If any memory allocation in resize_stripes fails we will return
-ENOMEM, but in some cases we update conf->pool_size anyway.
This means that if we try again, the allocations will be assumed
to be larger than they are, and badness results.
So only update pool_size if there is no error.
This bug was introduced in 2.6.17 and the patch is suitable for
-stable.
Fixes: ad01c9e3752f ("[PATCH] md: Allow stripes to be expanded in preparation for expanding an array") Signed-off-by: NeilBrown <neilb@suse.de> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit c20694a054e903745591f4f85b39ecbce4e58349)
Using the indenting we can see the curly braces were obviously intended.
This is a static checker fix, but my guess is that we don't read enough
bytes, because we don't calculate "t_len" correctly.
Fixes: f1d82698029b ('memstick: use fully asynchronous request processing') Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Cc: Alex Dubov <oakad@yahoo.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 0fd0b9f448e1ff459ea4f718def61a197b15bb4c)
ptrace_resume() is called when the tracee is still __TASK_TRACED. We set
tracee->exit_code and then wake_up_state() changes tracee->state. If the
tracer's sub-thread does wait() in between, task_stopped_code(ptrace => T)
wrongly looks like another report from tracee.
This confuses debugger, and since wait_task_stopped() clears ->exit_code
the tracee can miss a signal.
Note for stable: the bug is very old, but without 9899d11f6544 "ptrace:
ensure arch_ptrace/ptrace_request can never race with SIGKILL" the fix
should use lock_task_sighand(child).
Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reported-by: Pavel Labath <labath@google.com> Tested-by: Pavel Labath <labath@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit e3f81ba2f0546f030fc234f7aade3016532c75b1)
Sebastian reported a crash caused by a jump label mismatch after resume.
This happens because we do not save the kernel text section during suspend
and therefore also do not restore it during resume, but use the kernel image
that restores the old system.
This means that after a suspend/resume cycle we lost all modifications done
to the kernel text section.
The reason for this is the pfn_is_nosave() function, which incorrectly
returns that read-only pages don't need to be saved. This is incorrect since
we mark the kernel text section read-only.
We still need to make sure to not save and restore pages contained within
NSS and DCSS segment.
To fix this add an extra case for the kernel text section and only save
those pages if they are not contained within an NSS segment.
Fixes the following crash (and the above bugs as well):
This problem appears to have been introduced in 2.6.29 by commit 93197a36a9c1 "Rewrite sysfs processor cache info code".
This caused lscpu to error out on at least e500v2 devices, eg:
error: cannot open /sys/devices/system/cpu/cpu0/cache/index2/size: No such file or directory
Some embedded powerpc systems use cache-size in DTS for the unified L2
cache size, not d-cache-size, so we need to allow for both DTS names.
Added a new CACHE_TYPE_UNIFIED_D cache_type_info structure to handle
this.
Fixes: 93197a36a9c1 ("powerpc: Rewrite sysfs processor cache info code") Signed-off-by: Dave Olson <olson@cumulusnetworks.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
[bwh: Backported to 3.2:
- Adjust context
- Preserve __cpuinit attribute on cache_do_one_devnode_unified()] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit d6de5ca93fd6425dcdb21cc341365991c0c444cc) Signed-off-by: Willy Tarreau <w@1wt.eu>
There is a race condition between e1000_change_mtu's cleanups and
netpoll, when we change the MTU across jumbo size:
Changing MTU frees all the rx buffers:
e1000_change_mtu -> e1000_down -> e1000_clean_all_rx_rings ->
e1000_clean_rx_ring
Then, close to the end of e1000_change_mtu:
pr_info -> ... -> netpoll_poll_dev -> e1000_clean ->
e1000_clean_rx_irq -> e1000_alloc_rx_buffers -> e1000_alloc_frag
And when we come back to do the rest of the MTU change:
e1000_up -> e1000_configure -> e1000_configure_rx ->
e1000_alloc_jumbo_rx_buffers
alloc_jumbo finds the buffers already != NULL, since data (shared with
page in e1000_rx_buffer->rxbuf) has been re-alloc'd, but it's garbage,
or at least not what is expected when in jumbo state.
This results in an unusable adapter (packets don't get through), and a
NULL pointer dereference on the next call to e1000_clean_rx_ring
(other mtu change, link down, shutdown):
BUG: unable to handle kernel NULL pointer dereference at (null)
IP: [<ffffffff81194d6e>] put_compound_page+0x7e/0x330
We have two problems in UDP stack related to bogus checksums :
1) We return -EAGAIN to application even if receive queue is not empty.
This breaks applications using edge trigger epoll()
2) Under UDP flood, we can loop forever without yielding to other
processes, potentially hanging the host, especially on non SMP.
This patch is an attempt to make things better.
We might in the future add extra support for rt applications
wanting to better control time spent doing a recv() in a hostile
environment. For example we could validate checksums before queuing
packets in socket receive queue.
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Willem de Bruijn <willemb@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
CVE-2015-5364
CVE-2015-5366
Check length of extended attributes and allocation descriptors when
loading inodes from disk. Otherwise corrupted filesystems could confuse
the code and make the kernel oops.
Reported-by: Carl Henrik Lunde <chlunde@ping.uio.no> Signed-off-by: Jan Kara <jack@suse.cz>
[bwh: Backported to 2.6.32: use make_bad_inode() instead of returning error] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
CVE-2015-4167
Store blocksize in a local variable in udf_fill_inode() since it is used
a lot of times.
Signed-off-by: Jan Kara <jack@suse.cz>
[bwh: Needed for the following fix. Backported to 2.6.32: adjust context.] Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
Check that length specified in a component of a symlink fits in the
input buffer we are reading. Also properly ignore component length for
component types that do not use it. Otherwise we read memory after end
of buffer for corrupted udf image.
Reported-by: Carl Henrik Lunde <chlunde@ping.uio.no> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
CVE-2014-9728, CVE-2014-9730
Symlink reading code does not check whether the resulting path fits into
the page provided by the generic code. This isn't as easy as just
checking the symlink size because of various encoding conversions we
perform on path. So we have to check whether there is still enough space
in the buffer on the fly.
Reported-by: Carl Henrik Lunde <chlunde@ping.uio.no> Signed-off-by: Jan Kara <jack@suse.cz>
[bwh: Backported to 2.6.32: adjust context, indentation] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
CVE-2014-9731
Currently, we ignore symlink component of type 2. But mkisofs and other OS'
seem to treat it as / so do the same for compatibility.
Reported-by: "Gábor S." <otnaccess@hotmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
[bwh: Needed for the following fix] Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
UDF specification allows arbitrarily large symlinks. However we support
only symlinks at most one block large. Check the length of the symlink
so that we don't access memory beyond end of the symlink block.
