Linus Torvalds [Mon, 1 Jun 2020 18:30:28 +0000 (11:30 -0700)]
Merge tag 'tpmdd-next-20200522' of git://git.infradead.org/users/jjs/linux-tpmdd
Pull tpm updates from Jarkko Sakkinen.
* tag 'tpmdd-next-20200522' of git://git.infradead.org/users/jjs/linux-tpmdd:
tpm: eventlog: Replace zero-length array with flexible-array member
tpm/tpm_ftpm_tee: Use UUID API for exporting the UUID
Yes, staying withing 80 columns is certainly still _preferred_. But
it's not the hard limit that the checkpatch warnings imply, and other
concerns can most certainly dominate.
Increase the default limit to 100 characters. Not because 100
characters is some hard limit either, but that's certainly a "what are
you doing" kind of value and less likely to be about the occasional
slightly longer lines.
Miscellanea:
- to avoid unnecessary whitespace changes in files, checkpatch will no
longer emit a warning about line length when scanning files unless
--strict is also used
- Add a bit to coding-style about alignment to open parenthesis
Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Linus Torvalds [Sun, 31 May 2020 17:45:11 +0000 (10:45 -0700)]
Merge tag 'x86-urgent-2020-05-31' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Thomas Gleixner:
"A pile of x86 fixes:
- Prevent a memory leak in ioperm which was caused by the stupid
assumption that the exit cleanup is always called for current,
which is not the case when fork fails after taking a reference on
the ioperm bitmap.
- Fix an arithmething overflow in the DMA code on 32bit systems
- Fill gaps in the xstate copy with defaults instead of leaving them
uninitialized
- Revert: "Make __X32_SYSCALL_BIT be unsigned long" as it turned out
that existing user space fails to build"
* tag 'x86-urgent-2020-05-31' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/ioperm: Prevent a memory leak when fork fails
x86/dma: Fix max PFN arithmetic overflow on 32 bit systems
copy_xstate_to_kernel(): don't leave parts of destination uninitialized
x86/syscalls: Revert "x86/syscalls: Make __X32_SYSCALL_BIT be unsigned long"
14) Fix leak in inetdev_init(), from Yang Yingliang.
15) Don't try to use inet hash and unhash in l2tp code, results in
crashes. From Eric Dumazet"
* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (77 commits)
l2tp: add sk_family checks to l2tp_validate_socket
l2tp: do not use inet_hash()/inet_unhash()
net: qrtr: Allocate workqueue before kernel_bind
mptcp: remove msk from the token container at destruction time.
mptcp: fix race between MP_JOIN and close
mptcp: fix unblocking connect()
net/sched: act_ct: add nat mangle action only for NAT-conntrack
devinet: fix memleak in inetdev_init()
virtio_vsock: Fix race condition in virtio_transport_recv_pkt
drivers/net/ibmvnic: Update VNIC protocol version reporting
NFC: st21nfca: add missed kfree_skb() in an error path
neigh: fix ARP retransmit timer guard
bpf, selftests: Add a verifier test for assigning 32bit reg states to 64bit ones
bpf, selftests: Verifier bounds tests need to be updated
bpf: Fix a verifier issue when assigning 32bit reg states to 64bit ones
bpf: Fix use-after-free in fmod_ret check
net/mlx5e: replace EINVAL in mlx5e_flower_parse_meta()
net/mlx5e: Fix MLX5_TC_CT dependencies
net/mlx5e: Properly set default values when disabling adaptive moderation
net/mlx5e: Fix arch depending casting issue in FEC
...
Eric Dumazet [Fri, 29 May 2020 18:32:25 +0000 (11:32 -0700)]
l2tp: add sk_family checks to l2tp_validate_socket
syzbot was able to trigger a crash after using an ISDN socket
and fool l2tp.
Fix this by making sure the UDP socket is of the proper family.
BUG: KASAN: slab-out-of-bounds in setup_udp_tunnel_sock+0x465/0x540 net/ipv4/udp_tunnel.c:78
Write of size 1 at addr ffff88808ed0c590 by task syz-executor.5/3018
Memory state around the buggy address: ffff88808ed0c480: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ffff88808ed0c500: 00 00 00 fc fc fc fc fc fc fc fc fc fc fc fc fc
>ffff88808ed0c580: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
^ ffff88808ed0c600: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc ffff88808ed0c680: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
Fixes: 6b9f34239b00 ("l2tp: fix races in tunnel creation") Fixes: fd558d186df2 ("l2tp: Split pppol2tp patch into separate l2tp and ppp parts") Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: James Chapman <jchapman@katalix.com> Cc: Guillaume Nault <gnault@redhat.com> Reported-by: syzbot <syzkaller@googlegroups.com> Acked-by: Guillaume Nault <gnault@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Fri, 29 May 2020 18:20:53 +0000 (11:20 -0700)]
l2tp: do not use inet_hash()/inet_unhash()
syzbot recently found a way to crash the kernel [1]
Issue here is that inet_hash() & inet_unhash() are currently
only meant to be used by TCP & DCCP, since only these protocols
provide the needed hashinfo pointer.
L2TP uses a single list (instead of a hash table)
This old bug became an issue after commit 610236587600
("bpf: Add new cgroup attach type to enable sock modifications")
since after this commit, sk_common_release() can be called
while the L2TP socket is still considered 'hashed'.
Chris Lew [Thu, 28 May 2020 23:05:26 +0000 (16:05 -0700)]
net: qrtr: Allocate workqueue before kernel_bind
A null pointer dereference in qrtr_ns_data_ready() is seen if a client
opens a qrtr socket before qrtr_ns_init() can bind to the control port.
When the control port is bound, the ENETRESET error will be broadcasted
and clients will close their sockets. This results in DEL_CLIENT
packets being sent to the ns and qrtr_ns_data_ready() being called
without the workqueue being allocated.
Allocate the workqueue before setting sk_data_ready and binding to the
control port. This ensures that the work and workqueue structs are
allocated and initialized before qrtr_ns_data_ready can be called.
Fixes: 0c2204a4ad71 ("net: qrtr: Migrate nameservice to kernel from userspace") Signed-off-by: Chris Lew <clew@codeaurora.org> Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 31 May 2020 04:39:13 +0000 (21:39 -0700)]
Merge branch 'mptcp-a-bunch-of-fixes'
Paolo Abeni says:
====================
mptcp: a bunch of fixes
This patch series pulls together a few bugfixes for MPTCP bug observed while
doing stress-test with apache bench - forced to use MPTCP and multiple
subflows.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo Abeni [Fri, 29 May 2020 15:43:31 +0000 (17:43 +0200)]
mptcp: remove msk from the token container at destruction time.
Currently we remote the msk from the token container only
via mptcp_close(). The MPTCP master socket can be destroyed
also via other paths (e.g. if not yet accepted, when shutting
down the listener socket). When we hit the latter scenario,
dangling msk references are left into the token container,
leading to memory corruption and/or UaF.
This change addresses the issue by moving the token removal
into the msk destructor.
Fixes: 79c0949e9a09 ("mptcp: Add key generation and token tree") Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The MP_JOIN socket will be leaked. Additionally we can hit
UaF for the msk 'struct socket' referenced via the 'conn'
field.
This change try to address the issue introducing some
synchronization between the MP_JOIN 3whs and mptcp_close
via the join_list spinlock. If we detect the msk is closing
the MP_JOIN socket is closed, too.
Fixes: f296234c98a8 ("mptcp: Add handling of incoming MP_JOIN requests") Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo Abeni [Fri, 29 May 2020 15:43:29 +0000 (17:43 +0200)]
mptcp: fix unblocking connect()
Currently unblocking connect() on MPTCP sockets fails frequently.
