A revert of a recent change to the PTE bits for 32-bit BookS, which broke swap.
And a "fix" to disable STRICT_KERNEL_RWX for 64-bit in Kconfig, as it's causing
crashes for some people.
Thanks to:
Christophe Leroy, Rui Salvaterra.
-----BEGIN PGP SIGNATURE-----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=yBu+
-----END PGP SIGNATURE-----
Merge tag 'powerpc-5.7-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux
Pull powerpc fixes from Michael Ellerman:
- a revert of a recent change to the PTE bits for 32-bit BookS, which
broke swap.
- a "fix" to disable STRICT_KERNEL_RWX for 64-bit in Kconfig, as it's
causing crashes for some people.
Thanks to Christophe Leroy and Rui Salvaterra.
* tag 'powerpc-5.7-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
powerpc/64s: Disable STRICT_KERNEL_RWX
Revert "powerpc/32s: reorder Linux PTE bits to better match Hash PTE bits."
DMA transfers to and from the SD card stall for 10 seconds and run into
timeout on RTS5260 card readers after ASPM was enabled.
Adding a short msleep after disabling ASPM fixes the issue on several
Dell Precision 7530/7540 systems I tested.
This function is only called when waking up after the chip went into
power-save after not transferring data for a few seconds. The added
msleep does therefore not change anything in data transfer speed or
induce any excessive waiting while data transfers are running, or the
chip is sleeping. Only the transition from sleep to active is affected.
Signed-off-by: Klaus Doth <kdlnx@doth.eu>
Cc: stable <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/4434eaa7-2ee3-a560-faee-6cee63ebd6d4@doth.eu
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When attaching a flow dissector program to a network namespace with
bpf(BPF_PROG_ATTACH, ...) we grab a reference to bpf_prog.
If netns gets destroyed while a flow dissector is still attached, and there
are no other references to the prog, we leak the reference and the program
remains loaded.
Leak can be reproduced by running flow dissector tests from selftests/bpf:
# bpftool prog list
# ./test_flow_dissector.sh
...
selftests: test_flow_dissector [PASS]
# bpftool prog list
4: flow_dissector name _dissect tag e314084d332a5338 gpl
loaded_at 2020-05-20T18:50:53+0200 uid 0
xlated 552B jited 355B memlock 4096B map_ids 3,4
btf_id 4
#
Fix it by detaching the flow dissector program when netns is going away.
Fixes: d58e468b11 ("flow_dissector: implements flow dissector BPF hook")
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Stanislav Fomichev <sdf@google.com>
Link: https://lore.kernel.org/bpf/20200521083435.560256-1-jakub@cloudflare.com
In the function devm_platform_ioremap_resource(), if get resource
failed, the return value is ERR_PTR() not NULL. Thus it must be
replaced by IS_ERR(), or else it may result in crashes if a critical
error path is encountered.
Fixes: 0ce5ebd24d ("mfd: ioc3: Add driver for SGI IOC3 chip")
Signed-off-by: Zhang Shengju <zhangshengju@cmss.chinamobile.com>
Signed-off-by: Tang Bin <tangbin@cmss.chinamobile.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In case we can't find a ->dumpit callback for the requested
(family,type) pair, we fall back to (PF_UNSPEC,type). In effect, we're
in the same situation as if userspace had requested a PF_UNSPEC
dump. For RTM_GETROUTE, that handler is rtnl_dump_all, which calls all
the registered RTM_GETROUTE handlers.
The requested table id may or may not exist for all of those
families. commit ae677bbb44 ("net: Don't return invalid table id
error when dumping all families") fixed the problem when userspace
explicitly requests a PF_UNSPEC dump, but missed the fallback case.
For example, when we pass ipv6.disable=1 to a kernel with
CONFIG_IP_MROUTE=y and CONFIG_IP_MROUTE_MULTIPLE_TABLES=y,
the (PF_INET6, RTM_GETROUTE) handler isn't registered, so we end up in
rtnl_dump_all, and listing IPv6 routes will unexpectedly print:
# ip -6 r
Error: ipv4: MR table does not exist.
Dump terminated
commit ae677bbb44 introduced the dump_all_families variable, which
gets set when userspace requests a PF_UNSPEC dump. However, we can't
simply set the family to PF_UNSPEC in rtnetlink_rcv_msg in the
fallback case to get dump_all_families == true, because some messages
types (for example RTM_GETRULE and RTM_GETNEIGH) only register the
PF_UNSPEC handler and use the family to filter in the kernel what is
dumped to userspace. We would then export more entries, that userspace
would have to filter. iproute does that, but other programs may not.
Instead, this patch removes dump_all_families and updates the
RTM_GETROUTE handlers to check if the family that is being dumped is
their own. When it's not, which covers both the intentional PF_UNSPEC
dumps (as dump_all_families did) and the fallback case, ignore the
missing table id error.
Fixes: cb167893f4 ("net: Plumb support for filtering ipv4 and ipv6 multicast route dumps")
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In case of error with MPLS support the code is misusing AF_INET
instead of AF_MPLS.