Reported-by: Carl Henrik Lunde <chlunde@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
[bwh: Backported to 2.6.32: adjust context] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
CVE-2014-9728
Verify that inode size is sane when loading inode with data stored in
ICB. Otherwise we may get confused later when working with the inode and
inode size is too big.
Reported-by: Carl Henrik Lunde <chlunde@ping.uio.no> Signed-off-by: Jan Kara <jack@suse.cz>
[bwh: Backported to 2.6.32: on error, call make_bad_inode() then return] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
CVE-2014-9728, CVE-2014-9729
There are a few very theoretical off-by-one bugs in report descriptor size
checking when performing a pre-parsing fixup. Fix those.
Reported-by: Ben Hawkes <hawkes@google.com> Reviewed-by: Benjamin Tissoires <benjamin.tissoires@redhat.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
[bwh: Backported to 2.6.32:
- Adjust context
- Drop change to a quirk in hid-lg.c that doesn't exist here] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
CVE-2014-3184
Dmitry Chernenkov used KASAN to discover that eCryptfs writes past the
end of the allocated buffer during encrypted filename decoding. This
fix corrects the issue by getting rid of the unnecessary 0 write when
the current bit offset is 2.
Non-root user-space processes can send Netlink messages to other
processes that are well-known for being subscribed to Netlink
asynchronous notifications. This allows ilegitimate non-root
process to send forged messages to Netlink subscribers.
The userspace process usually verifies the legitimate origin in
two ways:
a) Socket credentials. If UID != 0, then the message comes from
some ilegitimate process and the message needs to be dropped.
b) Netlink portID. In general, portID == 0 means that the origin
of the messages comes from the kernel. Thus, discarding any
message not coming from the kernel.
However, ctnetlink sets the portID in event messages that has
been triggered by some user-space process, eg. conntrack utility.
So other processes subscribed to ctnetlink events, eg. conntrackd,
know that the event was triggered by some user-space action.
Neither of the two ways to discard ilegitimate messages coming
from non-root processes can help for ctnetlink.
This patch adds capability validation in case that dst_pid is set
in netlink_sendmsg(). This approach is aggressive since existing
applications using any Netlink bus to deliver messages between
two user-space processes will break. Note that the exception is
NETLINK_USERSOCK, since it is reserved for netlink-to-netlink
userspace communication.
Still, if anyone wants that his Netlink bus allows netlink-to-netlink
userspace, then they can set NL_NONROOT_SEND. However, by default,
I don't think it makes sense to allow to use NETLINK_ROUTE to
communicate two processes that are sending no matter what information
that is not related to link/neighbouring/routing. They should be using
NETLINK_USERSOCK instead for that.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: David S. Miller <davem@davemloft.net>
[bwh: Backported to 2.6.32:
- Adjust context
- NETLINK_USERSOCK does not exist, so drop that part] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
CVE-2012-6689
When tty_driver_lookup_tty fails in tty_open, we forget to drop a
reference to the tty driver. This was added by commit 4a2b5fddd5 (Move
tty lookup/reopen to caller).
Fix that by adding tty_driver_kref_put to the fail path.
I will refactor the code later. This is for the ease of backporting to
stable.
Introduced-in: v2.6.28-rc2 Signed-off-by: Jiri Slaby <jslaby@suse.cz> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Acked-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
[bwh: Backported to 2.6.32: adjust filename] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
CVE-2011-5321
unfortunately, allowing an arbitrary 16bit value means a possibility of
overflow in the calculation of total number of pages in bio_map_user_iov() -
we rely on there being no more than PAGE_SIZE members of sum in the
first loop there. If that sum wraps around, we end up allocating
too small array of pointers to pages and it's easy to overflow it in
the second loop.
X-Coverup: TINC (and there's no lumber cartel either) Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
[bwh: s/MAX_UIOVEC/UIO_MAXIOV/. This was fixed upstream by commit fdc81f45e9f5 ("sg_start_req(): use import_iovec()"), but we don't have
that function.] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 07213eed86c17c544bb10568fc04e49e03730ab7) Signed-off-by: Willy Tarreau <w@1wt.eu>
Ben Hutchings [Mon, 15 Jun 2015 02:51:55 +0000 (03:51 +0100)]
pipe: iovec: Fix memory corruption when retrying atomic copy as non-atomic
pipe_iov_copy_{from,to}_user() may be tried twice with the same iovec,
the first time atomically and the second time not. The second attempt
needs to continue from the iovec position, pipe buffer offset and
remaining length where the first attempt failed, but currently the
pipe buffer offset and remaining length are reset. This will corrupt
the piped data (possibly also leading to an information leak between
processes) and may also corrupt kernel memory.
This was fixed upstream by commits f0d1bec9d58d ("new helper:
copy_page_from_iter()") and 637b58c2887e ("switch pipe_read() to
copy_page_to_iter()"), but those aren't suitable for stable. This fix
for older kernel versions was made by Seth Jennings for RHEL and I
have extracted it from their update.
CVE-2015-1805
References: https://bugzilla.redhat.com/show_bug.cgi?id=1202855 Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 75cf667b7fac08a7b21694adca7dff07361be68a) Signed-off-by: Willy Tarreau <w@1wt.eu>
We found that after v3.10.73, recvmsg might return -EFAULT while -EINVAL
was expected.
We tested it through the recvmsg01 testcase come from LTP testsuit. It set
msg->msg_namelen to -1 and the recvmsg syscall returned errno 14, which is
unexpected (errno 22 is expected):
Linux mainline has no this bug for commit 08adb7dab fixes it accidentally.
However, it is too large and complex to be backported to LTS 3.10.
Commit 281c9c36 (net: compat: Update get_compat_msghdr() to match
copy_msghdr_from_user() behaviour) made get_compat_msghdr() return
error if msg_sys->msg_namelen was negative, which changed the behaviors
of recvmsg and sendmsg syscall in a lib32 system:
Before commit 281c9c36, get_compat_msghdr() wouldn't fail and it would
return -EINVAL in move_addr_to_user() or somewhere if msg_sys->msg_namelen
was invalid and then syscall returned -EINVAL, which is correct.
And now, when msg_sys->msg_namelen is negative, get_compat_msghdr() will
fail and wants to return -EINVAL, however, the outer syscall will return
-EFAULT directly, which is unexpected.
This patch gets the return value of get_compat_msghdr() as well as
copy_msghdr_from_user(), then returns this expected value if
get_compat_msghdr() fails.
