| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
There have been some findings during Eric Biggers rework of the
paes implementation which this patch tries to address:
A very minor finding within paes ctr where when the cpacf instruction
returns with only partially data en/decrytped the walk_done() was
mistakenly done with the all data counter. Please note this can only
happen when the kmctr returns because the protected key became invalid
in the middle of the operation. And this is only with suspend and
resume on a system with different effective wrapping key.
Eric Biggers mentioned that the context struct within the tfm struct
may be shared among multiple kernel threads. So here now a rework
which uses a spinlock per context to protect the read and write of the
protected key blob value. The en/decrypt functions copy the protected
key(s) at the beginning into a param struct and do not work with the
protected key within the context any more. If the protected key in the
param struct becomes invalid, the key material is again converted to
protected key(s) and the context gets this update protected by the
spinlock. Race conditions are still possible and may result in writing
the very same protected key value more than once. So the spinlock
needs to make sure the protected key(s) within the context are
consistent updated.
The ctr page is now locked by a mutex instead of a spinlock. A similar
patch went into the aes_s390 code as a result of a complain "sleeping
function called from invalid context at ...algapi.h". See
commit 1c2c7029c008 ("s390/crypto: fix possible sleep during spinlock
aquired")' for more.
During testing with instrumented code another issue with the xts
en/decrypt function revealed. The retry cleared the running iv value
and thus let to wrong en/decrypted data.
Tested and verified with additional testcases via AF_ALG interface and
additional selftests within the kernel (which will be made available
as soon as possible).
Reported-by: Eric Biggers <ebiggers@kernel.org>
Signed-off-by: Harald Freudenberger <freude@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Instead of using our own kprobes-on-ftrace handling convert the
code to support KPROBES_ON_FTRACE.
Signed-off-by: Sven Schnelle <svens@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Commit ee71d16d22bb ("s390/mm: make TASK_SIZE independent from the number
of page table levels") changed the logic of TASK_SIZE and also removed the
arch_mmap_check() implementation for s390. This combination has a subtle
effect on how get_unmapped_area() for hugetlbfs pages works. It is now
possible that a user process establishes a hugetlbfs mapping at an address
above 4 TB, without triggering a dynamic pagetable upgrade from 3 to 4
levels.
This is because hugetlbfs mappings will not use mm->get_unmapped_area, but
rather file->f_op->get_unmapped_area, which currently is the generic
implementation of hugetlb_get_unmapped_area() that does not know about s390
dynamic pagetable upgrades, but with the new definition of TASK_SIZE, it
will now allow mappings above 4 TB.
Subsequent access to such a mapped address above 4 TB will result in a page
fault loop, because the CPU cannot translate such a large address with 3
pagetable levels. The fault handler will try to map in a hugepage at the
address, but due to the folded pagetable logic it will end up with creating
entries in the 3 level pagetable, possibly overwriting existing mappings,
and then it all repeats when the access is retried.
Apart from the page fault loop, this can have various nasty effects, e.g.
kernel panic from one of the BUG_ON() checks in memory management code,
or even data loss if an existing mapping gets overwritten.
Fix this by implementing HAVE_ARCH_HUGETLB_UNMAPPED_AREA support for s390,
providing an s390 version for hugetlb_get_unmapped_area() with pagetable
upgrade support similar to arch_get_unmapped_area(), which will then be
used instead of the generic version.
Fixes: ee71d16d22bb ("s390/mm: make TASK_SIZE independent from the number of page table levels")
Cc: <stable@vger.kernel.org> # 4.12+
Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Add default MDIO_BCM_IPROC Kconfig setting such that it is default
on for IPROC family of devices.
Signed-off-by: Scott Branden <scott.branden@broadcom.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Multicast and broadcast packets can be looped from egress to ingress
pre segmentation with dev_loopback_xmit. That function unconditionally
sets ip_summed to CHECKSUM_UNNECESSARY.
udp_rcv_segment segments gso packets in the udp rx path. Segmentation
usually executes on egress, and does not expect packets of this type.
__udp_gso_segment interprets !CHECKSUM_PARTIAL as CHECKSUM_NONE. But
the offsets are not correct for gso_make_checksum.
UDP GSO packets are of type CHECKSUM_PARTIAL, with their uh->check set
to the correct pseudo header checksum. Reset ip_summed to this type.
(CHECKSUM_PARTIAL is allowed on ingress, see comments in skbuff.h)
Reported-by: syzbot <syzkaller@googlegroups.com>
Fixes: cf329aa42b66 ("udp: cope with UDP GRO packet misdirection")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The old netem mailing list was inactive and recently was targeted by
spammers. Switch to just using netdev mailing list which is where all
the real change happens.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
Harald Freudenberger
Sven Schnelle
Gerald Schaefer
Scott Branden
Willem de Bruijn
Stephen Hemminger