Fixes: 1b69e7e6c4 ("ipip: support MPLS over IPv4")
Signed-off-by: Vadim Fedorenko <vfedorenko@novek.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
Vadim Fedorenko says:
====================
net/tls: fix encryption error path
The problem with data stream corruption was found in KTLS
transmit path with small socket send buffers and large
amount of data. bpf_exec_tx_verdict() frees open record
on any type of error including EAGAIN, ENOMEM and ENOSPC
while callers are able to recover this transient errors.
Also wrong error code was returned to user space in that
case. This patchset fixes the problems.
====================
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
We cannot free record on any transient error because it leads to
losing previos data. Check socket error to know whether record must
be freed or not.
Fixes: d10523d0b3 ("net/tls: free the record on encryption error")
Signed-off-by: Vadim Fedorenko <vfedorenko@novek.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
bpf_exec_tx_verdict() can return negative value for copied
variable. In that case this value will be pushed back to caller
and the real error code will be lost. Fix it using signed type and
checking for positive value.
Fixes: d10523d0b3 ("net/tls: free the record on encryption error")
Fixes: d3b18ad31f ("tls: add bpf support to sk_msg handling")
Signed-off-by: Vadim Fedorenko <vfedorenko@novek.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
Wei Yongjun says:
====================
net: ethernet: ti: fix some return value check
This patchset convert cpsw_ale_create() to return PTR_ERR() only, and
changed all the caller to check IS_ERR() instead of NULL.
Since v2:
1) rebased on net.git, as Jakub's suggest
2) split am65-cpsw-nuss.c changes, as Grygorii's suggest
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Convert to using IS_ERR() instead of NULL test for cpsw_ale_create()
error handling. Also fix to return negative error code from this error
handling case instead of 0 in.
Fixes: 93a7653031 ("net: ethernet: ti: introduce am65x/j721e gigabit eth subsystem driver")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
cpsw_ale_create() can return both NULL and PTR_ERR(), but all of
the caller only check NULL for error handling. This patch convert
it to only return PTR_ERR() in all error cases, and the caller using
IS_ERR() instead of NULL test.
Fixes: 4b41d34367 ("net: ethernet: ti: cpsw: allow untagged traffic on host port")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Once the traversal of the list is completed with list_for_each_entry(),
the iterator (node) will point to an invalid object. So passing this to
qrtr_local_enqueue() which is outside of the iterator block is erroneous
eventhough the object is not used.
So fix this by passing NULL to qrtr_local_enqueue().
Fixes: bdabad3e36 ("net: Add Qualcomm IPC router")
Reported-by: kbuild test robot <lkp@intel.com>
Reported-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
As ethnl_request_ops::reply_size handlers do not include common header
size into calculated/estimated reply size, it needs to be added in
ethnl_default_doit() and ethnl_default_notify() before allocating the
message. On the other hand, strset_reply_size() should not add common
header size.
Fixes: 728480f124 ("ethtool: default handlers for GET requests")
Reported-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
As Ubuntu and Fedora release new version used kernel version equal to or
higher than v5.4, They started to support kernel exfat filesystem.
Linus reported a mount error with new version of exfat on Fedora:
exfat: Unknown parameter 'namecase'
This is because there is a difference in mount option between old
staging/exfat and new exfat. And utf8, debug, and codepage options as
well as namecase have been removed from new exfat.
This patch add the dummy mount options as deprecated option to be
backward compatible with old one.
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In the implementation of aa_audit_rule_init(), when aa_label_parse()
fails the allocated memory for rule is released using
aa_audit_rule_free(). But after this release, the return statement
tries to access the label field of the rule which results in
use-after-free. Before releasing the rule, copy errNo and return it
after release.
Fixes: 52e8c38001 ("apparmor: Fix memory leak of rule on error exit path")
Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
policy_update() invokes begin_current_label_crit_section(), which
returns a reference of the updated aa_label object to "label" with
increased refcount.
When policy_update() returns, "label" becomes invalid, so the refcount
should be decreased to keep refcount balanced.
The reference counting issue happens in one exception handling path of
policy_update(). When aa_may_manage_policy() returns not NULL, the
refcnt increased by begin_current_label_crit_section() is not decreased,
causing a refcnt leak.
Fix this issue by jumping to "end_section" label when
aa_may_manage_policy() returns not NULL.
Fixes: 5ac8c355ae ("apparmor: allow introspecting the loaded policy pre internal transform")
Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn>
Signed-off-by: Xin Tan <tanxin.ctf@gmail.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
aa_change_profile() invokes aa_get_current_label(), which returns
a reference of the current task's label.
According to the comment of aa_get_current_label(), the returned
reference must be put with aa_put_label().
However, when the original object pointed by "label" becomes
unreachable because aa_change_profile() returns or a new object
is assigned to "label", reference count increased by
aa_get_current_label() is not decreased, causing a refcnt leak.
Fix this by calling aa_put_label() before aa_change_profile() return
and dropping unnecessary aa_get_current_label().