If mptcp_stream_connect() is invoked to complete a previously
attempted unblocking connection, it will still try to create
the first subflow via __mptcp_socket_create(). If the 3whs is
completed and the 'can_ack' flag is already set, the latter
will fail with -EINVAL.
This change addresses the issue checking for pending connect and
delegating the completion to the first subflow. Additionally
do msk addresses and sk_state changes only when needed.
Fixes: 2303f994b3e1 ("mptcp: Associate MPTCP context with TCP socket") Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
wenxu [Sat, 30 May 2020 05:54:51 +0000 (13:54 +0800)]
net/sched: act_ct: add nat mangle action only for NAT-conntrack
Currently add nat mangle action with comparing invert and orig tuple.
It is better to check IPS_NAT_MASK flags first to avoid non necessary
memcmp for non-NAT conntrack.
Signed-off-by: wenxu <wenxu@ucloud.cn> Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Yang Yingliang [Sat, 30 May 2020 03:34:33 +0000 (11:34 +0800)]
devinet: fix memleak in inetdev_init()
When devinet_sysctl_register() failed, the memory allocated
in neigh_parms_alloc() should be freed.
Fixes: 20e61da7ffcf ("ipv4: fail early when creating netdev named all or default") Signed-off-by: Yang Yingliang <yangyingliang@huawei.com> Acked-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The race condition is as follows:
Task1 Task2
===== =====
__sock_release virtio_transport_recv_pkt
__vsock_release vsock_find_bound_socket (found sk)
lock_sock_nested
vsock_remove_sock
sock_orphan
sk_set_socket(sk, NULL)
sk->sk_shutdown = SHUTDOWN_MASK
...
release_sock
lock_sock
virtio_transport_recv_connecting
sk->sk_socket->state (panic!)
The root cause is that vsock_find_bound_socket can't hold the lock_sock,
so there is a small race window between vsock_find_bound_socket() and
lock_sock(). If __vsock_release() is running in another task,
sk->sk_socket will be set to NULL inadvertently.
This fixes it by checking sk->sk_shutdown(suggested by Stefano) after
lock_sock since sk->sk_shutdown is set to SHUTDOWN_MASK under the
protection of lock_sock_nested.
Signed-off-by: Jia He <justin.he@arm.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Linus Torvalds [Sat, 30 May 2020 19:28:44 +0000 (12:28 -0700)]
Merge tag 'powerpc-5.7-6' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux
Pull powerpc fixes from Michael Ellerman:
- a fix for the recent change to how we restore non-volatile GPRs,
which broke our emulation of reading from the DSCR (Data Stream
Control Register).
- a fix for the recent rewrite of interrupt/syscall exit in C, we need
to exclude KCOV from that code, otherwise it can lead to
unrecoverable faults.
Thanks to Daniel Axtens.
* tag 'powerpc-5.7-6' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
powerpc/64s: Disable sanitisers for C syscall/interrupt entry/exit code
powerpc/64s: Fix restore of NV GPRs after facility unavailable exception
Linus Torvalds [Sat, 30 May 2020 19:26:21 +0000 (12:26 -0700)]
Merge tag 'gpio-v5.7-3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio
Pull GPIO fixes from Linus Walleij:
"Here are some (very) late fixes for GPIO, none of them very serious
except the one tagged for stable for enabling IRQ on open drain lines:
- Fix probing of mvebu chips without PWM
- Fix error path on ida_get_simple() on the exar driver
- Notify userspace properly about line status changes when flags are
changed on lines.
- Fix a sleeping while holding spinlock in the mellanox driver.
- Fix return value of the PXA and Kona probe calls.
- Fix IRQ locking of open drain lines, it is fine to have IRQs on
open drain lines flagged for output"
* tag 'gpio-v5.7-3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:
gpio: fix locking open drain IRQ lines
gpio: bcm-kona: Fix return value of bcm_kona_gpio_probe()
gpio: pxa: Fix return value of pxa_gpio_probe()
gpio: mlxbf2: Fix sleeping while holding spinlock
gpiolib: notify user-space about line status changes after flags are set
gpio: exar: Fix bad handling for ida_simple_get error path
gpio: mvebu: Fix probing for chips without PWM
Thomas Falcon [Thu, 28 May 2020 16:19:17 +0000 (11:19 -0500)]
drivers/net/ibmvnic: Update VNIC protocol version reporting
VNIC protocol version is reported in big-endian format, but it
is not byteswapped before logging. Fix that, and remove version
comparison as only one protocol version exists at this time.
Signed-off-by: Thomas Falcon <tlfalcon@linux.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Chuhong Yuan [Thu, 28 May 2020 10:20:37 +0000 (18:20 +0800)]
NFC: st21nfca: add missed kfree_skb() in an error path
st21nfca_tm_send_atr_res() misses to call kfree_skb() in an error path.
Add the missed function call to fix it.
Fixes: 1892bf844ea0 ("NFC: st21nfca: Adding P2P support to st21nfca in Initiator & Target mode") Signed-off-by: Chuhong Yuan <hslester96@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Hangbin Liu [Thu, 28 May 2020 07:15:13 +0000 (15:15 +0800)]
neigh: fix ARP retransmit timer guard
In commit 19e16d220f0a ("neigh: support smaller retrans_time settting")
we add more accurate control for ARP and NS. But for ARP I forgot to
update the latest guard in neigh_timer_handler(), then the next
retransmit would be reset to jiffies + HZ/2 if we set the retrans_time
less than 500ms. Fix it by setting the time_before() check to HZ/100.
IPv6 does not have this issue.
Reported-by: Jianwen Ji <jiji@redhat.com> Fixes: 19e16d220f0a ("neigh: support smaller retrans_time settting") Signed-off-by: Hangbin Liu <liuhangbin@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 29 May 2020 23:31:22 +0000 (16:31 -0700)]
Merge tag 'mlx5-fixes-2020-05-28' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5 fixes 2020-05-28
This series introduces some fixes to mlx5 driver.
v1->v2:
- Fix bad sha1, Jakub.
- Added one more patch by Pablo.
net/mlx5e: replace EINVAL in mlx5e_flower_parse_meta()
Nothing major, the only patch worth mentioning is the suspend/resume crash
fix by adding the missing pci device handlers, the fix is very straight
forward and as Dexuan already expressed, the patch is important for Azure
users to avoid crash on VM hibernation, patch is marked for -stable v4.6
below.
Conflict note:
('net/mlx5e: Fix MLX5_TC_CT dependencies') has a trivial one line conflict
with current net-next, which can be resolved by simply using the line from
net-next.
Please pull and let me know if there is any problem.