Fixes: 281c9c36 (net: compat: Update get_compat_msghdr() to match copy_msghdr_from_user() behaviour) Signed-off-by: Junling Zheng <zhengjunling@huawei.com> Signed-off-by: Hanbing Xu <xuhanbing@huawei.com> Cc: Li Zefan <lizefan@huawei.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: David Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
Willy Tarreau [Wed, 3 Jun 2015 15:03:26 +0000 (17:03 +0200)]
net: fix incorrect backport of tcp_send_fin in 2.6.32.66
Eric forwarded this bug report happening since 2.6.32.66, found that the
backport of commit 845704a5 ("tcp: avoid looping in tcp_send_fin()") was
incorrect and proposed this patch to fix it. The bug was also reported by
starlight.2015q2@binnacle.cx who confirmed the fix.
> Date: Fri, 29 May 2015 09:12:45 +0000
> From: "bugzilla-daemon@bugzilla.kernel.org" <bugzilla-daemon@bugzilla.kernel.org>
> To: "shemminger@linux-foundation.org" <shemminger@linux-foundation.org>
> Subject: [Bug 99161] New: 2.6.32.66 PPC Oops in tcp_send_fin
>
>
> https://bugzilla.kernel.org/show_bug.cgi?id=99161
>
> Bug ID: 99161
> Summary: 2.6.32.66 PPC Oops in tcp_send_fin
> Product: Networking
> Version: 2.5
> Kernel Version: 2.6.32.66
> Hardware: PPC-32
> OS: Linux
> Tree: Mainline
> Status: NEW
> Severity: normal
> Priority: P1
> Component: IPV4
> Assignee: shemminger@linux-foundation.org
> Reporter: varenet@parisc-linux.org
> Regression: No
>
> I just updated my trusty old PPC box to longterm 2.6.32.66 (was running .65
> before that with zero issue) and it started spewing oopses at me like hell
> broke loose. This machine is primarily used as a DNS and MX (albeit under low
> pressure).
(...)
Cc: Eric Dumazet <edumazet@google.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
Catalin Marinas [Fri, 20 Mar 2015 16:48:13 +0000 (16:48 +0000)]
net: compat: Update get_compat_msghdr() to match copy_msghdr_from_user() behaviour
Commit db31c55a6fb2 (net: clamp ->msg_namelen instead of returning an
error) introduced the clamping of msg_namelen when the unsigned value
was larger than sizeof(struct sockaddr_storage). This caused a
msg_namelen of -1 to be valid. The native code was subsequently fixed by
commit dbb490b96584 (net: socket: error on a negative msg_namelen).
In addition, the native code sets msg_namelen to 0 when msg_name is
NULL. This was done in commit (6a2a2b3ae075 net:socket: set msg_namelen
to 0 if msg_name is passed as NULL in msghdr struct from userland) and
subsequently updated by 08adb7dabd48 (fold verify_iovec() into
copy_msghdr_from_user()).
This patch brings the get_compat_msghdr() in line with
copy_msghdr_from_user().
Fixes: db31c55a6fb2 (net: clamp ->msg_namelen instead of returning an error) Cc: David S. Miller <davem@davemloft.net> Cc: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit 91edd096e224941131f896b86838b1e59553696a) Cc: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
sound/oss: fix deadlock in sequencer_ioctl(SNDCTL_SEQ_OUTOFBAND)
A deadlock can be initiated by userspace via ioctl(SNDCTL_SEQ_OUTOFBAND)
on /dev/sequencer with TMR_ECHO midi event.
In this case the control flow is:
sound_ioctl()
-> case SND_DEV_SEQ:
case SND_DEV_SEQ2:
sequencer_ioctl()
-> case SNDCTL_SEQ_OUTOFBAND:
spin_lock_irqsave(&lock,flags);
play_event();
-> case EV_TIMING:
seq_timing_event()
-> case TMR_ECHO:
seq_copy_to_input()
-> spin_lock_irqsave(&lock,flags);
It seems that spin_lock_irqsave() around play_event() is not necessary,
because the only other call location in seq_startplay() makes the call
without acquiring spinlock.
So, the patch just removes spinlocks around play_event().
By the way, it removes unreachable code in seq_timing_event(),
since (seq_mode == SEQ_2) case is handled in the beginning.
Compile tested only.
Found by Linux Driver Verification project (linuxtesting.org).
Fix B-tree corruption when a new record is inserted at position 0 in the node
in hfs_brec_insert(). In this case a hfs_brec_update_parent() is called to
update the parent index node (if exists) and it is passed hfs_find_data with
a search_key containing a newly inserted key instead of the key to be updated.
This results in an inconsistent index node. The bug reproduces on my machine
after an extents overflow record for the catalog file (CNID=4) is inserted into
the extents overflow B-tree. Because of a low (reserved) value of CNID=4, it
has to become the first record in the first leaf node.
The resulting first leaf node is correct:
----------------------------------------------------
| key0.CNID=4 | key1.CNID=123 | key2.CNID=456, ... |
----------------------------------------------------
But the parent index key0 still contains the previous key CNID=123:
-----------------------
| key0.CNID=123 | ... |
-----------------------
A change in hfs_brec_insert() makes hfs_brec_update_parent() work correctly
by preventing it from getting fd->record=-1 value from __hfs_brec_find().
Along the way, I removed duplicate code with unification of the if condition.
The resulting code is equivalent to the original code because node is never 0.
Also hfs_brec_update_parent() will now return an error after getting a negative
fd->record value. However, the return value of hfs_brec_update_parent() is not
checked anywhere in the file and I'm leaving it unchanged by this patch.
brec.c lacks error checking after some other calls too, but this issue is of
less importance than the one being fixed by this patch.
Cc: stable@vger.kernel.org Cc: Joe Perches <joe@perches.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Vyacheslav Dubeyko <slava@dubeyko.com> Cc: Hin-Tak Leung <htl10@users.sourceforge.net> Cc: Anton Altaparmakov <aia21@cam.ac.uk> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christoph Hellwig <hch@infradead.org> Signed-off-by: Sergei Antonov <saproj@gmail.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
If userland creates a timer without specifying a sigevent info, we'll
create one ourself, using a stack local variable. Particularly will we
use the timer ID as sival_int. But as sigev_value is a union containing
a pointer and an int, that assignment will only partially initialize
sigev_value on systems where the size of a pointer is bigger than the
size of an int. On such systems we'll copy the uninitialized stack bytes
from the timer_create() call to userland when the timer actually fires
and we're going to deliver the signal.
Initialize sigev_value with 0 to plug the stack info leak.
Found in the PaX patch, written by the PaX Team.