Fixes: 9fcf78cca1 ("apparmor: update domain transitions that are subsets of confinement at nnp")
Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn>
Signed-off-by: Xin Tan <tanxin.ctf@gmail.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
The Intel kernel build robot recently pointed out that I missed the
register keyword on this one when I refactored the code to remove local
register variables (which aren't supported by LLVM). GCC's manual
indicates that global register variables must have the register keyword,
As far as I can tell lacking the register keyword causes GCC to ignore
the __asm__ and treat this as a regular variable, but I'm not sure how
that didn't show up as some sort of failure.
Fixes: 52e7c52d2d ("RISC-V: Stop relying on GCC's register allocator's hueristics")
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
Fix a couple of build warnings.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
-----BEGIN PGP SIGNATURE-----
iQFDBAABCAAtFiEEXQn9CHHI+FuUyooNKB8NuNKNVGkFAl6+tzkPHG1zdEByZWRo
YXQuY29tAAoJECgfDbjSjVRpV5QH/jx6Jj16Hzwy6YV9caV4QeySWgZrI3y8fWTK
YlKdzmBE3YNJDwdV6EM5lT6hmJNGf392cF8akGk339IemiYJaHPFLt409ubLvfhZ
ejo0zY7NStOd2DZJfPQdissME7bgiLRpNDvaXRofJwZ87yK7nSNbPWVYp0Jz0Rie
BFnx5XOSqyTkOovylHZajHfodl5eHtdAOYI1+6SZH6gA1YKrhdDqB0gdyejXg4EQ
Ijg0oiDovU/bLfvaF+8jZZJvNsy8mouFidF5NJhCzBewQwx49tl2tLVNOQP/PwGF
Yf8DN7zH8yw+hsUbruj5lFKILvY7Rn2RXhp7ikTaUPYnSKUcYgo=
=uvKy
-----END PGP SIGNATURE-----
Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Pull virtio fixes from Michael Tsirkin:
"Fix a couple of build warnings"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
vhost: missing __user tags
vdpasim: remove unused variable 'ret'
Zoned block device specification do not define the behavior of
discard/trim command as this command is generally replaced by the reset
write pointer (zone reset) command. Emulate this in null_blk by making
zoned and discard options mutually exclusive.
Suggested-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Several strange crashes have been eventually traced back to
STRICT_KERNEL_RWX and its interaction with code patching.
Various paths in our ftrace, kprobes and other patching code need to
be hardened against patching failures, otherwise we can end up running
with partially/incorrectly patched ftrace paths, kprobes or jump
labels, which can then cause strange crashes.
Although fixes for those are in development, they're not -rc material.
There also seem to be problems with the underlying strict RWX logic,
which needs further debugging.
So for now disable STRICT_KERNEL_RWX on 64-bit to prevent people from
enabling the option and tripping over the bugs.
Fixes: 1e0fc9d1eb ("powerpc/Kconfig: Enable STRICT_KERNEL_RWX for some configs")
Cc: stable@vger.kernel.org # v4.13+
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200520133605.972649-1-mpe@ellerman.id.au
In the function kobject_cleanup(), kobject_del(kobj) is
called before the kobj->release(). That makes it possible to
release the parent of the kobject before the kobject itself.
To fix that, adding function __kboject_del() that does
everything that kobject_del() does except release the parent
reference. kobject_cleanup() then calls __kobject_del()
instead of kobject_del(), and separately decrements the
reference count of the parent kobject after kobj->release()
has been called.
Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Reported-by: kernel test robot <rong.a.chen@intel.com>
Fixes: 7589238a8c ("Revert "software node: Simplify software_node_release() function"")
Suggested-by: "Rafael J. Wysocki" <rafael@kernel.org>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Brendan Higgins <brendanhiggins@google.com>
Tested-by: Brendan Higgins <brendanhiggins@google.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Link: https://lore.kernel.org/r/20200513151840.36400-1-heikki.krogerus@linux.intel.com
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 21c27f0658 ("driver core: Fix SYNC_STATE_ONLY device link
implementation") didn't completely fix STATELESS + SYNC_STATE_ONLY
handling.
What looks like an optimization in that commit is actually a bug that
causes an if condition to always take the else path. This prevents
reordering of devices in the dpm_list when a DL_FLAG_STATELESS device
link is create on top of an existing DL_FLAG_SYNC_STATE_ONLY device
link.
Fixes: 21c27f0658 ("driver core: Fix SYNC_STATE_ONLY device link implementation")
Signed-off-by: Saravana Kannan <saravanak@google.com>
Cc: stable <stable@vger.kernel.org>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20200520043626.181820-1-saravanak@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fixes data remnant seen when we fail to reserve space for a
nexthop group during a larger dump.
If we fail the reservation, we goto nla_put_failure and
cancel the message.