For -stable v4.6
('net/mlx5: Fix crash upon suspend/resume')
For -stable v5.6
('net/mlx5e: replace EINVAL in mlx5e_flower_parse_meta()')
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Linus Torvalds [Fri, 29 May 2020 23:10:07 +0000 (16:10 -0700)]
Merge tag 'armsoc-fixes-v5.7' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Pull ARM SoC fixes from Arnd Bergmann:
"This time there is one fix for the error path in the mediatek cmdq
driver (used by their video driver) and a couple of devicetree fixes,
mostly for 32-bit ARM, and fairly harmless:
- On OMAP2 there were a few regressions in the ethernet drivers, one
of them leading to an external abort trap
- One Raspberry Pi version had a misconfigured LED
- Interrupts on Broadcom NSP were slightly misconfigured
- One i.MX6q board had issues with graphics mode setting
- On mmp3 there are some minor fixes that were submitted for v5.8
with a cc:stable tag, so I ended up picking them up here as well
- The Mediatek Video Codec needs to run at a higher frequency than
configured originally"
* tag 'armsoc-fixes-v5.7' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc:
ARM: dts: mmp3: Drop usb-nop-xceiv from HSIC phy
ARM: dts: mmp3-dell-ariel: Fix the SPI devices
ARM: dts: mmp3: Use the MMP3 compatible string for /clocks
ARM: dts: bcm: HR2: Fix PPI interrupt types
ARM: dts: bcm2835-rpi-zero-w: Fix led polarity
ARM: dts/imx6q-bx50v3: Set display interface clock parents
soc: mediatek: cmdq: return send msg error code
arm64: dts: mt8173: fix vcodec-enc clock
ARM: dts: Fix wrong mdio clock for dm814x
ARM: dts: am437x: fix networking on boards with ksz9031 phy
ARM: dts: am57xx: fix networking on boards with ksz9031 phy
Linus Torvalds [Fri, 29 May 2020 20:59:54 +0000 (13:59 -0700)]
Merge tag 'ceph-for-5.7-rc8' of git://github.com/ceph/ceph-client
Pull ceph fixes from Ilya Dryomov:
"Cache tiering and cap handling fixups, both marked for stable"
* tag 'ceph-for-5.7-rc8' of git://github.com/ceph/ceph-client:
ceph: flush release queue when handling caps for unknown inode
libceph: ignore pool overlay and cache logic on redirects
Linus Torvalds [Fri, 29 May 2020 20:41:33 +0000 (13:41 -0700)]
Merge tag 'iommu-fixes-v5.7-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu
Pull iommu fixes from Joerg Roedel:
- Two build fixes for issues introduced during the merge window
- A fix for a reference count leak in an error path of
iommu_group_alloc()
* tag 'iommu-fixes-v5.7-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
iommu: Fix reference count leak in iommu_group_alloc.
x86: Hide the archdata.iommu field behind generic IOMMU_API
ia64: Hide the archdata.iommu field behind generic IOMMU_API
Linus Torvalds [Fri, 29 May 2020 20:39:26 +0000 (13:39 -0700)]
Merge tag 'block-5.7-2020-05-29' of git://git.kernel.dk/linux-block
Pull block fixes from Jens Axboe:
"Two small fixes:
- Revert a block change that mixed up the return values for non-mq
devices
- NVMe poll race fix"
* tag 'block-5.7-2020-05-29' of git://git.kernel.dk/linux-block:
Revert "block: end bio with BLK_STS_AGAIN in case of non-mq devs and REQ_NOWAIT"
nvme-pci: avoid race between nvme_reap_pending_cqes() and nvme_poll()
Linus Torvalds [Fri, 29 May 2020 20:35:45 +0000 (13:35 -0700)]
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
Pull rdma fixes from Jason Gunthorpe:
"Nothing profound here, just a last set of long standing bug fixes:
- Incorrect error unwind in qib and pvrdma
- User triggerable NULL pointer crash in mlx5 with ODP prefetch
- syzkaller RCU race in uverbs
- Rare double free crash in ipoib"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
IB/ipoib: Fix double free of skb in case of multicast traffic in CM mode
RDMA/core: Fix double destruction of uobject
RDMA/pvrdma: Fix missing pci disable in pvrdma_pci_probe()
RDMA/mlx5: Fix NULL pointer dereference in destroy_prefetch_work
IB/qib: Call kobject_put() when kobject_init_and_add() fails
After the fix at insn 10 because we have 's32_min_value < 0' the following
step 11 now has 'smax_value=U32_MAX' where before we pulled the s32_max_value
bound into the smax_value as seen above in 11 with smax_value=2147483647.
In the original step 14 'smin_value=72057594021150720' this trips the logic
in the verifier function check_reg_sane_offset(),
if (smin >= BPF_MAX_VAR_OFF || smin <= -BPF_MAX_VAR_OFF) {
verbose(env, "value %lld makes %s pointer be out of bounds\n",
smin, reg_type_str[type]);
return false;
}
Specifically, the 'smin <= -BPF_MAX_VAR_OFF' check. But with the fix
at step 14 we have bounds 'smin_value=0' so the above check is not tripped
because BPF_MAX_VAR_OFF=1<<29.
We have a smin_value=0 here because at step 10 the smaller smin_value=0 means
the subtractions at steps 11 and 12 bring the smin_value negative.
Then the shift clears the top bit and smin_value is set to 0. Note we still
have the smax_value in the fixed code so any reads will fail. An alternative
would be to have reg_sane_check() do both smin and smax value tests.
To fix the test we can omit the 'r1 >>=8' at line 13. This will change the
err string, but keeps the intention of the test as suggseted by the title,
"check after truncation of boundary-crossing range". If the verifier logic
changes a different value is likely to be thrown in the error or the error
will no longer be thrown forcing this test to be examined. With this change
we see the new state at step 13.
Giving the expected out of bounds error, "value -4294967168 makes map_value
pointer be out of bounds" However, for unpriv case we see a different error
now because of the mixed signed bounds pointer arithmatic. This seems OK so
I've only added the unpriv_errstr for this. Another optino may have been to
do addition on r1 instead of subtraction but I favor the approach above
slightly.
In real test_verif_scale1 code, "w9 = 0" and "w2 = w9" are in
different basic blocks.
In the above, after "r6 += r2", r6 becomes a scalar, which eventually
caused the memory access error. The correct register state should be
a pkt pointer.
The inprecise register state starts at "w2 = w9".
The 32bit register w9 is 0, in __reg_assign_32_into_64(),
the 64bit reg->smax_value is assigned to be U32_MAX.
The 64bit reg->smin_value is 0 and the 64bit register
itself remains constant based on reg->var_off.
In adjust_ptr_min_max_vals(), the verifier checks for a known constant,
smin_val must be equal to smax_val. Since they are not equal,
the verifier decides r6 is a unknown scalar, which caused later failure.
The llvm10 does not have this issue as it generates different code:
w9 = 0 /* R9_w=inv0 */
r8 = *(u32 *)(r1 + 80) /* __sk_buff->data_end */
r7 = *(u32 *)(r1 + 76) /* __sk_buff->data */
......
r6 = r7 /* R6_w=pkt(id=0,off=0,r=0,imm=0) */
r6 += r9 /* R6_w=pkt(id=0,off=0,r=0,imm=0) */
r3 = r6 /* R3_w=pkt(id=0,off=0,r=0,imm=0) */
r3 += 14 /* R3_w=pkt(id=0,off=14,r=0,imm=0) */
if r3 > r8 goto end
...
To fix the above issue, we can include zero in the test condition for
assigning the s32_max_value and s32_min_value to their 64-bit equivalents
smax_value and smin_value.
Further, fix the condition to avoid doing zero extension bounds checks
when s32_min_value <= 0. This could allow for the case where bounds
32-bit bounds (-1,1) get incorrectly translated to (0,1) 64-bit bounds.
When in-fact the -1 min value needs to force U32_MAX bound.
Linus Torvalds [Fri, 29 May 2020 20:31:01 +0000 (13:31 -0700)]
Merge tag 'sound-5.7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Pull sound fixes from Takashi Iwai:
"Only a few last-minute small fixes: the change in ALSA core hwdep is
about the undefined behavior of bit shift, which is almost harmless
but still worth to pick up quickly.
The rest are all device-specific fixes for HD-audio and USB-audio, and
safe to apply at the late stage"
* tag 'sound-5.7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
ALSA: hda/realtek - Add new codec supported for ALC287
ALSA: usb-audio: Quirks for Gigabyte TRX40 Aorus Master onboard audio
ALSA: usb-audio: mixer: volume quirk for ESS Technology Asus USB DAC
ALSA: hda/realtek - Add a model for Thinkpad T570 without DAC workaround
ALSA: hwdep: fix a left shifting 1 by 31 UB bug
Linus Torvalds [Fri, 29 May 2020 20:29:20 +0000 (13:29 -0700)]
Merge tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux
Pull clk fixes from Stephen Boyd:
"Two fixes for the new SM8150 and SM8250 Qualcomm clk drivers to fix a
randconfig build error and an incorrect parent mapping"
* tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux:
clk: qcom: gcc: Fix parent for gpll0_out_even
clk: qcom: sm8250 gcc depends on QCOM_GDSC
prog->aux->trampoline = tr should be set only when prog is valid.