Fixes: 5a9fa7307285 ("posix-timers: kill ->it_sigev_signo and...") Signed-off-by: Mathias Krause <minipli@googlemail.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Brad Spengler <spender@grsecurity.net> Cc: PaX Team <pageexec@freemail.hu> Link: http://lkml.kernel.org/r/1412456799-32339-1-git-send-email-minipli@googlemail.com Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
[bwh: Backported to 3.2: adjust filename] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 3cd3a349aa3519b88d29845c0bc36bcbae158e93)
When sg_scsi_ioctl() fails to prepare request to submit in
blk_rq_map_kern() we jump to a label where we just end up copying
(luckily zeroed-out) kernel buffer to userspace instead of reporting
error. Fix the problem by jumping to the right label.
CC: Jens Axboe <axboe@kernel.dk> CC: linux-scsi@vger.kernel.org
Coverity-id: 1226871 Signed-off-by: Jan Kara <jack@suse.cz>
Fixed up the, now unused, out label.
Signed-off-by: Jens Axboe <axboe@fb.com> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit d73b032b63e8967462e1cf5763858ed89e97880f)
If rpc.statd is restarted, upcalls to monitor hosts can fail with
ECONNREFUSED. In that case force a lookup of statd's new port and retry the
upcall.
Signed-off-by: Benjamin Coddington <bcodding@redhat.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
[bwh: Backported to 3.2: not using RPC_TASK_SOFTCONN] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 3aabe891f32c209a2be7cd5581d2634020e801c1)
As pointed by recent post[1] on exploiting DRAM physical imperfection,
/proc/PID/pagemap exposes sensitive information which can be used to do
attacks.
This disallows anybody without CAP_SYS_ADMIN to read the pagemap.
fib_nh_match does not match nexthops correctly. Example:
ip route add 172.16.10/24 nexthop via 192.168.122.12 dev eth0 \
nexthop via 192.168.122.13 dev eth0
ip route del 172.16.10/24 nexthop via 192.168.122.14 dev eth0 \
nexthop via 192.168.122.15 dev eth0
Del command is successful and route is removed. After this patch
applied, the route is correctly matched and result is:
RTNETLINK answers: No such process
Please consider this for stable trees as well.
Fixes: 4e902c57417c4 ("[IPv4]: FIB configuration using struct fib_config") Signed-off-by: Jiri Pirko <jiri@resnulli.us> Acked-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 0aba46add2915b344580569e87d9c41274b9c475)
`spidev_message()` sums the lengths of the individual SPI transfers to
determine the overall SPI message length. It restricts the total
length, returning an error if too long, but it does not check for
arithmetic overflow. For example, if the SPI message consisted of two
transfers and the first has a length of 10 and the second has a length
of (__u32)(-1), the total length would be seen as 9, even though the
second transfer is actually very long. If the second transfer specifies
a null `rx_buf` and a non-null `tx_buf`, the `copy_from_user()` could
overrun the spidev's pre-allocated tx buffer before it reaches an
invalid user memory address. Fix it by checking that neither the total
nor the individual transfer lengths exceed the maximum allowed value.
Thanks to Dan Carpenter for reporting the potential integer overflow.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Mark Brown <broonie@kernel.org>
[Ian Abbott: Note: original commit compares the lengths to INT_MAX
instead of bufsiz due to changes in earlier commits.] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 7499401e4a0b01ee43cff768de4ca630dcd0bc64) Signed-off-by: Willy Tarreau <w@1wt.eu>
Presence of an unbound loop in tcp_send_fin() had always been hard
to explain when analyzing crash dumps involving gigantic dying processes
with millions of sockets.
Lets try a different strategy :
In case of memory pressure, try to add the FIN flag to last packet
in write queue, even if packet was already sent. TCP stack will
be able to deliver this FIN after a timeout event. Note that this
FIN being delivered by a retransmit, it also carries a Push flag
given our current implementation.
By checking sk_under_memory_pressure(), we anticipate that cooking
many FIN packets might deplete tcp memory.
In the case we could not allocate a packet, even with __GFP_WAIT
allocation, then not sending a FIN seems quite reasonable if it allows
to get rid of this socket, free memory, and not block the process from
eventually doing other useful work.
Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
[bwh: Backported to 3.2:
- Drop inapplicable change to sk_forced_wmem_schedule()
- s/sk_under_memory_pressure(sk)/tcp_memory_pressure/] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 82241580d7734af2207ad0bb1720904f569dac3a)
[wt: backported to 2.6.32: s/TCPHDR_FIN/TCPCB_FLAG_FIN/] Signed-off-by: Willy Tarreau <w@1wt.eu>
Forwarded frames should not have a socket attached. Especially
tw sockets will lead to panics later-on in the stack.
This was observed with TPROXY assigning a tw socket and broken
policy routing (misconfigured). As a result frame enters
forwarding path instead of input. We cannot solve this in
TPROXY as it cannot know that policy routing is broken.
v2:
Remove useless comment
Signed-off-by: Sebastian Poehn <sebastian.poehn@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit fccb908d23fbae3e941e9294590dd94de6b1d822)
While working on sk_forward_alloc problems reported by Denys
Fedoryshchenko, we found that tcp connect() (and fastopen) do not call
sk_wmem_schedule() for SYN packet (and/or SYN/DATA packet), so
sk_forward_alloc is negative while connect is in progress.
We can fix this by calling regular sk_stream_alloc_skb() both for the
SYN packet (in tcp_connect()) and the syn_data packet in
tcp_send_syn_data()
Then, tcp_send_syn_data() can avoid copying syn_data as we simply
can manipulate syn_data->cb[] to remove SYN flag (and increment seq)
Instead of open coding memcpy_fromiovecend(), simply use this helper.
This leaves in socket write queue clean fast clone skbs.
This was tested against our fastopen packetdrill tests.
Reported-by: Denys Fedoryshchenko <nuclearcat@nuclearcat.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Acked-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
[bwh: Backported to 3.2:
- Drop the Fast Open changes
- Adjust context] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 3e2eb8946907b2d53eb906e13e01d273c6534f5c)
[I would really like an ACK on that one from dhowells; it appears to be
quite straightforward, but...]
MSG_PEEK isn't passed to ->recvmsg() via msg->msg_flags; as the matter of
fact, neither the kernel users of rxrpc, nor the syscalls ever set that bit
in there. It gets passed via flags; in fact, another such check in the same
function is done correctly - as flags & MSG_PEEK.