Reproduce with the following iproute2 commands:
=====================
ip link add dummy1 type dummy
ip link add dummy2 type dummy
ip link add dummy3 type dummy
ip link add dummy4 type dummy
ip link add dummy5 type dummy
ip link add dummy6 type dummy
ip link add dummy7 type dummy
ip link add dummy8 type dummy
ip link add dummy9 type dummy
ip link add dummy10 type dummy
ip link add dummy11 type dummy
ip link add dummy12 type dummy
ip link add dummy13 type dummy
ip link add dummy14 type dummy
ip link add dummy15 type dummy
ip link add dummy16 type dummy
ip link add dummy17 type dummy
ip link add dummy18 type dummy
ip link add dummy19 type dummy
ip link add dummy20 type dummy
ip link add dummy21 type dummy
ip link add dummy22 type dummy
ip link add dummy23 type dummy
ip link add dummy24 type dummy
ip link add dummy25 type dummy
ip link add dummy26 type dummy
ip link add dummy27 type dummy
ip link add dummy28 type dummy
ip link add dummy29 type dummy
ip link add dummy30 type dummy
ip link add dummy31 type dummy
ip link add dummy32 type dummy
ip link set dummy1 up
ip link set dummy2 up
ip link set dummy3 up
ip link set dummy4 up
ip link set dummy5 up
ip link set dummy6 up
ip link set dummy7 up
ip link set dummy8 up
ip link set dummy9 up
ip link set dummy10 up
ip link set dummy11 up
ip link set dummy12 up
ip link set dummy13 up
ip link set dummy14 up
ip link set dummy15 up
ip link set dummy16 up
ip link set dummy17 up
ip link set dummy18 up
ip link set dummy19 up
ip link set dummy20 up
ip link set dummy21 up
ip link set dummy22 up
ip link set dummy23 up
ip link set dummy24 up
ip link set dummy25 up
ip link set dummy26 up
ip link set dummy27 up
ip link set dummy28 up
ip link set dummy29 up
ip link set dummy30 up
ip link set dummy31 up
ip link set dummy32 up
ip link set dummy33 up
ip link set dummy34 up
ip link set vrf-red up
ip link set vrf-blue up
ip link set dummyVRFred up
ip link set dummyVRFblue up
ip ro add 1.1.1.1/32 dev dummy1
ip ro add 1.1.1.2/32 dev dummy2
ip ro add 1.1.1.3/32 dev dummy3
ip ro add 1.1.1.4/32 dev dummy4
ip ro add 1.1.1.5/32 dev dummy5
ip ro add 1.1.1.6/32 dev dummy6
ip ro add 1.1.1.7/32 dev dummy7
ip ro add 1.1.1.8/32 dev dummy8
ip ro add 1.1.1.9/32 dev dummy9
ip ro add 1.1.1.10/32 dev dummy10
ip ro add 1.1.1.11/32 dev dummy11
ip ro add 1.1.1.12/32 dev dummy12
ip ro add 1.1.1.13/32 dev dummy13
ip ro add 1.1.1.14/32 dev dummy14
ip ro add 1.1.1.15/32 dev dummy15
ip ro add 1.1.1.16/32 dev dummy16
ip ro add 1.1.1.17/32 dev dummy17
ip ro add 1.1.1.18/32 dev dummy18
ip ro add 1.1.1.19/32 dev dummy19
ip ro add 1.1.1.20/32 dev dummy20
ip ro add 1.1.1.21/32 dev dummy21
ip ro add 1.1.1.22/32 dev dummy22
ip ro add 1.1.1.23/32 dev dummy23
ip ro add 1.1.1.24/32 dev dummy24
ip ro add 1.1.1.25/32 dev dummy25
ip ro add 1.1.1.26/32 dev dummy26
ip ro add 1.1.1.27/32 dev dummy27
ip ro add 1.1.1.28/32 dev dummy28
ip ro add 1.1.1.29/32 dev dummy29
ip ro add 1.1.1.30/32 dev dummy30
ip ro add 1.1.1.31/32 dev dummy31
ip ro add 1.1.1.32/32 dev dummy32
ip next add id 1 via 1.1.1.1 dev dummy1
ip next add id 2 via 1.1.1.2 dev dummy2
ip next add id 3 via 1.1.1.3 dev dummy3
ip next add id 4 via 1.1.1.4 dev dummy4
ip next add id 5 via 1.1.1.5 dev dummy5
ip next add id 6 via 1.1.1.6 dev dummy6
ip next add id 7 via 1.1.1.7 dev dummy7
ip next add id 8 via 1.1.1.8 dev dummy8
ip next add id 9 via 1.1.1.9 dev dummy9
ip next add id 10 via 1.1.1.10 dev dummy10
ip next add id 11 via 1.1.1.11 dev dummy11
ip next add id 12 via 1.1.1.12 dev dummy12
ip next add id 13 via 1.1.1.13 dev dummy13
ip next add id 14 via 1.1.1.14 dev dummy14
ip next add id 15 via 1.1.1.15 dev dummy15
ip next add id 16 via 1.1.1.16 dev dummy16
ip next add id 17 via 1.1.1.17 dev dummy17
ip next add id 18 via 1.1.1.18 dev dummy18
ip next add id 19 via 1.1.1.19 dev dummy19
ip next add id 20 via 1.1.1.20 dev dummy20
ip next add id 21 via 1.1.1.21 dev dummy21
ip next add id 22 via 1.1.1.22 dev dummy22
ip next add id 23 via 1.1.1.23 dev dummy23
ip next add id 24 via 1.1.1.24 dev dummy24
ip next add id 25 via 1.1.1.25 dev dummy25
ip next add id 26 via 1.1.1.26 dev dummy26
ip next add id 27 via 1.1.1.27 dev dummy27
ip next add id 28 via 1.1.1.28 dev dummy28
ip next add id 29 via 1.1.1.29 dev dummy29
ip next add id 30 via 1.1.1.30 dev dummy30
ip next add id 31 via 1.1.1.31 dev dummy31
ip next add id 32 via 1.1.1.32 dev dummy32
i=100
while [ $i -le 200 ]
do
ip next add id $i group 1/2/3/4/5/6/7/8/9/10/11/12/13/14/15/16/17/18/19
echo $i
((i++))
done
ip next add id 999 group 1/2/3/4/5/6
ip next ls
========================
Fixes: ab84be7e54 ("net: Initial nexthop code")
Signed-off-by: Stephen Worley <sworley@cumulusnetworks.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jason A. Donenfeld says:
====================
wireguard fixes for 5.7-rc7
Hopefully these are the last fixes for 5.7:
1) A trivial bump in the selftest harness to support gcc-10.