Otherwise prog freeing will try to put trampoline via prog->aux->trampoline,
but it may not point to a valid trampoline.
net/mlx5e: replace EINVAL in mlx5e_flower_parse_meta()
The drivers reports EINVAL to userspace through netlink on invalid meta
match. This is confusing since EINVAL is usually reserved for malformed
netlink messages. Replace it by more meaningful codes.
Vlad Buslov [Mon, 25 May 2020 13:57:51 +0000 (16:57 +0300)]
net/mlx5e: Fix MLX5_TC_CT dependencies
Change MLX5_TC_CT config dependencies to include MLX5_ESWITCH instead of
MLX5_CORE_EN && NET_SWITCHDEV, which are already required by MLX5_ESWITCH.
Without this change mlx5 fails to compile if user disables MLX5_ESWITCH
without also manually disabling MLX5_TC_CT.
Aya Levin [Mon, 13 Apr 2020 08:31:00 +0000 (11:31 +0300)]
net/mlx5e: Fix arch depending casting issue in FEC
Change type of active_fec to u32 to match the type expected by
mlx5e_get_fec_mode. Copy active_fec and configured_fec values to
unsigned long before preforming bitwise manipulations.
Take the same approach when configuring FEC over 50G link modes: copy
the policy into an unsigned long and only than preform bitwise
operations.
Maor Dickman [Sun, 24 May 2020 06:45:44 +0000 (09:45 +0300)]
net/mlx5e: Remove warning "devices are not on same switch HW"
On tunnel decap rule insertion, the indirect mechanism will attempt to
offload the rule on all uplink representors which will trigger the
"devices are not on same switch HW, can't offload forwarding" message
for the uplink which isn't on the same switch HW as the VF representor.
The above flow is valid and shouldn't cause warning message,
fix by removing the warning and only report this flow using extack.
1) Several fixes for ESP gro/gso in transport and beet mode when
IPv6 extension headers are present. From Xin Long.
2) Fix a wrong comment on XFRMA_OFFLOAD_DEV.
From Antony Antony.
3) Fix sk_destruct callback handling on ESP in TCP encapsulation.
From Sabrina Dubroca.
4) Fix a use after free in xfrm_output_gso when used with vxlan.
From Xin Long.
5) Fix secpath handling of VTI when used wiuth IPCOMP.
From Xin Long.
6) Fix an oops when deleting a x-netns xfrm interface.
From Nicolas Dichtel.
7) Fix a possible warning on policy updates. We had a case where it was
possible to add two policies with the same lookup keys.
From Xin Long.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
* tag 'drm-fixes-2020-05-29-1' of git://anongit.freedesktop.org/drm/drm:
drm/amd/display: Fix potential integer wraparound resulting in a hang
drm/amd/display: drop cursor position check in atomic test
gpu/drm: Ingenic: Fix opaque pointer casted to wrong type
gpu/drm: ingenic: Fix bogus crtc_atomic_check callback
Fix several issues in the previous gfs2_find_jhead fix:
* When updating @blocks_submitted, @block refers to the first block block not
submitted yet, not the last block submitted, so fix an off-by-one error.
* We want to ensure that @blocks_submitted is far enough ahead of @blocks_read
to guarantee that there is in-flight I/O. Otherwise, we'll eventually end up
waiting for pages that haven't been submitted, yet.
* It's much easier to compare the number of blocks added with the number of
blocks submitted to limit the maximum bio size.
* Even with bio chaining, we can keep adding blocks until we reach the maximum
bio size, as long as we stop at a page boundary. This simplifies the logic.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com> Reviewed-by: Bob Peterson <rpeterso@redhat.com>
Helge Deller [Thu, 28 May 2020 20:29:25 +0000 (22:29 +0200)]
parisc: Fix kernel panic in mem_init()
The Debian kernel v5.6 triggers this kernel panic:
Kernel panic - not syncing: Bad Address (null pointer deref?)
Bad Address (null pointer deref?): Code=26 (Data memory access rights trap) at addr 0000000000000000
CPU: 0 PID: 0 Comm: swapper Not tainted 5.6.0-2-parisc64 #1 Debian 5.6.14-1
IAOQ[0]: mem_init+0xb0/0x150
IAOQ[1]: mem_init+0xb4/0x150
RP(r2): start_kernel+0x6c8/0x1190
Backtrace:
[<0000000040101ab4>] start_kernel+0x6c8/0x1190
[<0000000040108574>] start_parisc+0x158/0x1b8
on a HP-PARISC rp3440 machine with this memory layout:
Memory Ranges:
0) Start 0x0000000000000000 End 0x000000003fffffff Size 1024 MB
1) Start 0x0000004040000000 End 0x00000040ffdfffff Size 3070 MB
Fix the crash by avoiding virt_to_page() and similar functions in
mem_init() until the memory zones have been fully set up.
Qiushi Wu [Wed, 27 May 2020 21:00:19 +0000 (16:00 -0500)]
iommu: Fix reference count leak in iommu_group_alloc.
kobject_init_and_add() takes reference even when it fails.
Thus, when kobject_init_and_add() returns an error,
kobject_put() must be called to properly clean up the kobject.
Linus Walleij [Wed, 27 May 2020 14:07:58 +0000 (16:07 +0200)]
gpio: fix locking open drain IRQ lines
We provided the right semantics on open drain lines being
by definition output but incidentally the irq set up function
would only allow IRQs on lines that were "not output".
Fix the semantics to allow output open drain lines to be used
for IRQs.
Reported-by: Hans Verkuil <hverkuil@xs4all.nl> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl> Tested-by: Hans Verkuil <hverkuil@xs4all.nl> Cc: Russell King <linux@armlinux.org.uk> Cc: stable@vger.kernel.org # v5.3+ Link: https://lore.kernel.org/r/20200527140758.162280-1-linus.walleij@linaro.org Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
This is caused by __sanitizer_cov_trace_pc() causing an SLB fault
after MSR[RI] has been cleared by __hard_EE_RI_disable(), which we
can not recover from.
Do not instrument the new syscall/interrupt entry/exit code with KCOV,
GCOV or UBSAN.
Reported-by: syzbot-ppc64 <ozlabsyz@au1.ibm.com> Fixes: 68b34588e202 ("powerpc/64/sycall: Implement syscall entry/exit logic in C") Signed-off-by: Daniel Axtens <dja@axtens.net> Acked-by: Andrew Donnellan <ajd@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This occurred when sending a v4 skb over vxlan6 over ipsec, in which case
skb->protocol == htons(ETH_P_IPV6) while skb->sk->sk_family == AF_INET in
xfrm_local_error(). Then it will go to xfrm6_local_error() where it tries
to get ipv6 info from a ipv4 sk.
This issue was actually fixed by Commit 628e341f319f ("xfrm: make local
error reporting more robust"), but brought back by Commit 844d48746e4b
("xfrm: choose protocol family by skb protocol").
So to fix it, we should call xfrm6_local_error() only when skb->protocol
is htons(ETH_P_IPV6) and skb->sk->sk_family is AF_INET6.
Fixes: 844d48746e4b ("xfrm: choose protocol family by skb protocol") Reported-by: Xiumei Mu <xmu@redhat.com> Signed-off-by: Xin Long <lucien.xin@gmail.com> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Eric Dumazet [Thu, 28 May 2020 21:57:47 +0000 (14:57 -0700)]
net: be more gentle about silly gso requests coming from user
Recent change in virtio_net_hdr_to_skb() broke some packetdrill tests.