It had been that way (effectively disabled) for 8 years, though, so the patch
needs beating up - that case had never been tested. If it is correct, it's
-stable fodder.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 10c82cd7d46e4c525b046c399fcd285ce138198e)
The rds_iw_update_cm_id function stores a large 'struct rds_sock' object
on the stack in order to pass a pair of addresses. This happens to just
fit withint the 1024 byte stack size warning limit on x86, but just
exceed that limit on ARM, which gives us this warning:
net/rds/iw_rdma.c:200:1: warning: the frame size of 1056 bytes is larger than 1024 bytes [-Wframe-larger-than=]
As the use of this large variable is basically bogus, we can rearrange
the code to not do that. Instead of passing an rds socket into
rds_iw_get_device, we now just pass the two addresses that we have
available in rds_iw_update_cm_id, and we change rds_iw_get_mr accordingly,
to create two address structures on the stack there.
Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Sowmini Varadhan <sowmini.varadhan@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 3fe2d645fe4ea7ff6cba9020685e46c1a1dff9c0)
sysctl has sysctl.net.core.rmem_*/wmem_* parameters which can be
set to incorrect values. Given that 'struct sk_buff' allocates from
rcvbuf, incorrectly set buffer length could result to memory
allocation failures. For example, set them as follows:
Moreover, the possible minimum is 1, so we can get another kernel panic:
...
BUG: unable to handle kernel paging request at ffff88013caee5c0
IP: [<ffffffff815604cf>] __alloc_skb+0x12f/0x1f0
...
Signed-off-by: Alexey Kodanev <alexey.kodanev@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
[bwh: Backported to 3.2: delete now-unused 'one' variable] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 2d6dfb109bfbf3abd5f762173b1d73fd321dbe37) Signed-off-by: Willy Tarreau <w@1wt.eu>
I found if we write a larger than 4GB value to some sysctl
variables, the sending syscall will hang up forever, because these
variables are 32 bits, such large values make them overflow to 0 or
negative.
This patch try to fix overflow or prevent from zero value setup
of below sysctl variables:
net.core.wmem_default
net.core.rmem_default
net.core.rmem_max
net.core.wmem_max
net.ipv4.udp_rmem_min
net.ipv4.udp_wmem_min
net.ipv4.tcp_wmem
net.ipv4.tcp_rmem
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: Li Yu <raise.sail@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
[bwh: Backported to 3.2:
- Adjust context
- Delete now-unused 'zero' variable] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 98eee187cdee2807bd80e6c02180c5c2abae6453)
[wt: backported to 2.6.32: set strategy to sysctl_intvec where relevant] Signed-off-by: Willy Tarreau <w@1wt.eu>
If an over-MTU UDP datagram is sent through a SOCK_RAW socket to a
UFO-capable device, ip_ufo_append_data() sets skb->ip_summed to
CHECKSUM_PARTIAL unconditionally as all GSO code assumes transport layer
checksum is to be computed on segmentation. However, in this case,
skb->csum_start and skb->csum_offset are never set as raw socket
transmit path bypasses udp_send_skb() where they are usually set. As a
result, driver may access invalid memory when trying to calculate the
checksum and store the result (as observed in virtio_net driver).
Moreover, the very idea of modifying the userspace provided UDP header
is IMHO against raw socket semantics (I wasn't able to find a document
clearly stating this or the opposite, though). And while allowing
CHECKSUM_NONE in the UFO case would be more efficient, it would be a bit
too intrusive change just to handle a corner case like this. Therefore
disallowing UFO for packets from SOCK_DGRAM seems to be the best option.
Signed-off-by: Michal Kubecek <mkubecek@suse.cz> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 332640b2821f75381b1049a904d93d4fb846334f) Signed-off-by: Willy Tarreau <w@1wt.eu>
Steffen Klassert [Wed, 29 Jun 2011 23:19:32 +0000 (23:19 +0000)]
ipv4: Don't use ufo handling on later transformed packets
We might call ip_ufo_append_data() for packets that will be IPsec
transformed later. This function should be used just for real
udp packets. So we check for rt->dst.header_len which is only
nonzero on IPsec handling and call ip_ufo_append_data() just
if rt->dst.header_len is zero.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit c146066ab80267c3305de5dda6a4083f06df9265) Signed-off-by: Willy Tarreau <w@1wt.eu>
colons are used as a separator in netdev device lookup in dev_ioctl.c
Specific functions are SIOCGIFTXQLEN SIOCETHTOOL SIOCSIFNAME
Signed-off-by: Matthew Thode <mthode@mthode.org> Signed-off-by: David S. Miller <davem@davemloft.net>
[bwh: Backported to 3.2: adjust context] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit d501ebeb7da7531e92e3c8d194730341c314ff2d)
In tcf_em_validate(), after calling request_module() to load the
kind-specific module, set em->ops to NULL before returning -EAGAIN, so
that module_put() is not called again by tcf_em_tree_destroy().
Signed-off-by: Ignacy GawÄ\99dzki <ignacy.gawedzki@green-communications.fr> Acked-by: Cong Wang <cwang@twopensource.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 9405be73262273a6904688fb2f1e60a13117cf1b)
When we've run out of space in the output buffer to store more data, we
will call zlib_deflate with a NULL output buffer until we've consumed
remaining input.
When this happens, olen contains the size the output buffer would have
consumed iff we'd have had enough room.
This can later cause skb_over_panic when ppp_generic skb_put()s
the returned length.
Reported-by: Iain Douglas <centos@1n6.org.uk> Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 8bcd64423836bad3638684677f6d740bc7c9297f) Signed-off-by: Willy Tarreau <w@1wt.eu>
Linux manpage for recvmsg and sendmsg calls does not explicitly mention setting msg_namelen to 0 when
msg_name passed set as NULL. When developers don't set msg_namelen member in msghdr, it might contain garbage
value which will fail the validation check and sendmsg and recvmsg calls from kernel will return EINVAL. This will
break old binaries and any code for which there is no access to source code.
To fix this, we set msg_namelen to 0 when msg_name is passed as NULL from userland.
Signed-off-by: Ani Sinha <ani@arista.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit d29f1f53e5299e0bbb3e33ef8d35ed657fa633b6)
A local route may have a lower hop_limit set than global routes do.
RFC 3756, Section 4.2.7, "Parameter Spoofing"
> 1. The attacker includes a Current Hop Limit of one or another small
> number which the attacker knows will cause legitimate packets to
> be dropped before they reach their destination.
> As an example, one possible approach to mitigate this threat is to
> ignore very small hop limits. The nodes could implement a
> configurable minimum hop limit, and ignore attempts to set it below
> said limit.
Signed-off-by: D.S. Ljungmark <ljungmark@modio.se> Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: David S. Miller <davem@davemloft.net>
[bwh: Backported to 3.2: adjust ND_PRINTK() usage] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit f10f7d2a8200fe33c5030c7e32df3a2b3561f3cd)
Max unacked packets/bytes is an int while sizeof(long) was used in the
sysctl table.
This means that when they were getting read we'd also leak kernel memory
to userspace along with the timeout values.
Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 3760b67b3e419b9ac42a45417491360a14a35357)
The timeout entries are sizeof(int) rather than sizeof(long), which
means that when they were getting read we'd also leak kernel memory
to userspace along with the timeout values.
Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 88fe14be08a475ad0eea4ca7c51f32437baf41af)
Properly verify that the resulting page aligned end address is larger
than both the start address and the length of the memory area requested.
Both the start and length arguments for ib_umem_get are controlled by
the user. A misbehaving user can provide values which will cause an
integer overflow when calculating the page aligned end address.
This overflow can cause also miscalculation of the number of pages
mapped, and additional logic issues.
Addresses: CVE-2014-8159 Signed-off-by: Shachar Raindel <raindel@mellanox.com> Signed-off-by: Jack Morgenstein <jackm@mellanox.com> Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com> Signed-off-by: Roland Dreier <roland@purestorage.com> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 485f16b743d98527620396639b73d7214006f3c7)
When hitting an INIT collision case during the 4WHS with AUTH enabled, as
already described in detail in commit 1be9a950c646 ("net: sctp: inherit
auth_capable on INIT collisions"), it can happen that we occasionally
still remotely trigger the following panic on server side which seems to
have been uncovered after the fix from commit 1be9a950c646 ...
Since this only triggers in some collision-cases with AUTH, the problem at
heart is that sctp_auth_key_put() on asoc->asoc_shared_key is called twice
when having refcnt 1, once directly in sctp_assoc_update() and yet again
from within sctp_auth_asoc_init_active_key() via sctp_assoc_update() on
the already kzfree'd memory, which is also consistent with the observation
of the poison decrease from 0x6b to 0x6a (note: the overwrite is detected
at a later point in time when poison is checked on new allocation).
Reference counting of auth keys revisited:
Shared keys for AUTH chunks are being stored in endpoints and associations
in endpoint_shared_keys list. On endpoint creation, a null key is being
added; on association creation, all endpoint shared keys are being cached
and thus cloned over to the association. struct sctp_shared_key only holds
a pointer to the actual key bytes, that is, struct sctp_auth_bytes which
keeps track of users internally through refcounting. Naturally, on assoc
or enpoint destruction, sctp_shared_key are being destroyed directly and
the reference on sctp_auth_bytes dropped.
User space can add keys to either list via setsockopt(2) through struct
sctp_authkey and by passing that to sctp_auth_set_key() which replaces or
adds a new auth key. There, sctp_auth_create_key() creates a new sctp_auth_bytes
with refcount 1 and in case of replacement drops the reference on the old
sctp_auth_bytes. A key can be set active from user space through setsockopt()
on the id via sctp_auth_set_active_key(), which iterates through either
endpoint_shared_keys and in case of an assoc, invokes (one of various places)
sctp_auth_asoc_init_active_key().
sctp_auth_asoc_init_active_key() computes the actual secret from local's
and peer's random, hmac and shared key parameters and returns a new key
directly as sctp_auth_bytes, that is asoc->asoc_shared_key, plus drops
the reference if there was a previous one. The secret, which where we
eventually double drop the ref comes from sctp_auth_asoc_set_secret() with
intitial refcount of 1, which also stays unchanged eventually in
sctp_assoc_update(). This key is later being used for crypto layer to
set the key for the hash in crypto_hash_setkey() from sctp_auth_calculate_hmac().
To close the loop: asoc->asoc_shared_key is freshly allocated secret
material and independant of the sctp_shared_key management keeping track
of only shared keys in endpoints and assocs. Hence, also commit 4184b2a79a76
("net: sctp: fix memory leak in auth key management") is independant of
this bug here since it concerns a different layer (though same structures
being used eventually). asoc->asoc_shared_key is reference dropped correctly
on assoc destruction in sctp_association_free() and when active keys are
being replaced in sctp_auth_asoc_init_active_key(), it always has a refcount
of 1. Hence, it's freed prematurely in sctp_assoc_update(). Simple fix is
to remove that sctp_auth_key_put() from there which fixes these panics.
Fixes: 730fc3d05cd4 ("[SCTP]: Implete SCTP-AUTH parameter processing") Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Willy Tarreau <w@1wt.eu>
A very minimal and simple user space application allocating an SCTP
socket, setting SCTP_AUTH_KEY setsockopt(2) on it and then closing
the socket again will leak the memory containing the authentication
key from user space:
This is bad because of two things, we can bring down a machine from
user space when auth_enable=1, but also we would leave security sensitive
keying material in memory without clearing it after use. The issue is
that sctp_auth_create_key() already sets the refcount to 1, but after
allocation sctp_auth_set_key() does an additional refcount on it, and
thus leaving it around when we free the socket.
Fixes: 65b07e5d0d0 ("[SCTP]: API updates to suport SCTP-AUTH extensions.") Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Cc: Vlad Yasevich <vyasevich@gmail.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 3af10169145c8eed7b3591c0644da4298405efbc)
We didn't check length of rock ridge ER records before printing them.
Thus corrupted isofs image can cause us to access and print some memory
behind the buffer with obvious consequences.
Reported-and-tested-by: Carl Henrik Lunde <chlunde@ping.uio.no> CC: stable@vger.kernel.org Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Willy Tarreau <w@1wt.eu>
Rock Ridge extensions define so called Continuation Entries (CE) which
define where is further space with Rock Ridge data. Corrupted isofs
image can contain arbitrarily long chain of these, including a one
containing loop and thus causing kernel to end in an infinite loop when
traversing these entries.
Limit the traversal to 32 entries which should be more than enough space
to store all the Rock Ridge data.
Reported-by: P J P <ppandit@redhat.com> CC: stable@vger.kernel.org Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Willy Tarreau <w@1wt.eu>
-P FORWARD DROP
-A FORWARD -m sctp --dport 9 -j ACCEPT
-A FORWARD -p tcp --dport 80 -j ACCEPT
-A FORWARD -p tcp -m conntrack -m state ESTABLISHED,RELATED -j ACCEPT
One would assume that this allows SCTP on port 9 and TCP on port 80.
Unfortunately, if the SCTP conntrack module is not loaded, this allows
*all* SCTP communication, to pass though, i.e. -p sctp -j ACCEPT,
which we think is a security issue.
This is because on the first SCTP packet on port 9, we create a dummy
"generic l4" conntrack entry without any port information (since
conntrack doesn't know how to extract this information).
All subsequent packets that are unknown will then be in established
state since they will fallback to proto_generic and will match the
'generic' entry.
Our originally proposed version [1] completely disabled generic protocol
tracking, but Jozsef suggests to not track protocols for which a more
suitable helper is available, hence we now mitigate the issue for in
tree known ct protocol helpers only, so that at least NAT and direction
information will still be preserved for others.