build.wireguard.com is still on gcc-9 but I'll probably switch to
gcc-10 in the coming weeks.
2) A concurrency fix regarding userspace modifying the pre-shared key at
the same time as packets are being processed, reported by Matt
Dunwoodie.
3) We were previously clearing skb->hash on egress, which broke
fq_codel, cake, and other things that actually make use of the flow
hash for queueing, reported by Dave Taht and Toke Høiland-Jørgensen.
4) A fix for the increased memory usage caused by (3). This can be
thought of as part of patch (3), but because of the separate
reasoning and breadth of it I thought made it a bit cleaner to put in
a standalone commit.
Fixes (2), (3), and (4) are -stable material.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
In "wireguard: queueing: preserve flow hash across packet scrubbing", we
were required to slightly increase the size of the receive replay
counter to something still fairly small, but an increase nonetheless.
It turns out that we can recoup some of the additional memory overhead
by splitting up the prior union type into two distinct types. Before, we
used the same "noise_counter" union for both sending and receiving, with
sending just using a simple atomic64_t, while receiving used the full
replay counter checker. This meant that most of the memory being
allocated for the sending counter was being wasted. Since the old
"noise_counter" type increased in size in the prior commit, now is a
good time to split up that union type into a distinct "noise_replay_
counter" for receiving and a boring atomic64_t for sending, each using
neither more nor less memory than required.
Also, since sometimes the replay counter is accessed without
necessitating additional accesses to the bitmap, we can reduce cache
misses by hoisting the always-necessary lock above the bitmap in the
struct layout. We also change a "noise_replay_counter" stack allocation
to kmalloc in a -DDEBUG selftest so that KASAN doesn't trigger a stack
frame warning.
All and all, removing a bit of abstraction in this commit makes the code
simpler and smaller, in addition to the motivating memory usage
recuperation. For example, passing around raw "noise_symmetric_key"
structs is something that really only makes sense within noise.c, in the
one place where the sending and receiving keys can safely be thought of
as the same type of object; subsequent to that, it's important that we
uniformly access these through keypair->{sending,receiving}, where their
distinct roles are always made explicit. So this patch allows us to draw
that distinction clearly as well.
Fixes: e7096c131e ("net: WireGuard secure network tunnel")
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It's important that we clear most header fields during encapsulation and
decapsulation, because the packet is substantially changed, and we don't
want any info leak or logic bug due to an accidental correlation. But,
for encapsulation, it's wrong to clear skb->hash, since it's used by
fq_codel and flow dissection in general. Without it, classification does
not proceed as usual. This change might make it easier to estimate the
number of innerflows by examining clustering of out of order packets,
but this shouldn't open up anything that can't already be inferred
otherwise (e.g. syn packet size inference), and fq_codel can be disabled
anyway.
Furthermore, it might be the case that the hash isn't used or queried at
all until after wireguard transmits the encrypted UDP packet, which
means skb->hash might still be zero at this point, and thus no hash
taken over the inner packet data. In order to address this situation, we
force a calculation of skb->hash before encrypting packet data.
Of course this means that fq_codel might transmit packets slightly more
out of order than usual. Toke did some testing on beefy machines with
high quantities of parallel flows and found that increasing the
reply-attack counter to 8192 takes care of the most pathological cases
pretty well.
Reported-by: Dave Taht <dave.taht@gmail.com>
Reviewed-and-tested-by: Toke Høiland-Jørgensen <toke@toke.dk>
Fixes: e7096c131e ("net: WireGuard secure network tunnel")
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Prior we read the preshared key after dropping the handshake lock, which
isn't an actual crypto issue if it races, but it's still not quite
correct. So copy that part of the state into a temporary like we do with
the rest of the handshake state variables. Then we can release the lock,
operate on the temporary, and zero it out at the end of the function. In
performance tests, the impact of this was entirely unnoticable, probably
because those bytes are coming from the same cacheline as other things
that are being copied out in the same manner.