When --mss=XXX option is set, packetdrill always provide gso_type & gso_size
for its inbound packets, regardless of packet size.
if (packet->tcp && packet->mss) {
if (packet->ipv4)
gso.gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
else
gso.gso_type = VIRTIO_NET_HDR_GSO_TCPV6;
gso.gso_size = packet->mss;
}
Since many other programs could do the same, relax virtio_net_hdr_to_skb()
to no longer return an error, but instead ignore gso settings.
This keeps Willem intent to make sure no malicious packet could
reach gso stack.
Note that TCP stack has a special logic in tcp_set_skb_tso_segs()
to clear gso_size for small packets.
Fixes: 6dd912f82680 ("net: check untrusted gso_size at kernel entry") Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Willem de Bruijn <willemb@google.com> Acked-by: Willem de Bruijn <willemb@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jonas Falkevik [Wed, 27 May 2020 09:56:40 +0000 (11:56 +0200)]
sctp: check assoc before SCTP_ADDR_{MADE_PRIM, ADDED} event
Make sure SCTP_ADDR_{MADE_PRIM,ADDED} are sent only for associations
that have been established.
These events are described in rfc6458#section-6.1
SCTP_PEER_ADDR_CHANGE:
This tag indicates that an address that is
part of an existing association has experienced a change of
state (e.g., a failure or return to service of the reachability
of an endpoint via a specific transport address).
Signed-off-by: Jonas Falkevik <jonas.falkevik@gmail.com> Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> Reviewed-by: Xin Long <lucien.xin@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jay Lang [Sun, 24 May 2020 16:27:39 +0000 (12:27 -0400)]
x86/ioperm: Prevent a memory leak when fork fails
In the copy_process() routine called by _do_fork(), failure to allocate
a PID (or further along in the function) will trigger an invocation to
exit_thread(). This is done to clean up from an earlier call to
copy_thread_tls(). Naturally, the child task is passed into exit_thread(),
however during the process, io_bitmap_exit() nullifies the parent's
io_bitmap rather than the child's.
As copy_thread_tls() has been called ahead of the failure, the reference
count on the calling thread's io_bitmap is incremented as we would expect.
However, io_bitmap_exit() doesn't accept any arguments, and thus assumes
it should trash the current thread's io_bitmap reference rather than the
child's. This is pretty sneaky in practice, because in all instances but
this one, exit_thread() is called with respect to the current task and
everything works out.
A determined attacker can issue an appropriate ioctl (i.e. KDENABIO) to
get a bitmap allocated, and force a clone3() syscall to fail by passing
in a zeroed clone_args structure. The kernel handles the erroneous struct
and the buggy code path is followed, and even though the parent's reference
to the io_bitmap is trashed, the child still holds a reference and thus
the structure will never be freed.
Fix this by tweaking io_bitmap_exit() and its subroutines to accept a
task_struct argument which to operate on.
Fixes: ea5f1cd7ab49 ("x86/ioperm: Remove bitmap if all permissions dropped") Signed-off-by: Jay Lang <jaytlang@mit.edu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable#@vger.kernel.org Link: https://lkml.kernel.org/r/20200524162742.253727-1-jaytlang@mit.edu
io_uring does do the right thing for this case, and we're still returning
-EAGAIN to userspace for the cases we don't support. Revert this change
to avoid doing endless spins of resubmits.
fs/binfmt_elf.c: allocate initialized memory in fill_thread_core_info()
KMSAN reported uninitialized data being written to disk when dumping
core. As a result, several kilobytes of kmalloc memory may be written
to the core file and then read by a non-privileged user.
Reported-by: sam <sunhaoyl@outlook.com> Signed-off-by: Alexander Potapenko <glider@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Kees Cook <keescook@chromium.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Alexey Dobriyan <adobriyan@gmail.com> Cc: <stable@vger.kernel.org> Link: http://lkml.kernel.org/r/20200419100848.63472-1-glider@google.com Link: https://github.com/google/kmsan/issues/76 Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm: remove VM_BUG_ON(PageSlab()) from page_mapcount()
Replace superfluous VM_BUG_ON() with comment about correct usage.
Technically reverts commit 1d148e218a0d ("mm: add VM_BUG_ON_PAGE() to
page_mapcount()"), but context lines have changed.
Function isolate_migratepages_block() runs some checks out of lru_lock
when choose pages for migration. After checking PageLRU() it checks
extra page references by comparing page_count() and page_mapcount().
Between these two checks page could be removed from lru, freed and taken
by slab.
As a result this race triggers VM_BUG_ON(PageSlab()) in page_mapcount().
Race window is tiny. For certain workload this happens around once a
year.
The code in isolate_migratepages_block() was added in commit 119d6d59dcc0 ("mm, compaction: avoid isolating pinned pages") before
adding VM_BUG_ON into page_mapcount().
This race has been predicted in 2015 by Vlastimil Babka (see link
below).
Hugh Dickins [Thu, 28 May 2020 05:20:43 +0000 (22:20 -0700)]
mm,thp: stop leaking unreleased file pages
When collapse_file() calls try_to_release_page(), it has already isolated
the page: so if releasing buffers happens to fail (as it sometimes does),
remember to putback_lru_page(): otherwise that page is left unreclaimable
and unfreeable, and the file extent uncollapsible.
Fixes: 99cb0dbd47a1 ("mm,thp: add read-only THP support for (non-shmem) FS") Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Song Liu <songliubraving@fb.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Rik van Riel <riel@surriel.com> Cc: <stable@vger.kernel.org> [5.4+] Link: http://lkml.kernel.org/r/alpine.LSU.2.11.2005231837500.1766@eggly.anvils Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Qian Cai [Thu, 28 May 2020 05:20:40 +0000 (22:20 -0700)]
mm/z3fold: silence kmemleak false positives of slots
Kmemleak reported many leaks while under memory pressue in,
slots = alloc_slots(pool, gfp);
which is referenced by "zhdr" in init_z3fold_page(),
zhdr->slots = slots;
However, "zhdr" could be gone without freeing slots as the later will be
freed separately when the last "handle" off of "handles" array is freed.
It will be within "slots" which is always aligned.
Alexander Dahl [Tue, 26 May 2020 17:57:49 +0000 (19:57 +0200)]
x86/dma: Fix max PFN arithmetic overflow on 32 bit systems
The intermediate result of the old term (4UL * 1024 * 1024 * 1024) is
4 294 967 296 or 0x100000000 which is no problem on 64 bit systems.
The patch does not change the later overall result of 0x100000 for
MAX_DMA32_PFN (after it has been shifted by PAGE_SHIFT). The new
calculation yields the same result, but does not require 64 bit
arithmetic.
On 32 bit systems the old calculation suffers from an arithmetic
overflow in that intermediate term in braces: 4UL aka unsigned long int
is 4 byte wide and an arithmetic overflow happens (the 0x100000000 does
not fit in 4 bytes), the in braces result is truncated to zero, the
following right shift does not alter that, so MAX_DMA32_PFN evaluates to
0 on 32 bit systems.
That wrong value is a problem in a comparision against MAX_DMA32_PFN in
the init code for swiotlb in pci_swiotlb_detect_4gb() to decide if
swiotlb should be active. That comparison yields the opposite result,
when compiling on 32 bit systems.
This was not possible before
1b7e03ef7570 ("x86, NUMA: Enable emulation on 32bit too")
when that MAX_DMA32_PFN was first made visible to x86_32 (and which
landed in v3.0).
In practice this wasn't a problem, unless CONFIG_SWIOTLB is active on
x86-32.
However if one has set CONFIG_IOMMU_INTEL, since
c5a5dc4cbbf4 ("iommu/vt-d: Don't switch off swiotlb if bounce page is used")
there's a dependency on CONFIG_SWIOTLB, which was not necessarily
active before. That landed in v5.4, where we noticed it in the fli4l
Linux distribution. We have CONFIG_IOMMU_INTEL active on both 32 and 64
bit kernel configs there (I could not find out why, so let's just say
historical reasons).
