nr) (1UL << (nr))
^~
drivers/infiniband/hw/bnxt_re/bnxt_re.h:61:34: note: in expansion of macro 'BIT'
#define BNXT_RE_MAX_MR_SIZE_HIGH BIT(39)
^~~
drivers/infiniband/hw/bnxt_re/bnxt_re.h:62:30: note: in expansion of macro 'BNXT_RE_MAX_MR_SIZE_HIGH'
#define BNXT_RE_MAX_MR_SIZE BNXT_RE_MAX_MR_SIZE_HIGH
^~~~~~~~~~~~~~~~~~~~~~~~
drivers/infiniband/hw/bnxt_re/ib_verbs.c:149:25: note: in expansion of macro 'BNXT_RE_MAX_MR_SIZE'
ib_attr->max_mr_size = BNXT_RE_MAX_MR_SIZE;
^~~~~~~~~~~~~~~~~~~
Fixes: 872f3578241d ("RDMA/bnxt_re: Add support for MRs with Huge pages") Signed-off-by: Arnd Bergmann <[email protected]> Signed-off-by: Jason Gunthorpe <[email protected]>
Arnd Bergmann [Tue, 20 Feb 2018 20:56:26 +0000 (21:56 +0100)]
infiniband: qplib_fp: fix pointer cast
Building for a 32-bit target results in a couple of warnings from casting
between a 32-bit pointer and a 64-bit integer:
drivers/infiniband/hw/bnxt_re/qplib_fp.c: In function 'bnxt_qplib_service_nq':
drivers/infiniband/hw/bnxt_re/qplib_fp.c:333:23: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
bnxt_qplib_arm_srq((struct bnxt_qplib_srq *)q_handle,
^
drivers/infiniband/hw/bnxt_re/qplib_fp.c:336:12: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
(struct bnxt_qplib_srq *)q_handle,
^
In file included from include/linux/byteorder/little_endian.h:5,
from arch/arm/include/uapi/asm/byteorder.h:22,
from include/asm-generic/bitops/le.h:6,
from arch/arm/include/asm/bitops.h:342,
from include/linux/bitops.h:38,
from include/linux/kernel.h:11,
from include/linux/interrupt.h:6,
from drivers/infiniband/hw/bnxt_re/qplib_fp.c:39:
drivers/infiniband/hw/bnxt_re/qplib_fp.c: In function 'bnxt_qplib_create_srq':
include/uapi/linux/byteorder/little_endian.h:31:43: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast]
#define __cpu_to_le64(x) ((__force __le64)(__u64)(x))
^
include/linux/byteorder/generic.h:86:21: note: in expansion of macro '__cpu_to_le64'
#define cpu_to_le64 __cpu_to_le64
^~~~~~~~~~~~~
drivers/infiniband/hw/bnxt_re/qplib_fp.c:569:19: note: in expansion of macro 'cpu_to_le64'
req.srq_handle = cpu_to_le64(srq);
Using a uintptr_t as an intermediate works on all architectures.
Fixes: 37cb11acf1f7 ("RDMA/bnxt_re: Add SRQ support for Broadcom adapters") Signed-off-by: Arnd Bergmann <[email protected]> Signed-off-by: Jason Gunthorpe <[email protected]>
Dan Williams [Wed, 14 Mar 2018 04:36:22 +0000 (21:36 -0700)]
x86, memremap: fix altmap accounting at free
Commit 24b6d4164348 "mm: pass the vmem_altmap to vmemmap_free" converted
the vmemmap_free() path to pass the altmap argument all the way through
the call chain rather than looking it up based on the page.
Unfortunately that ends up over freeing altmap allocated pages in some
cases since free_pagetable() is used to free both memmap space and pte
space, where only the memmap stored in huge pages uses altmap
allocations.
Given that altmap allocations for memmap space are special cased in
vmemmap_populate_hugepages() add a symmetric / special case
free_hugepage_table() to handle altmap freeing, and cleanup the unneeded
passing of altmap to leaf functions that do not require it.