Reported-by: Matt Dunwoodie <ncon@noconroy.net>
Fixes: e7096c131e ("net: WireGuard secure network tunnel")
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
gcc-10 switched to defaulting to -fno-common, which broke iproute2-5.4.
This was fixed in iproute-5.6, so switch to that. Because we're after a
stable testing surface, we generally don't like to bump these
unnecessarily, but in this case, being able to actually build is a basic
necessity.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As discussed in [0], it's dangerous to allow mapping BPF map, that's meant to
be frozen and is read-only on BPF program side, because that allows user-space
to actually store a writable view to the page even after it is frozen. This is
exacerbated by BPF verifier making a strong assumption that contents of such
frozen map will remain unchanged. To prevent this, disallow mapping
BPF_F_RDONLY_PROG mmap()'able BPF maps as writable, ever.
[0] https://lore.kernel.org/bpf/CAEf4BzYGWYhXdp6BJ7_=9OQPJxQpgug080MMjdSB72i9R+5c6g@mail.gmail.com/
Fixes: fc9702273e ("bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY")
Suggested-by: Jann Horn <jannh@google.com>
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Jann Horn <jannh@google.com>
Link: https://lore.kernel.org/bpf/20200519053824.1089415-1-andriin@fb.com
secid_to_secctx is not stackable, and since the BPF LSM registers this
hook by default, the call_int_hook logic is not suitable which
"bails-on-fail" and casues issues when other LSMs register this hook and
eventually breaks Audit.
In order to fix this, directly iterate over the security hooks instead
of using call_int_hook as suggested in:
https: //lore.kernel.org/bpf/9d0eb6c6-803a-ff3a-5603-9ad6d9edfc00@schaufler-ca.com/#t
Fixes: 98e828a065 ("security: Refactor declaration of LSM hooks")
Fixes: 625236ba38 ("security: Fix the default value of secid_to_secctx hook")
Reported-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: KP Singh <kpsingh@google.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: James Morris <jamorris@linux.microsoft.com>
Link: https://lore.kernel.org/bpf/20200520125616.193765-1-kpsingh@chromium.org
arch/riscv/mm/init.c: In function ‘print_vm_layout’:
arch/riscv/mm/init.c:68:37: error: ‘FIXADDR_START’ undeclared (first use in this function);
arch/riscv/mm/init.c:69:20: error: ‘FIXADDR_TOP’ undeclared
arch/riscv/mm/init.c:70:37: error: ‘PCI_IO_START’ undeclared
arch/riscv/mm/init.c:71:20: error: ‘PCI_IO_END’ undeclared
arch/riscv/mm/init.c:72:38: error: ‘VMEMMAP_START’ undeclared
arch/riscv/mm/init.c:73:20: error: ‘VMEMMAP_END’ undeclared (first use in this function);
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
[Why]
We dropped the delay after changed the cursor functions locking the
entire pipe to locking just the CURSOR registers to fix page flip
stuttering - this introduced cursor stuttering instead, and an underflow
issue.
The cursor update can be delayed indefinitely if the cursor update
repeatedly happens right around VUPDATE.
The underflow issue can happen if we do a viewport update on a pipe
on the same frame where a cursor update happens around VUPDATE - the
old cursor registers are retained which can be in an invalid position.
This can cause a pipe hang and indefinite underflow.
[How]
The complex, ideal solution to the problem would be a software
triple buffering mechanism from the DM layer to program only one cursor
update per frame just before VUPDATE.
The simple workaround until we have that infrastructure in place is
this change - bring back the delay until VUPDATE before locking, but
with some corrections to the calculations.
This didn't work for all timings before because the calculation for
VUPDATE was wrong - it was using the offset from VSTARTUP instead and
didn't correctly handle the case where VUPDATE could be in the back
porch.
Add a new hardware sequencer function to use the existing helper to
calculate the real VUPDATE start and VUPDATE end - VUPDATE can last
multiple lines after all.
Change the udelay to incorporate the width of VUPDATE as well.
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
During the rework for removing the FPU issues, I found the following
warning:
[..] dml_common_defs.o: warning: objtool: dml_round()+0x9: FPU
instruction outside of kernel_fpu_{begin,end}()
This file has a single function that does not need to be in a specific
file. This commit drop dml_common_defs file, and move dml_round function
to dml_inline_defs.
CC: Christian König <christian.koenig@amd.com>
CC: Alexander Deucher <Alexander.Deucher@amd.com>
CC: Peter Zijlstra <peterz@infradead.org>
CC: Tony Cheng <tony.cheng@amd.com>
CC: Harry Wentland <hwentlan@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
DP training sequence to set SCRAMBLING_DISABLE bit properly based on
training pattern - per DP Spec.
[How]
Update dpcd_pattern.v1_4.SCRAMBLING_DISABLE with 1 for TPS1, TPS2, TPS3,
but not for TPS4.