The effect is at boot time 64 MiB (default size) were allocated for
bounce buffers now, which is a noticeable amount of memory on small
systems like pcengines ALIX 2D3 with 256 MiB memory, which are still
frequently used as home routers.
We noticed this effect when migrating from kernel v4.19 (LTS) to v5.4
(LTS) in fli4l and got that kernel messages for example:
Linux version 5.4.22 (buildroot@buildroot) (gcc version 7.3.0 (Buildroot 2018.02.8)) #1 SMP Mon Nov 26 23:40:00 CET 2018
…
Memory: 183484K/261756K available (4594K kernel code, 393K rwdata, 1660K rodata, 536K init, 456K bss , 78272K reserved, 0K cma-reserved, 0K highmem)
…
PCI-DMA: Using software bounce buffering for IO (SWIOTLB)
software IO TLB: mapped [mem 0x0bb78000-0x0fb78000] (64MB)
The initial analysis and the suggested fix was done by user 'sourcejedi'
at stackoverflow and explicitly marked as GPLv2 for inclusion in the
Linux kernel:
https://unix.stackexchange.com/a/520525/50007
The new calculation, which does not suffer from that overflow, is the
same as for arch/mips now as suggested by Robin Murphy.
The fix was tested by fli4l users on round about two dozen different
systems, including both 32 and 64 bit archs, bare metal and virtualized
machines.
Qiushi Wu [Thu, 28 May 2020 03:10:29 +0000 (22:10 -0500)]
bonding: Fix reference count leak in bond_sysfs_slave_add.
kobject_init_and_add() takes reference even when it fails.
If this function returns an error, kobject_put() must be called to
properly clean up the memory associated with the object. Previous
commit "b8eb718348b8" fixed a similar problem.
Fixes: 07699f9a7c8d ("bonding: add sysfs /slave dir for bond slave devices.") Signed-off-by: Qiushi Wu <wu000273@umn.edu> Acked-by: Jay Vosburgh <jay.vosburgh@canonical.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Petr Mladek [Wed, 27 May 2020 12:28:44 +0000 (14:28 +0200)]
powerpc/bpf: Enable bpf_probe_read{, str}() on powerpc again
The commit 0ebeea8ca8a4d1d453a ("bpf: Restrict bpf_probe_read{, str}() only
to archs where they work") caused that bpf_probe_read{, str}() functions
were not longer available on architectures where the same logical address
might have different content in kernel and user memory mapping. These
architectures should use probe_read_{user,kernel}_str helpers.
For backward compatibility, the problematic functions are still available
on architectures where the user and kernel address spaces are not
overlapping. This is defined CONFIG_ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE.
At the moment, these backward compatible functions are enabled only on x86_64,
arm, and arm64. Let's do it also on powerpc that has the non overlapping
address space as well.
Fixes: 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") Signed-off-by: Petr Mladek <pmladek@suse.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/lkml/20200527122844.19524-1-pmladek@suse.com
arm64/kernel: Fix return value when cpu_online() fails in __cpu_up()
If boot_secondary() was successful, and cpu_online() was an error in
__cpu_up(), -EIO was returned, but 0 is returned by commit d22b115cbfbb7
("arm64/kernel: Simplify __cpu_up() by bailing out early").
Therefore, bringup_wait_for_ap() causes the primary core to wait for a
long time, which may cause boot failure.
This commit sets -EIO to return code under the same conditions.
Fixes: d22b115cbfbb ("arm64/kernel: Simplify __cpu_up() by bailing out early") Signed-off-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp> Tested-by: Yuji Ishikawa <yuji2.ishikawa@toshiba.co.jp> Acked-by: Will Deacon <will@kernel.org> Cc: Gavin Shan <gshan@redhat.com> Cc: Mark Rutland <mark.rutland@arm.com> Link: https://lore.kernel.org/r/20200527233457.2531118-1-nobuhiro1.iwamatsu@toshiba.co.jp
[catalin.marinas@arm.com: return -EIO at the end of the function] Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Guo Ren [Sun, 24 May 2020 12:14:11 +0000 (12:14 +0000)]
csky: Coding convention in entry.S
There is no fixup or feature in the patch, we only cleanup with:
- Remove unnecessary reg used (r11, r12), just use r9 & r10 &
syscallid regs as temp useage.
- Add _TIF_SYSCALL_WORK and _TIF_WORK_MASK to gather macros.
The preempt_count of reg could be destroyed after csky_do_IRQ without reload
from memory.
After reference to other architectures (arm64, riscv), we move preempt entry
into ret_from_exception and disable irq at the beginning of
ret_from_exception instead of RESTORE_ALL.
Signed-off-by: Guo Ren <guoren@linux.alibaba.com> Reported-by: Lu Baoquan <lu.baoquan@intellif.com>
Valentine Fatiev [Wed, 27 May 2020 13:47:05 +0000 (16:47 +0300)]
IB/ipoib: Fix double free of skb in case of multicast traffic in CM mode
When connected mode is set, and we have connected and datagram traffic in
parallel, ipoib might crash with double free of datagram skb.
The current mechanism assumes that the order in the completion queue is
the same as the order of sent packets for all QPs. Order is kept only for
specific QP, in case of mixed UD and CM traffic we have few QPs (one UD and
few CM's) in parallel.
The problem:
----------------------------------------------------------
Transmit queue:
-----------------
UD skb pointer kept in queue itself, CM skb kept in spearate queue and
uses transmit queue as a placeholder to count the number of total
transmitted packets.
1. CM1 'wc' processing
- skb freed in cm separate ring.
- tx_tail of transmit queue increased although UD2 is not freed.
Now driver assumes UD2 index is already freed and it could be used for
new transmitted skb.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 .........127
------------------------------------------------------------
NL NL UD2 CM1 ud3 cm2 cm3 ud4 cm4 ud5 NL NL NL ...........
------------------------------------------------------------
^ ^ ^
(Bad)tail head
(Bad - Could be used for new SKB)
In this case (due to heavy load) UD2 skb pointer could be replaced by new
transmitted packet UD_NEW, as the driver assumes its free. At this point
we will have to process two 'wc' with same index but we have only one
pointer to free.
During second attempt to free the same skb we will have NULL pointer
exception.
2. UD2 'wc' processing
- skb freed according the index we got from 'wc', but it was already
overwritten by mistake. So actually the skb that was released is the
skb of the new transmitted packet and not the original one.
The fix:
-----------------------------------------------------------------------
The fix is to stop using the UD ring as a placeholder for CM packets, the
cyclic ring variables tx_head and tx_tail will manage the UD tx_ring, a
new cyclic variables global_tx_head and global_tx_tail are introduced for
managing and counting the overall outstanding sent packets, then the send
queue will be stopped and waken based on these variables only.
Note that no locking is needed since global_tx_head is updated in the xmit
flow and global_tx_tail is updated in the NAPI flow only. A previous
attempt tried to use one variable to count the outstanding sent packets,
but it did not work since xmit and NAPI flows can run at the same time and
the counter will be updated wrongly. Thus, we use the same simple cyclic
head and tail scheme that we have today for the UD tx_ring.
Fixes: 2c104ea68350 ("IB/ipoib: Get rid of the tx_outstanding variable in all modes") Link: https://lore.kernel.org/r/20200527134705.480068-1-leon@kernel.org Signed-off-by: Valentine Fatiev <valentinef@mellanox.com> Signed-off-by: Alaa Hleihel <alaa@mellanox.com> Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Acked-by: Doug Ledford <dledford@redhat.com> Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Simon Ser [Sat, 23 May 2020 11:53:41 +0000 (11:53 +0000)]
drm/amd/display: drop cursor position check in atomic test
get_cursor_position already handles the case where the cursor has
negative off-screen coordinates by not setting
dc_cursor_position.enabled.