Without this change the sanity check accounting in
devm_memremap_pages_release() will throw a warning with the following
signature.
nd_pmem pfn10.1: devm_memremap_pages_release: failed to free all reserved pages
WARNING: CPU: 44 PID: 3539 at kernel/memremap.c:310 devm_memremap_pages_release+0x1c7/0x220
CPU: 44 PID: 3539 Comm: ndctl Tainted: G L 4.16.0-rc1-linux-stable #7
RIP: 0010:devm_memremap_pages_release+0x1c7/0x220
[..]
Call Trace:
release_nodes+0x225/0x270
device_release_driver_internal+0x15d/0x210
bus_remove_device+0xe2/0x160
device_del+0x130/0x310
? klist_release+0x56/0x100
? nd_region_notify+0xc0/0xc0 [libnvdimm]
device_unregister+0x16/0x60
This was missed in testing since not all configurations will trigger
this warning.
wrote 65536/65536 bytes at offset 0
64 KiB, 4 ops; 0.0000 sec (121.359 MiB/sec and 7766.9903 ops/sec)
/media/scratch/file5:
EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS
0: [0..127]: 24576..24703 128 0x2001
/media/scratch/file5:
EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS
0: [0..127]: 24576..24703 128 0x1
This is because btrfs_check_shared calls find_parent_nodes
repeatedly in a loop, passing a share_check struct to report
the count of shared extent. But btrfs_check_shared does not
re-initialize the count value to zero for subsequent calls
from the loop, resulting in a false share count value. This
is a regressive behavior from 4.13.
With proper re-initialization the test result is as follows:
wrote 65536/65536 bytes at offset 0
64 KiB, 4 ops; 0.0000 sec (110.035 MiB/sec and 7042.2535 ops/sec)
/media/scratch/file5:
EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS
0: [0..127]: 24576..24703 128 0x1
/media/scratch/file5:
EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS
0: [0..127]: 24576..24703 128 0x1
which corrects the regression.
Fixes: 3ec4d3238ab ("btrfs: allow backref search checks for shared extents") Signed-off-by: Edmund Nadolski <[email protected]>
[ add text from cover letter to changelog ] Signed-off-by: David Sterba <[email protected]>
This is reproducible in a cycle, where a series of writes is followed by
SCSI device delete command. The test may take up to few minutes.
Fixes: 74d46992e0d9 ("block: replace bi_bdev with a gendisk pointer and partitions index")
[ no signed-off-by provided ]
Author: Dmitriy Gorokh <[email protected]> Reviewed-by: Liu Bo <[email protected]> Reviewed-by: David Sterba <[email protected]> Signed-off-by: David Sterba <[email protected]>
Mark Bloch [Wed, 14 Mar 2018 07:14:15 +0000 (09:14 +0200)]
IB/mlx5: Fix cleanup order on unload
On load we create private CQ/QP/PD in order to be used by UMR, we create
those resources after we register ourself as an IB device, and we destroy
them after we unregister as an IB device. This was changed by commit 16c1975f1032 ("IB/mlx5: Create profile infrastructure to add and remove
stages") which moved the destruction before we unregistration. This
allowed to trigger an invalid memory access when unloading mlx5_ib while
there are open resources:
We restore the original behavior by breaking the UMR stage into two parts,
pre and post IB registration stages, this way we can restore the original
functionality and maintain clean separation of logic between stages.
Fixes: 16c1975f1032 ("IB/mlx5: Create profile infrastructure to add and remove stages") Signed-off-by: Mark Bloch <[email protected]> Signed-off-by: Leon Romanovsky <[email protected]> Signed-off-by: Doug Ledford <[email protected]>
Alex Deucher [Thu, 8 Mar 2018 14:56:01 +0000 (09:56 -0500)]
drm/amdgpu: save/restore backlight level in legacy dce code
Save/restore the backlight level scratch register in S3/S4 so the
backlight level comes back at the previously requested level.
Bug: https://bugzilla.kernel.org/show_bug.cgi?id=199047 Fixes: 4ec6ecf48c64d (drm/amdgpu: drop scratch regs save and restore from S3/S4 handling) Acked-by: Michel Dänzer <[email protected]> Reviewed-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
Leon Romanovsky [Tue, 13 Mar 2018 16:37:27 +0000 (18:37 +0200)]
RDMA/ucma: Don't allow join attempts for unsupported AF family
Users can provide garbage while calling to ucma_join_ip_multicast(),
it will indirectly cause to rdma_addr_size() return 0, making the
call to ucma_process_join(), which had the right checks, but it is
better to check the input as early as possible.