Ben Hutchings [Thu, 29 Jan 2015 02:50:33 +0000 (02:50 +0000)]
splice: Apply generic position and size checks to each write
We need to check the position and size of file writes against various
limits, using generic_write_check(). This was not being done for
the splice write path. It was fixed upstream by commit 8d0207652cbe
("->splice_write() via ->write_iter()") but we can't apply that.
CVE-2014-7822
Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
Robert Baldyga [Mon, 24 Nov 2014 06:56:21 +0000 (07:56 +0100)]
serial: samsung: wait for transfer completion before clock disable
This patch adds waiting until transmit buffer and shifter will be empty
before clock disabling.
Without this fix it's possible to have clock disabled while data was
not transmited yet, which causes unproper state of TX line and problems
in following data transfers.
Cc: stable@vger.kernel.org Signed-off-by: Robert Baldyga <r.baldyga@samsung.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit 1ff383a4c3eda8893ec61b02831826e1b1f46b41) Signed-off-by: Willy Tarreau <w@1wt.eu>
On virtual environments, apic_read could take a long time. As a
result, under certain conditions the ack pending loop may exit
without any queued irqs left, but after more than one second. A
warning will be printed needlessly in this case.
If the loop is about to exit regardless of max_loops, don't
update it.
'ret_from_fork' checks TIF_IA32 to determine whether 'pt_regs' and
the related state make sense for 'ret_from_sys_call'. This is
entirely the wrong check. TS_COMPAT would make a little more
sense, but there's really no point in keeping this optimization
at all.
This fixes a return to the wrong user CS if we came from int
0x80 in a 64-bit task.
Signed-off-by: Andy Lutomirski <luto@amacapital.net> Cc: Borislav Petkov <bp@alien8.de> Cc: Denys Vlasenko <dvlasenk@redhat.com> Cc: H. Peter Anvin <hpa@zytor.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Thomas Gleixner <tglx@linutronix.de> Link: http://lkml.kernel.org/r/4710be56d76ef994ddf59087aad98c000fbab9a4.1424989793.git.luto@amacapital.net
[ Backported from tip:x86/asm. ] Signed-off-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 159891c0953a89a28f793fc52373b031262c44d2)
This adds the workaround for erratum 793 as a precaution in case not
every BIOS implements it. This addresses CVE-2013-6885.
Erratum text:
[Revision Guide for AMD Family 16h Models 00h-0Fh Processors,
document 51810 Rev. 3.04 November 2013]
793 Specific Combination of Writes to Write Combined Memory Types and
Locked Instructions May Cause Core Hang
Description
Under a highly specific and detailed set of internal timing
conditions, a locked instruction may trigger a timing sequence whereby
the write to a write combined memory type is not flushed, causing the
locked instruction to stall indefinitely.
Potential Effect on System
Processor core hang.
Suggested Workaround
BIOS should set MSR
C001_1020[15] = 1b.
Fix Planned
No fix planned
[ hpa: updated description, fixed typo in MSR name ]
Signed-off-by: Borislav Petkov <bp@suse.de> Link: http://lkml.kernel.org/r/20140114230711.GS29865@pd.tnic Tested-by: Aravind Gopalakrishnan <aravind.gopalakrishnan@amd.com> Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
[bwh: Backported to 3.2:
- Adjust filename
- Venkatesh Srinivas pointed out we should use {rd,wr}msrl_safe() to
avoid crashing on KVM. This was fixed upstream by commit 8f86a7373a1c
("x86, AMD: Convert to the new bit access MSR accessors") but that's too
much trouble to backport. Here we must use {rd,wr}msrl_amd_safe().] Signed-off-by: Willy Tarreau <w@1wt.eu>
Note that, it declares the "random_variable" variable as "unsigned int". Since
the result of the shifting operation between STACK_RND_MASK (which is
0x3fffff on x86_64, 22 bits) and PAGE_SHIFT (which is 12 on x86_64):
random_variable <<= PAGE_SHIFT;
then the two leftmost bits are dropped when storing the result in the
"random_variable". This variable shall be at least 34 bits long to hold the
(22+12) result.
These two dropped bits have an impact on the entropy of process stack.
Concretely, the total stack entropy is reduced by four: from 2^28 to 2^30 (One
fourth of expected entropy).
This patch restores back the entropy by correcting the types involved in the
operations in the functions randomize_stack_top() and stack_maxrandom_size().
The theory behind vdso randomization is that it's mapped at a random
offset above the top of the stack. To avoid wasting a page of
memory for an extra page table, the vdso isn't supposed to extend
past the lowest PMD into which it can fit. Other than that, the
address should be a uniformly distributed address that meets all of
the alignment requirements.
The current algorithm is buggy: the vdso has about a 50% probability
of being at the very end of a PMD. The current algorithm also has a
decent chance of failing outright due to incorrect handling of the
case where the top of the stack is near the top of its PMD.
This fixes the implementation. The paxtest estimate of vdso
"randomisation" improves from 11 bits to 18 bits. (Disclaimer: I
don't know what the paxtest code is actually calculating.)
It's worth noting that this algorithm is inherently biased: the vdso
is more likely to end up near the end of its PMD than near the
beginning. Ideally we would either nix the PMD sharing requirement
or jointly randomize the vdso and the stack to reduce the bias.
In the mean time, this is a considerable improvement with basically
no risk of compatibility issues, since the allowed outputs of the
algorithm are unchanged.
As an easy test, doing this:
for i in `seq 10000`
do grep -P vdso /proc/self/maps |cut -d- -f1
done |sort |uniq -d
used to produce lots of output (1445 lines on my most recent run).
A tiny subset looks like this:
Note the suspicious fe000 endings. With the fix, I get a much more
palatable 76 repeated addresses.
Reviewed-by: Kees Cook <keescook@chromium.org> Cc: stable@vger.kernel.org Signed-off-by: Andy Lutomirski <luto@amacapital.net>
[bwh: Backported to 2.6.32:
- The whole file is only built for x86_64; adjust context and comment for this
- We don't have align_vdso_addr()] Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
- Disables the F00F bug workaround warning. There is no F00F bug
workaround any more because Linux's standard IDT handling already
works around the F00F bug, but the warning still exists. This
is only cosmetic, and, in any event, there is no such thing as
KVM on a CPU with the F00F bug.
- Disables 32-bit APM BIOS detection. On a KVM paravirt system,
there should be no APM BIOS anyway.
- Disables tboot. I think that the tboot code should check the
CPUID hypervisor bit directly if it matters.
- paravirt_enabled disables espfix32. espfix32 should *not* be
disabled under KVM paravirt.