Signed-off-by: Simon Ser <contact@emersion.fr> Fixes: 626bf90fe03f ("drm/amd/display: add basic atomic check for cursor plane") Cc: Alex Deucher <alexander.deucher@amd.com> Cc: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org
Vladimir Oltean [Wed, 27 May 2020 18:08:05 +0000 (21:08 +0300)]
net: dsa: declare lockless TX feature for slave ports
Be there a platform with the following layout:
Regular NIC
|
+----> DSA master for switch port
|
+----> DSA master for another switch port
After changing DSA back to static lockdep class keys in commit 1a33e10e4a95 ("net: partially revert dynamic lockdep key changes"), this
kernel splat can be seen:
Lockdep keys themselves were added in commit ab92d68fc22f ("net: core:
add generic lockdep keys"), and it's very likely that this splat existed
since then, but I have no real way to check, since this stacked platform
wasn't supported by mainline back then.
>From Taehee's own words:
This patch was considered that all stackable devices have LLTX flag.
But the dsa doesn't have LLTX, so this splat happened.
After this patch, dsa shares the same lockdep class key.
On the nested dsa interface architecture, which you illustrated,
the same lockdep class key will be used in __dev_queue_xmit() because
dsa doesn't have LLTX.
So that lockdep detects deadlock because the same lockdep class key is
used recursively although actually the different locks are used.
There are some ways to fix this problem.
1. using NETIF_F_LLTX flag.
If possible, using the LLTX flag is a very clear way for it.
But I'm so sorry I don't know whether the dsa could have LLTX or not.
2. using dynamic lockdep again.
It means that each interface uses a separate lockdep class key.
So, lockdep will not detect recursive locking.
But this way has a problem that it could consume lockdep class key
too many.
Currently, lockdep can have 8192 lockdep class keys.
- you can see this number with the following command.
cat /proc/lockdep_stats
lock-classes: 1251 [max: 8192]
...
The [max: 8192] means that the maximum number of lockdep class keys.
If too many lockdep class keys are registered, lockdep stops to work.
So, using a dynamic(separated) lockdep class key should be considered
carefully.
In addition, updating lockdep class key routine might have to be existing.
(lockdep_register_key(), lockdep_set_class(), lockdep_unregister_key())
3. Using lockdep subclass.
A lockdep class key could have 8 subclasses.
The different subclass is considered different locks by lockdep
infrastructure.
But "lock-classes" is not counted by subclasses.
So, it could avoid stopping lockdep infrastructure by an overflow of
lockdep class keys.
This approach should also have an updating lockdep class key routine.
(lockdep_set_subclass())
4. Using nonvalidate lockdep class key.
The lockdep infrastructure supports nonvalidate lockdep class key type.
It means this lockdep is not validated by lockdep infrastructure.
So, the splat will not happen but lockdep couldn't detect real deadlock
case because lockdep really doesn't validate it.
I think this should be used for really special cases.
(lockdep_set_novalidate_class())
Further discussion here:
https://patchwork.ozlabs.org/project/netdev/patch/20200503052220.4536-2-xiyou.wangcong@gmail.com/
There appears to be no negative side-effect to declaring lockless TX for
the DSA virtual interfaces, which means they handle their own locking.
So that's what we do to make the splat go away.
Patch tested in a wide variety of cases: unicast, multicast, PTP, etc.
Fixes: ab92d68fc22f ("net: core: add generic lockdep keys") Suggested-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Wed, 27 May 2020 16:48:03 +0000 (19:48 +0300)]
net: dsa: felix: send VLANs on CPU port as egress-tagged
As explained in other commits before (b9cd75e66895 and 87b0f983f66f),
ocelot switches have a single egress-untagged VLAN per port, and the
driver would deny adding a second one while an egress-untagged VLAN
already exists.
But on the CPU port (where the VLAN configuration is implicit, because
there is no net device for the bridge to control), the DSA core attempts
to add a VLAN using the same flags as were used for the front-panel
port. This would make adding any untagged VLAN fail due to the CPU port
rejecting the configuration:
bridge vlan add dev swp0 vid 100 pvid untagged
[ 1865.854253] mscc_felix 0000:00:00.5: Port already has a native VLAN: 1
[ 1865.860824] mscc_felix 0000:00:00.5: Failed to add VLAN 100 to port 5: -16
(note that port 5 is the CPU port and not the front-panel swp0).
So this hardware will send all VLANs as tagged towards the CPU.
Fixes: 56051948773e ("net: dsa: ocelot: add driver for Felix switch family") Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Dongli Zhang [Wed, 27 May 2020 16:13:52 +0000 (09:13 -0700)]
nvme-pci: avoid race between nvme_reap_pending_cqes() and nvme_poll()
There may be a race between nvme_reap_pending_cqes() and nvme_poll(), e.g.,
when doing live reset while polling the nvme device.
CPU X CPU Y
nvme_poll()
nvme_dev_disable()
-> nvme_stop_queues()
-> nvme_suspend_io_queues()
-> nvme_suspend_queue()
-> spin_lock(&nvmeq->cq_poll_lock);
-> nvme_reap_pending_cqes()
-> nvme_process_cq() -> nvme_process_cq()
In the above scenario, the nvme_process_cq() for the same queue may be
running on both CPU X and CPU Y concurrently.
It is much more easier to reproduce the issue when CONFIG_PREEMPT is
enabled in kernel. When CONFIG_PREEMPT is disabled, it would take longer
time for nvme_stop_queues()-->blk_mq_quiesce_queue() to wait for grace
period.
This patch protects nvme_process_cq() with nvmeq->cq_poll_lock in
nvme_reap_pending_cqes().
Fixes: fa46c6fb5d61 ("nvme/pci: move cqe check after device shutdown") Signed-off-by: Dongli Zhang <dongli.zhang@oracle.com> Reviewed-by: Ming Lei <ming.lei@redhat.com> Reviewed-by: Keith Busch <kbusch@kernel.org> Signed-off-by: Christoph Hellwig <hch@lst.de>
The accept(2) is an "input" socket interface, so we should use
SO_RCVTIMEO instead of SO_SNDTIMEO to set the timeout.
So this patch replace sock_sndtimeo() with sock_rcvtimeo() to
use the right timeout in the vsock_accept().
Fixes: d021c344051a ("VSOCK: Introduce VM Sockets") Signed-off-by: Stefano Garzarella <sgarzare@redhat.com> Reviewed-by: Jorgen Hansen <jhansen@vmware.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Heinrich Kuhn [Wed, 27 May 2020 07:44:20 +0000 (09:44 +0200)]
nfp: flower: fix used time of merge flow statistics
Prior to this change the correct value for the used counter is calculated
but not stored nor, therefore, propagated to user-space. In use-cases such
as OVS use-case at least this results in active flows being removed from
the hardware datapath. Which results in both unnecessary flow tear-down
and setup, and packet processing on the host.
This patch addresses the problem by saving the calculated used value
which allows the value to propagate to user-space.
Found by inspection.
Fixes: aa6ce2ea0c93 ("nfp: flower: support stats update for merge flows") Signed-off-by: Heinrich Kuhn <heinrich.kuhn@netronome.com> Signed-off-by: Simon Horman <simon.horman@netronome.com> Signed-off-by: David S. Miller <davem@davemloft.net>
we can't accept 65536 as a valid number for 'nflows', because the loop on
'idx' in fq_pie_init() will never end. The extack message is correct, but
it doesn't say that 0 is not a valid number for 'flows': while at it, fix
this also. Add a tdc selftest to check correct validation of 'flows'.