Linus Torvalds [Wed, 14 Mar 2018 20:01:14 +0000 (13:01 -0700)]
Merge tag 'platform-drivers-x86-v4.16-7' of git://git.infradead.org/linux-platform-drivers-x86
Pull x86 platform drives fixes from Darren Hart:
- DELL_SMBIOS conditionally depends on ACPI_WMI in the same way it
depends on DCDBAS, update the Kconfig accordingly.
- fix the dell driver init order to ensure that the driver dependencies
are met, avoiding race conditions resulting in boot failure on
certain systems when the drivers are built-in.
* tag 'platform-drivers-x86-v4.16-7' of git://git.infradead.org/linux-platform-drivers-x86:
platform/x86: Fix dell driver init order
platform/x86: dell-smbios: Resolve dependency error on ACPI_WMI
Tatyana Nikolova [Mon, 12 Mar 2018 22:14:02 +0000 (17:14 -0500)]
RDMA/core: Do not use invalid destination in determining port reuse
cma_port_is_unique() allows local port reuse if the quad (source
address and port, destination address and port) for this connection
is unique. However, if the destination info is zero or unspecified, it
can't make a correct decision but still allows port reuse. For example,
sometimes rdma_bind_addr() is called with unspecified destination and
reusing the port can lead to creating a connection with a duplicate quad,
after the destination is resolved. The issue manifests when MPI scale-up
tests hang after the duplicate quad is used.
Set the destination address family and add checks for zero destination
address and port to prevent source port reuse based on invalid destination.
Steffen Maier [Wed, 14 Mar 2018 14:33:06 +0000 (15:33 +0100)]
dm mpath: fix passing integrity data
After v4.12 commit e2460f2a4bc7 ("dm: mark targets that pass integrity
data"), dm-multipath, e.g. on DIF+DIX SCSI disk paths, does not support
block integrity any more. So add it to the whitelist.
This is also a pre-requisite to use block integrity with other dm layer(s)
on top of multipath, such as kpartx partitions (dm-linear) or LVM.
Leon Romanovsky [Mon, 12 Mar 2018 19:26:37 +0000 (21:26 +0200)]
RDMA/mlx5: Fix crash while accessing garbage pointer and freed memory
The failure in rereg_mr flow caused to set garbage value (error value)
into mr->umem pointer. This pointer is accessed at the release stage
and it causes to the following crash.
There is not enough to simply change umem to point to NULL, because the
MR struct is needed to be accessed during MR deregistration phase, so
delay kfree too.
Toshi Kani [Tue, 13 Mar 2018 17:03:46 +0000 (11:03 -0600)]
x86/mm: Fix vmalloc_fault to use pXd_large
Gratian Crisan reported that vmalloc_fault() crashes when CONFIG_HUGETLBFS
is not set since the function inadvertently uses pXn_huge(), which always
return 0 in this case. ioremap() does not depend on CONFIG_HUGETLBFS.
Tejun Heo [Wed, 14 Mar 2018 19:10:18 +0000 (12:10 -0700)]
RDMAVT: Fix synchronization around percpu_ref
rvt_mregion uses percpu_ref for reference counting and RCU to protect
accesses from lkey_table. When a rvt_mregion needs to be freed, it
first gets unregistered from lkey_table and then rvt_check_refs() is
called to wait for in-flight usages before the rvt_mregion is freed.
rvt_check_refs() seems to have a couple issues.
* It has a fast exit path which tests percpu_ref_is_zero(). However,
a percpu_ref reading zero doesn't mean that the object can be
released. In fact, the ->release() callback might not even have
started executing yet. Proceeding with freeing can lead to
use-after-free.
* lkey_table is RCU protected but there is no RCU grace period in the
free path. percpu_ref uses RCU internally but it's sched-RCU whose
grace periods are different from regular RCU. Also, it generally
isn't a good idea to depend on internal behaviors like this.