Signed-off-by: Vladimir Stempen <vladimir.stempen@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
- Fix a PM regression in brcmnand driver
- Propagate ECC information correctly on SPI-NAND
- Make sure no MTD name is used multiple time in nvmem
-----BEGIN PGP SIGNATURE-----
iQJKBAABCAA0FiEEdgfidid8lnn52cLTZvlZhesYu8EFAl7Fh/EWHHJpY2hhcmRA
c2lnbWEtc3Rhci5hdAAKCRBm+VmF6xi7we+tEADnoxI9nF+PU/4nCn2ppZJUE/Rq
dpnp7+gPIDZpnjANN/Xq7AaRkVdHRjAVv4F1n1rhikUcIULCGvkmDrb3YO8QLql4
cDGJCehSc26jOx84s8c6pi7gE9HWV9SEEGLpMjQyigogRfu7SkI9dJGZ9FzuvaBd
+upBOdcYRQ54T9iDAXQZV7suWLUbvqRGS+a+NUkQMdE+eneDX96VjGsRK3SUSEFa
IK8ppoQRKMMRWlkas4UowOesk3thRkl1P99Mu/zCnSSYIKePimizyPKkQNsO0ZQ3
vtaGJOMT6AA0fhT3ka7axQAkaVkZLKRTivj/z0fq4uqY9UNa9mZ3uJ1GD2+BtgYt
IzGEpvCwBac+1MIcdTUTxLjFysWc43enLTkv8aFtw0PrfpRiH/efJWH12KHs0LXO
H/clnW1ky7PzU2LWl8iJqGCYI1tdrEbJwnyjycwjsS1g10L6ItYRfxgu0dI4ANXJ
EH0yySLtpjYOPu9hU2LQjVizMOPPwIDOD7QnRjwR72OZj6mS7yr03bW24t8nyl1Z
/n4wuhHCjZdDkZzw5B9x5gm4lpYUVxp6IChcSz3t241R84Gw7poBbXWNFcX84VG4
ZfU8k64olzVvDVJx/mInpED8POWQXsg26HNKNwBXscoc0GlEs3kItz9ZkwWmsyRD
vmpzCxUJtcyEI5d15w==
=L8AH
-----END PGP SIGNATURE-----
Merge tag 'fixes-for-5.7-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux
Pull MTD fixes from Richard Weinberger:
- Fix a PM regression in brcmnand driver
- Propagate ECC information correctly on SPI-NAND
- Make sure no MTD name is used multiple time in nvmem
* tag 'fixes-for-5.7-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux:
mtd:rawnand: brcmnand: Fix PM resume crash
mtd: Fix mtd not registered due to nvmem name collision
mtd: spinand: Propagate ECC information to the MTD structure
- Correctly set next cursor for detailed_erase_block_info debugfs file
- Don't use crypto_shash_descsize() for digest size in UBIFS
- Remove broken lazytime support from UBIFS
-----BEGIN PGP SIGNATURE-----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=eCh7
-----END PGP SIGNATURE-----
Merge tag 'for-linus-5.7-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs
Pull UBI and UBIFS fixes from Richard Weinberger:
- Correctly set next cursor for detailed_erase_block_info debugfs file
- Don't use crypto_shash_descsize() for digest size in UBIFS
- Remove broken lazytime support from UBIFS
* tag 'for-linus-5.7-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs:
ubi: Fix seq_file usage in detailed_erase_block_info debugfs file
ubifs: fix wrong use of crypto_shash_descsize()
ubifs: remove broken lazytime support
- Two missing includes which cause build issues on recent systems
- Correctly set TRANS_GRE_LEN in our vector network driver
-----BEGIN PGP SIGNATURE-----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=p8hh
-----END PGP SIGNATURE-----
Merge tag 'for-linus-5.7-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml
Pull UML fixes from Richard Weinberger:
- Two missing includes which caused build issues on recent systems
- Correctly set TRANS_GRE_LEN in our vector network driver
* tag 'for-linus-5.7-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml:
um: Fix typo in vector driver transport option definition
um: syscall.c: include <asm/unistd.h>
um: Fix xor.h include
Make a recently introduced suspend-to-idle wakeup issue on Dell
XPS13 9360 go away.
-----BEGIN PGP SIGNATURE-----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=ik70
-----END PGP SIGNATURE-----
Merge tag 'pm-5.7-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull power management fix from Rafael Wysocki:
"This makes a recently introduced suspend-to-idle wakeup issue on Dell
XPS13 9360 go away"
* tag 'pm-5.7-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
ACPI: EC: PM: Avoid flushing EC work when EC GPE is inactive
-----BEGIN PGP SIGNATURE-----
iHUEABYIAB0WIQSQHSd0lITzzeNWNm3h3BK/laaZPAUCXsU6SgAKCRDh3BK/laaZ
PABRAP9MCZz/CLH2sEqHqH9KQHScNc4uf4bReiCU1hrLs7PbYwD/Y+vbRMffki7I
B/gt0Dg4kGxG5CV+ckeZK0+p2NWUUgQ=
=PPLW
-----END PGP SIGNATURE-----
Merge tag 'ovl-fixes-5.7-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs
Pull overlayfs fixes from Miklos Szeredi:
"Fix two bugs introduced in this cycle and one introduced in v5.5"
* tag 'ovl-fixes-5.7-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs:
ovl: potential crash in ovl_fid_to_fh()
ovl: clear ATTR_OPEN from attr->ia_valid
ovl: clear ATTR_FILE from attr->ia_valid
The Rx protocol has a "previousPacket" field in it that is not handled in
the same way by all protocol implementations. Sometimes it contains the
serial number of the last DATA packet received, sometimes the sequence
number of the last DATA packet received and sometimes the highest sequence
number so far received.