CC: Ivan Vecera <ivecera@redhat.com> Fixes: ec97ecf1ebe4 ("net: sched: add Flow Queue PIE packet scheduler") Signed-off-by: Davide Caratti <dcaratti@redhat.com> Reviewed-by: Ivan Vecera <ivecera@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Linus Torvalds [Wed, 27 May 2020 18:03:24 +0000 (11:03 -0700)]
Merge tag 'fsnotify_for_v5.7-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs
Pull fanotify FAN_DIR_MODIFY disabling from Jan Kara:
"A single patch that disables FAN_DIR_MODIFY support that was merged in
this merge window.
When discussing further functionality we realized it may be more
logical to guard it with a feature flag or to call things slightly
differently (or maybe not) so let's not set the API in stone for now."
* tag 'fsnotify_for_v5.7-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
fanotify: turn off support for FAN_DIR_MODIFY
Linus Torvalds [Wed, 27 May 2020 17:58:19 +0000 (10:58 -0700)]
Merge branch 'for-5.7-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup
Pull cgroup fixes from Tejun Heo:
- Reverted stricter synchronization for cgroup recursive stats which
was prepping it for event counter usage which never got merged. The
change was causing performation regressions in some cases.
- Restore bpf-based device-cgroup operation even when cgroup1 device
cgroup is disabled.
Jason Gunthorpe [Wed, 27 May 2020 13:55:34 +0000 (16:55 +0300)]
RDMA/core: Fix double destruction of uobject
Fix use after free when user user space request uobject concurrently for
the same object, within the RCU grace period.
In that case, remove_handle_idr_uobject() is called twice and we will have
an extra put on the uobject which cause use after free. Fix it by leaving
the uobject write locked after it was removed from the idr.
Call to rdma_lookup_put_uobject with UVERBS_LOOKUP_DESTROY instead of
UVERBS_LOOKUP_WRITE will do the work.
Fixes: 7452a3c745a2 ("IB/uverbs: Allow RDMA_REMOVE_DESTROY to work concurrently with disassociate") Link: https://lore.kernel.org/r/20200527135534.482279-1-leon@kernel.org Signed-off-by: Maor Gottlieb <maorg@mellanox.com> Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Amir Goldstein [Wed, 27 May 2020 12:54:55 +0000 (15:54 +0300)]
fanotify: turn off support for FAN_DIR_MODIFY
FAN_DIR_MODIFY has been enabled by commit 44d705b0370b ("fanotify:
report name info for FAN_DIR_MODIFY event") in 5.7-rc1. Now we are
planning further extensions to the fanotify API and during that we
realized that FAN_DIR_MODIFY may behave slightly differently to be more
consistent with extensions we plan. So until we finalize these
extensions, let's not bind our hands with exposing FAN_DIR_MODIFY to
userland.
Signed-off-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
Linus Torvalds [Wed, 27 May 2020 16:53:25 +0000 (09:53 -0700)]
Merge branch 'exec-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace
Pull execve fix from Eric Biederman:
"While working on my exec cleanups I found a bug in exec that winds up
miscomputing the ambient credentials during exec. Andy appears to have
to been confused as to why credentials are computed for both the
script and the interpreter
From the original patch description:
[3] Linux very confusingly processes both the script and the
interpreter if applicable, for reasons that elude me. The results
from thinking about a script's file capabilities and/or setuid
bits are mostly discarded.
The only value in struct cred that gets changed in cap_bprm_set_creds
that I could find that might persist between the script and the
interpreter was cap_ambient. Which is fixed with this trivial change"
* 'exec-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace:
exec: Always set cap_ambient in cap_bprm_set_creds
x86: Hide the archdata.iommu field behind generic IOMMU_API
There is a generic, kernel wide configuration symbol for enabling the
IOMMU specific bits: CONFIG_IOMMU_API. Implementations (including
INTEL_IOMMU and AMD_IOMMU driver) select it so use it here as well.
This makes the conditional archdata.iommu field consistent with other
platforms and also fixes any compile test builds of other IOMMU drivers,
when INTEL_IOMMU or AMD_IOMMU are not selected).
For the case when INTEL_IOMMU/AMD_IOMMU and COMPILE_TEST are not
selected, this should create functionally equivalent code/choice. With
COMPILE_TEST this field could appear if other IOMMU drivers are chosen
but neither INTEL_IOMMU nor AMD_IOMMU are not.
Reported-by: kbuild test robot <lkp@intel.com> Fixes: e93a1695d7fb ("iommu: Enable compile testing for some of drivers") Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org> Acked-by: Borislav Petkov <bp@suse.de> Link: https://lore.kernel.org/r/20200518120855.27822-2-krzk@kernel.org Signed-off-by: Joerg Roedel <jroedel@suse.de>
ia64: Hide the archdata.iommu field behind generic IOMMU_API
There is a generic, kernel wide configuration symbol for enabling the
IOMMU specific bits: CONFIG_IOMMU_API. Implementations (including
INTEL_IOMMU driver) select it so use it here as well.
This makes the conditional archdata.iommu field consistent with other
platforms and also fixes any compile test builds of other IOMMU drivers,
when INTEL_IOMMU is not selected).
For the case when INTEL_IOMMU and COMPILE_TEST are not selected, this
should create functionally equivalent code/choice. With COMPILE_TEST
this field could appear if other IOMMU drivers are chosen but
INTEL_IOMMU not.
Reported-by: kbuild test robot <lkp@intel.com> Fixes: e93a1695d7fb ("iommu: Enable compile testing for some of drivers") Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org> Link: https://lore.kernel.org/r/20200518120855.27822-1-krzk@kernel.org Signed-off-by: Joerg Roedel <jroedel@suse.de>
netfilter: nf_conntrack_pptp: fix compilation warning with W=1 build
>> include/linux/netfilter/nf_conntrack_pptp.h:13:20: warning: 'const' type qualifier on return type has no effect [-Wignored-qualifiers]
extern const char *const pptp_msg_name(u_int16_t msg);
^~~~~~
Reported-by: kbuild test robot <lkp@intel.com> Fixes: 4c559f15efcc ("netfilter: nf_conntrack_pptp: prevent buffer overflows in debug code") Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netfilter: conntrack: comparison of unsigned in cthelper confirmation
net/netfilter/nf_conntrack_core.c: In function nf_confirm_cthelper:
net/netfilter/nf_conntrack_core.c:2117:15: warning: comparison of unsigned expression in < 0 is always false [-Wtype-limits]
2117 | if (protoff < 0 || (frag_off & htons(~0x7)) != 0)
| ^
ipv6_skip_exthdr() returns a signed integer.
Reported-by: Colin Ian King <colin.king@canonical.com> Fixes: 703acd70f249 ("netfilter: nfnetlink_cthelper: unbreak userspace helper support") Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netfilter: conntrack: Pass value of ctinfo to __nf_conntrack_update
Clang warns:
net/netfilter/nf_conntrack_core.c:2068:21: warning: variable 'ctinfo' is
uninitialized when used here [-Wuninitialized]
nf_ct_set(skb, ct, ctinfo);
^~~~~~
net/netfilter/nf_conntrack_core.c:2024:2: note: variable 'ctinfo' is
declared here
enum ip_conntrack_info ctinfo;
^
1 warning generated.
nf_conntrack_update was split up into nf_conntrack_update and
__nf_conntrack_update, where the assignment of ctinfo is in
nf_conntrack_update but it is used in __nf_conntrack_update.
Pass the value of ctinfo from nf_conntrack_update to
__nf_conntrack_update so that uninitialized memory is not used
and everything works properly.
Fixes: ee04805ff54a ("netfilter: conntrack: make conntrack userspace helpers work again") Link: https://github.com/ClangBuiltLinux/linux/issues/1039 Signed-off-by: Nathan Chancellor <natechancellor@gmail.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>