To address the above issues, this patch removes the fast exit and adds
an explicit synchronize_rcu().
Tejun Heo [Wed, 14 Mar 2018 19:10:17 +0000 (12:10 -0700)]
fs/aio: Use RCU accessors for kioctx_table->table[]
While converting ioctx index from a list to a table, db446a08c23d
("aio: convert the ioctx list to table lookup v3") missed tagging
kioctx_table->table[] as an array of RCU pointers and using the
appropriate RCU accessors. This introduces a small window in the
lookup path where init and access may race.
Mark kioctx_table->table[] with __rcu and use the approriate RCU
accessors when using the field.
Tejun Heo [Wed, 14 Mar 2018 19:10:17 +0000 (12:10 -0700)]
fs/aio: Add explicit RCU grace period when freeing kioctx
While fixing refcounting, e34ecee2ae79 ("aio: Fix a trinity splat")
incorrectly removed explicit RCU grace period before freeing kioctx.
The intention seems to be depending on the internal RCU grace periods
of percpu_ref; however, percpu_ref uses a different flavor of RCU,
sched-RCU. This can lead to kioctx being freed while RCU read
protected dereferences are still in progress.
Fix it by updating free_ioctx() to go through call_rcu() explicitly.
Marc Zyngier [Tue, 6 Mar 2018 21:44:37 +0000 (21:44 +0000)]
kvm: arm/arm64: vgic-v3: Tighten synchronization for guests using v2 on v3
On guest exit, and when using GICv2 on GICv3, we use a dsb(st) to
force synchronization between the memory-mapped guest view and
the system-register view that the hypervisor uses.
This is incorrect, as the spec calls out the need for "a DSB whose
required access type is both loads and stores with any Shareability
attribute", while we're only synchronizing stores.
We also lack an isb after the dsb to ensure that the latter has
actually been executed before we start reading stuff from the sysregs.
The fix is pretty easy: turn dsb(st) into dsb(sy), and slap an isb()
just after.
Marc Zyngier [Tue, 6 Mar 2018 21:48:01 +0000 (21:48 +0000)]
KVM: arm/arm64: vgic: Don't populate multiple LRs with the same vintid
The vgic code is trying to be clever when injecting GICv2 SGIs,
and will happily populate LRs with the same interrupt number if
they come from multiple vcpus (after all, they are distinct
interrupt sources).
Unfortunately, this is against the letter of the architecture,
and the GICv2 architecture spec says "Each valid interrupt stored
in the List registers must have a unique VirtualID for that
virtual CPU interface.". GICv3 has similar (although slightly
ambiguous) restrictions.
This results in guests locking up when using GICv2-on-GICv3, for
example. The obvious fix is to stop trying so hard, and inject
a single vcpu per SGI per guest entry. After all, pending SGIs
with multiple source vcpus are pretty rare, and are mostly seen
in scenario where the physical CPUs are severely overcomitted.
But as we now only inject a single instance of a multi-source SGI per
vcpu entry, we may delay those interrupts for longer than strictly
necessary, and run the risk of injecting lower priority interrupts
in the meantime.
In order to address this, we adopt a three stage strategy:
- If we encounter a multi-source SGI in the AP list while computing
its depth, we force the list to be sorted
- When populating the LRs, we prevent the injection of any interrupt
of lower priority than that of the first multi-source SGI we've
injected.
- Finally, the injection of a multi-source SGI triggers the request
of a maintenance interrupt when there will be no pending interrupt
in the LRs (HCR_NPIE).
At the point where the last pending interrupt in the LRs switches
from Pending to Active, the maintenance interrupt will be delivered,
allowing us to add the remaining SGIs using the same process.
The KVM IDMAP is a mapping of a statically allocated kernel structure,
and so printing its physical address leaks the physical placement of
the kernel when physical KASLR in effect. So change the kvm_info() to
kvm_debug() to remove it from the log output.
While at it, trim the output a bit more: IRQ numbers can be found in
/proc/interrupts, and the HYP VA and vgic-v2 lines are not highly
informational either.