AF_RXRPC is using this to weed out ACKs that are out of date (it's possible
for ACK packets to get reordered on the wire), but this does not work with
OpenAFS which will just stick the sequence number of the last packet seen
into previousPacket.
The issue being seen is that big AFS FS.StoreData RPC (eg. of ~256MiB) are
timing out when partly sent. A trace was captured, with an additional
tracepoint to show ACKs being discarded in rxrpc_input_ack(). Here's an
excerpt showing the problem.
52873.203230: rxrpc_tx_data: c=000004ae DATA ed1a3584:00000002 0002449c q=00024499 fl=09
A DATA packet with sequence number 00024499 has been transmitted (the "q="
field).
...
52873.243296: rxrpc_rx_ack: c=000004ae 00012a2b DLY r=00024499 f=00024497 p=00024496 n=0
52873.243376: rxrpc_rx_ack: c=000004ae 00012a2c IDL r=0002449b f=00024499 p=00024498 n=0
52873.243383: rxrpc_rx_ack: c=000004ae 00012a2d OOS r=0002449d f=00024499 p=0002449a n=2
The Out-Of-Sequence ACK indicates that the server didn't see DATA sequence
number 00024499, but did see seq 0002449a (previousPacket, shown as "p=",
skipped the number, but firstPacket, "f=", which shows the bottom of the
window is set at that point).
52873.252663: rxrpc_retransmit: c=000004ae q=24499 a=02 xp=14581537
52873.252664: rxrpc_tx_data: c=000004ae DATA ed1a3584:00000002 000244bc q=00024499 fl=0b *RETRANS*
The packet has been retransmitted. Retransmission recurs until the peer
says it got the packet.
52873.271013: rxrpc_rx_ack: c=000004ae 00012a31 OOS r=000244a1 f=00024499 p=0002449e n=6
More OOS ACKs indicate that the other packets that are already in the
transmission pipeline are being received. The specific-ACK list is up to 6
ACKs and NAKs.
...
52873.284792: rxrpc_rx_ack: c=000004ae 00012a49 OOS r=000244b9 f=00024499 p=000244b6 n=30
52873.284802: rxrpc_retransmit: c=000004ae q=24499 a=0a xp=63505500
52873.284804: rxrpc_tx_data: c=000004ae DATA ed1a3584:00000002 000244c2 q=00024499 fl=0b *RETRANS*
52873.287468: rxrpc_rx_ack: c=000004ae 00012a4a OOS r=000244ba f=00024499 p=000244b7 n=31
52873.287478: rxrpc_rx_ack: c=000004ae 00012a4b OOS r=000244bb f=00024499 p=000244b8 n=32
At this point, the server's receive window is full (n=32) with presumably 1
NAK'd packet and 31 ACK'd packets. We can't transmit any more packets.
52873.287488: rxrpc_retransmit: c=000004ae q=24499 a=0a xp=61327980
52873.287489: rxrpc_tx_data: c=000004ae DATA ed1a3584:00000002 000244c3 q=00024499 fl=0b *RETRANS*
52873.293850: rxrpc_rx_ack: c=000004ae 00012a4c DLY r=000244bc f=000244a0 p=00024499 n=25
And now we've received an ACK indicating that a DATA retransmission was
received. 7 packets have been processed (the occupied part of the window
moved, as indicated by f= and n=).
52873.293853: rxrpc_rx_discard_ack: c=000004ae r=00012a4c 000244a0<00024499 00024499<000244b8
However, the DLY ACK gets discarded because its previousPacket has gone
backwards (from p=000244b8, in the ACK at 52873.287478 to p=00024499 in the
ACK at 52873.293850).
We then end up in a continuous cycle of retransmit/discard. kafs fails to
update its window because it's discarding the ACKs and can't transmit an
extra packet that would clear the issue because the window is full.
OpenAFS doesn't change the previousPacket value in the ACKs because no new
DATA packets are received with a different previousPacket number.
Fix this by altering the discard check to only discard an ACK based on
previousPacket if there was no advance in the firstPacket. This allows us
to transmit a new packet which will cause previousPacket to advance in the
next ACK.
The check, however, needs to allow for the possibility that previousPacket
may actually have had the serial number placed in it instead - in which
case it will go outside the window and we should ignore it.
Fixes: 1a2391c30c ("rxrpc: Fix detection of out of order acks")
Reported-by: Dave Botsch <botsch@cnf.cornell.edu>
Signed-off-by: David Howells <dhowells@redhat.com>