We currently don't allow resetting mapped IRQs from userspace, because
their state is controlled by the hardware. But we do need to reset the
state when the VM is reset, so we provide a function for the 'owner' of
the mapped interrupt to reset the interrupt state.
Currently only the timer uses mapped interrupts, so we call this
function from the timer reset logic.
Cc: [email protected] Fixes: 4c60e360d6df ("KVM: arm/arm64: Provide a get_input_level for the arch timer") Signed-off-by: Christoffer Dall <[email protected]> Signed-off-by: Marc Zyngier <[email protected]>
Christoffer Dall [Tue, 27 Feb 2018 11:33:50 +0000 (12:33 +0100)]
KVM: arm/arm64: Avoid vcpu_load for other vcpu ioctls than KVM_RUN
Calling vcpu_load() registers preempt notifiers for this vcpu and calls
kvm_arch_vcpu_load(). The latter will soon be doing a lot of heavy
lifting on arm/arm64 and will try to do things such as enabling the
virtual timer and setting us up to handle interrupts from the timer
hardware.
Loading state onto hardware registers and enabling hardware to signal
interrupts can be problematic when we're not actually about to run the
VCPU, because it makes it difficult to establish the right context when
handling interrupts from the timer, and it makes the register access
code difficult to reason about.
Luckily, now when we call vcpu_load in each ioctl implementation, we can
simply remove the call from the non-KVM_RUN vcpu ioctls, and our
kvm_arch_vcpu_load() is only used for loading vcpu content to the
physical CPU when we're actually going to run the vcpu.
Andre Przywara [Tue, 6 Mar 2018 09:21:06 +0000 (09:21 +0000)]
KVM: arm/arm64: vgic: Add missing irq_lock to vgic_mmio_read_pending
Our irq_is_pending() helper function accesses multiple members of the
vgic_irq struct, so we need to hold the lock when calling it.
Add that requirement as a comment to the definition and take the lock
around the call in vgic_mmio_read_pending(), where we were missing it
before.
Fixes: 96b298000db4 ("KVM: arm/arm64: vgic-new: Add PENDING registers handlers") Signed-off-by: Andre Przywara <[email protected]> Signed-off-by: Marc Zyngier <[email protected]>
Darren Hart [Sun, 11 Mar 2018 00:12:16 +0000 (16:12 -0800)]
platform/x86: dell-smbios: Resolve dependency error on ACPI_WMI
Similarly to DCDBAS for DELL_SMBIOS_SMM, if DELL_SMBIOS_WMI is enabled,
DELL_SMBIOS becomes dependent on ACPI_WMI. Update the depends lines to
prevent a configuration where DELL_SMBIOS=y and either backend
dependency =m. Update the comment accordingly.
Cathy Zhou [Wed, 14 Mar 2018 17:56:07 +0000 (10:56 -0700)]
sunvnet: does not support GSO for sctp
The NETIF_F_GSO_SOFTWARE implies support for GSO on SCTP, but the
sunvnet driver does not support GSO for sctp. Here we remove the
NETIF_F_GSO_SOFTWARE feature flag and only report NETIF_F_ALL_TSO
instead.
David S. Miller [Wed, 14 Mar 2018 17:50:47 +0000 (13:50 -0400)]
Merge tag 'linux-can-fixes-for-4.16-20180314' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can
Marc Kleine-Budde says:
====================
pull-request: can 2018-03-14
this is a pull request of two patches for net/master.
Both patches are by Andri Yngvason and fix problems in the cc770 driver,
that show up quite fast on RT systems, but also on non RT setups.
====================
Jonathan Toppins [Wed, 14 Mar 2018 16:36:25 +0000 (12:36 -0400)]
tg3: prevent scheduling while atomic splat
The problem was introduced in commit 506b0a395f26 ("[netdrv] tg3: APE heartbeat changes"). The bug occurs
because tp->lock spinlock is held which is obtained in tg3_start
by way of tg3_full_lock(), line 11571. The documentation for usleep_range()
specifically states it cannot be used inside a spinlock.
Fixes: 506b0a395f26 ("[netdrv] tg3: APE heartbeat changes") Signed-off-by: Jonathan Toppins <[email protected]> Acked-by: Michael Chan <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Sabrina Dubroca [Wed, 14 Mar 2018 09:21:14 +0000 (10:21 +0100)]
ipv4: lock mtu in fnhe when received PMTU < net.ipv4.route.min_pmtu
Prior to the rework of PMTU information storage in commit 2c8cec5c10bc ("ipv4: Cache learned PMTU information in inetpeer."),
when a PMTU event advertising a PMTU smaller than
net.ipv4.route.min_pmtu was received, we would disable setting the DF
flag on packets by locking the MTU metric, and set the PMTU to
net.ipv4.route.min_pmtu.
Since then, we don't disable DF, and set PMTU to
net.ipv4.route.min_pmtu, so the intermediate router that has this link
with a small MTU will have to drop the packets.
This patch reestablishes pre-2.6.39 behavior by splitting
rtable->rt_pmtu into a bitfield with rt_mtu_locked and rt_pmtu.
rt_mtu_locked indicates that we shouldn't set the DF bit on that path,
and is checked in ip_dont_fragment().
One possible workaround is to set net.ipv4.route.min_pmtu to a value low
enough to accommodate the lowest MTU encountered.
Fixes: 2c8cec5c10bc ("ipv4: Cache learned PMTU information in inetpeer.") Signed-off-by: Sabrina Dubroca <[email protected]> Reviewed-by: Stefano Brivio <[email protected]> Signed-off-by: David S. Miller <[email protected]>
David S. Miller [Wed, 14 Mar 2018 17:16:23 +0000 (13:16 -0400)]
Merge branch 'DPAA-Ethernet-fixes'
Madalin Bucur says:
====================
DPAA Ethernet fixes
This patch set is addressing several issues in the DPAA Ethernet
driver suite:
- module unload crash caused by wrong reference to device being left
in the cleanup code after the DSA related changes
- scheduling wile atomic bug in QMan code revealed during dpaa_eth
module unload
- a couple of error counter fixes, a duplicated init in dpaa_eth.
====================
Madalin Bucur [Wed, 14 Mar 2018 13:37:29 +0000 (08:37 -0500)]
dpaa_eth: fix error in dpaa_remove()
The recent changes that make the driver probing compatible with DSA
were not propagated in the dpa_remove() function, breaking the
module unload function. Using the proper device to address the issue.
Madalin Bucur [Wed, 14 Mar 2018 13:37:28 +0000 (08:37 -0500)]
soc/fsl/qbman: fix issue in qman_delete_cgr_safe()
The wait_for_completion() call in qman_delete_cgr_safe()
was triggering a scheduling while atomic bug, replacing the
kthread with a smp_call_function_single() call to fix it.
Linus Torvalds [Wed, 14 Mar 2018 17:05:59 +0000 (10:05 -0700)]
Merge tag 'usb-4.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Pull USB fixes from Greg KH:
"Here are a small clump of USB fixes for 4.16-rc6.
Nothing major, just a number of fixes in lots of different drivers, as
well as a PHY driver fix that snuck into this tree. Full details are
in the shortlog.
All of these have been in linux-next with no reported issues"
* tag 'usb-4.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (22 commits)
usb: musb: Fix external abort in musb_remove on omap2430
phy: qcom-ufs: add MODULE_LICENSE tag
usb: typec: tcpm: fusb302: Do not log an error on -EPROBE_DEFER
USB: OHCI: Fix NULL dereference in HCDs using HCD_LOCAL_MEM
usbip: vudc: fix null pointer dereference on udc->lock
xhci: Fix front USB ports on ASUS PRIME B350M-A
usb: host: xhci-plat: revert "usb: host: xhci-plat: enable clk in resume timing"
usb: usbmon: Read text within supplied buffer size
usb: host: xhci-rcar: add support for r8a77965
USB: storage: Add JMicron bridge 152d:2567 to unusual_devs.h
usb: xhci: dbc: Fix lockdep warning
xhci: fix endpoint context tracer output
Revert "typec: tcpm: Only request matching pdos"
usb: musb: call pm_runtime_{get,put}_sync before reading vbus registers
usb: quirks: add control message delay for 1b1c:1b20
uas: fix comparison for error code
usb: gadget: udc: renesas_usb3: add binging for r8a77965
usb: renesas_usbhs: add binding for r8a77965
usb: dwc2: fix STM32F7 USB OTG HS compatible
dt-bindings: usb: fix the STM32F7 DWC2 OTG HS core binding
...
Linus Torvalds [Wed, 14 Mar 2018 17:02:40 +0000 (10:02 -0700)]
Merge tag 'tty-4.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Pull tty/serial driver fixes from Greg KH:
"Here are some small tty core and serial driver fixes for 4.16-rc6.
They resolve some newly reported bugs, as well as some very old ones,
which is always nice to see. There is also a new device id added in
here for good measure.
All of these have been in linux-next for a while with no reported
issues"
* tag 'tty-4.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
serial: imx: fix bogus dev_err
serial: sh-sci: prevent lockup on full TTY buffers
serial: 8250_pci: Add Brainboxes UC-260 4 port serial device
earlycon: add reg-offset to physical address before mapping
serial: core: mark port as initialized in autoconfig
serial: 8250_pci: Don't fail on multiport card class
tty/serial: atmel: add new version check for usart
tty: make n_tty_read() always abort if hangup is in progress
Linus Torvalds [Wed, 14 Mar 2018 16:59:45 +0000 (09:59 -0700)]
Merge tag 'staging-4.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging
Pull staging fixes from Greg KH:
"Here are three staging driver fixes for 4.16-rc6
Two of them are lockdep fixes for the ashmem driver that have been
reported by a number of people recently. The last one is a fix for the
comedi driver core.
All of these have been in linux-next for a while with no reported
issues"
* tag 'staging-4.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
staging: android: ashmem: Fix possible deadlock in ashmem_ioctl
staging: comedi: fix comedi_nsamples_left.
staging: android: ashmem: Fix lockdep issue during llseek
Eric Dumazet [Wed, 14 Mar 2018 16:04:16 +0000 (09:04 -0700)]
net: use skb_to_full_sk() in skb_update_prio()
Andrei Vagin reported a KASAN: slab-out-of-bounds error in
skb_update_prio()
Since SYNACK might be attached to a request socket, we need to
get back to the listener socket.
Since this listener is manipulated without locks, add const
qualifiers to sock_cgroup_prioidx() so that the const can also
be used in skb_update_prio()
Also add the const qualifier to sock_cgroup_classid() for consistency.
Fixes: ca6fb0651883 ("tcp: attach SYNACK messages to request sockets instead of listener") Signed-off-by: Eric Dumazet <[email protected]> Reported-by: Andrei Vagin <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Josh Poimboeuf [Wed, 14 Mar 2018 15:24:20 +0000 (10:24 -0500)]
jump_label: Fix sparc64 warning
The kbuild test robot reported the following warning on sparc64:
kernel/jump_label.c: In function '__jump_label_update':
kernel/jump_label.c:376:51: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
WARN_ONCE(1, "can't patch jump_label at %pS", (void *)entry->code);
On sparc64, the jump_label entry->code field is of type u32, but
pointers are 64-bit. Silence the warning by casting entry->code to an
unsigned long before casting it to a pointer. This is also what the
sparc jump label code does.
Andri Yngvason [Wed, 14 Mar 2018 11:52:57 +0000 (11:52 +0000)]
can: cc770: Fix queue stall & dropped RTR reply
While waiting for the TX object to send an RTR, an external message with a
matching id can overwrite the TX data. In this case we must call the rx
routine and then try transmitting the message that was overwritten again.
The queue was being stalled because the RX event did not generate an
interrupt to wake up the queue again and the TX event did not happen
because the TXRQST flag is reset by the chip when new data is received.
According to the CC770 datasheet the id of a message object should not be
changed while the MSGVAL bit is set. This has been fixed by resetting the
MSGVAL bit before modifying the object in the transmit function and setting
it after. It is not enough to set & reset CPUUPD.
It is important to keep the MSGVAL bit reset while the message object is
being modified. Otherwise, during RTR transmission, a frame with matching
id could trigger an rx-interrupt, which would cause a race condition
between the interrupt routine and the transmit function.
projects / linux.git / log