The last point is the purpose of this patch. It fixes a leak of the
high 16 bits of the kernel stack address on 32-bit KVM paravirt
guests. Fixes CVE-2014-8134.
Suggested-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Andy Lutomirski <luto@amacapital.net> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
[bwh: Backported to 2.6.32: adjust indentation, context] Signed-off-by: Willy Tarreau <w@1wt.eu>
Users have no business installing custom code segments into the
GDT, and segments that are not present but are otherwise valid
are a historical source of interesting attacks.
For completeness, block attempts to set the L bit. (Prior to
this patch, the L bit would have been silently dropped.)
This is an ABI break. I've checked glibc, musl, and Wine, and
none of them look like they'll have any trouble.
Note to stable maintainers: this is a hardening patch that fixes
no known bugs. Given the possibility of ABI issues, this
probably shouldn't be backported quickly.
Signed-off-by: Andy Lutomirski <luto@amacapital.net> Acked-by: H. Peter Anvin <hpa@zytor.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: security@kernel.org <security@kernel.org> Cc: Willy Tarreau <w@1wt.eu> Signed-off-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit fbc3c534ddffeebba6f943945ac71ec83cfa04b8)
Strictly speaking, this code was never correct. It should have set
read_exec_only and seg_not_present to 1 to indicate that it wanted
to find a free slot without putting anything there, or it should
have put something sensible in the TLS slot if it wanted to allocate
a TLS entry for real. The actual effect of this code was to
allocate a bogus segment that could be used to exploit espfix.
The set_thread_area hardening patches changed the behavior, causing
set_thread_area to return -EINVAL and crashing the game.
This changes set_thread_area to interpret this as a request to find
a free slot and to leave it empty, which isn't *quite* what the game
expects but should be close enough to keep it working. In
particular, using the code above to allocate two segments will
allocate the same segment both times.
According to FrostbittenKing on Github, this fixes The Witcher 2.
If this somehow still causes problems, we could instead allocate
a limit==0 32-bit data segment, but that seems rather ugly to me.
32-bit programs don't have an lm bit in their ABI, so they can't
reliably cause LDT_empty to return true without resorting to memset.
They shouldn't need to do this.
This should fix a longstanding, if minor, issue in all 64-bit kernels
as well as a potential regression in the TLS hardening code.
Ben Hutchings [Sun, 7 Dec 2014 19:57:36 +0000 (19:57 +0000)]
proc connector: Delete spurious memset in proc_exit_connector()
Upstream commit e727ca82e0e9 ("proc connector: fix info leaks")
changed many functions that don't exist in 2.6.32.y. When it was
cherry-picked into 2.6.32.61, one extra memset() calls was inserted
into proc_exit_connector(). This results in clearing the cpu
field of exit events.
If the queue is dead blk_execute_rq_nowait() doesn't invoke the done()
callback function. That will result in blk_execute_rq() being stuck
in wait_for_completion(). Avoid this by initializing rq->end_io to the
done() callback before we check the queue state. Also, make sure the
queue lock is held around the invocation of the done() callback. Found
this through source code review.
Signed-off-by: Muthukumar Ratty <muthur@gmail.com> Signed-off-by: Bart Van Assche <bvanassche@acm.org> Reviewed-by: Tejun Heo <tj@kernel.org> Acked-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
[bwh: Backported to 2.6.32: adjust context] Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
blk_insert_cloned_request(), blk_execute_rq_nowait() and
blk_flush_plug_list() either didn't check whether the queue was dead
or did it without holding queue_lock. Update them so that dead state
is checked while holding queue_lock.
AFAICS, this plugs all holes (requeue doesn't matter as the request is
transitioning atomically from in_flight to queued).
Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
[bwh: Backported to 2.6.32:
- Drop inapplicable changes to queue_unplugged() and
blk_flush_plug_list()
- We don't have blk_queue_dead() so open-code it
- Adjust context] Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
Ben Hutchings [Sun, 7 Dec 2014 20:00:27 +0000 (20:00 +0000)]
md/raid6: Fix misapplied backport in 2.6.32.64
Upstream commit 9c4bdf697c39 ("md/raid6: avoid data corruption during
recovery of double-degraded RAID6") changes handle_stripe(), but we
have separate functions for RAID5 and RAID6 and need to apply the
change to handle_stripe6(). When cherry-picked, the change was
wrongly applied to handle_stripe5().
Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
When (de)configuring a vlan interface, the IFF_ALLMULTI ans IFF_PROMISC
flags are cleared or set on the underlying interface. So, if these flags
are changed on a vlan interface that is not up, the flags underlying
interface might be set or cleared twice.
Only propagating flag changes when a device is up makes sure this does
not happen. It also makes sure that an underlying device is not set to
promiscuous or allmulti mode for a vlan device that is down.
Signed-off-by: Matthijs Kooijman <matthijs@stdin.nl> Signed-off-by: David S. Miller <davem@davemloft.net>
[bwh: This is a dependency of commit d2615bf45069 ("net: core: Always
propagate flag changes to interfaces"), already backported in 2.6.32.62] Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
The "new" fragmentation code (since my rewrite almost 5 years ago)
erroneously sets skb->len rather than using skb_trim() to adjust
the length of the first fragment after copying out all the others.
This leaves the skb tail pointer pointing to after where the data
originally ended, and thus causes the encryption MIC to be written
at that point, rather than where it belongs: immediately after the
data.
The impact of this is that if software encryption is done, then
a) encryption doesn't work for the first fragment, the connection
becomes unusable as the first fragment will never be properly
verified at the receiver, the MIC is practically guaranteed to
be wrong
b) we leak up to 8 bytes of plaintext (!) of the packet out into
the air
This is only mitigated by the fact that many devices are capable
of doing encryption in hardware, in which case this can't happen
as the tail pointer is irrelevant in that case. Additionally,
fragmentation is not used very frequently and would normally have
to be configured manually.
Fix this by using skb_trim() properly.
Cc: stable@vger.kernel.org Fixes: 2de8e0d999b8 ("mac80211: rewrite fragmentation") Reported-by: Jouni Malinen <j@w1.fi> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
(backported from commit 338f977f4eb441e69bb9a46eaa0ac715c931a67f)
CVE-2014-8709 BugLink: http://bugs.launchpad.net/bugs/1392013 Signed-off-by: Luis Henriques <luis.henriques@canonical.com> Acked-by: Andy Whitcroft <apw@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Signed-off-by: Andy Whitcroft <apw@canonical.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
While the INIT chunk parameter verification dissects through many things
in order to detect malformed input, it misses to actually check parameters
inside of parameters. E.g. RFC5061, section 4.2.4 proposes a 'set primary
IP address' parameter in ASCONF, which has as a subparameter an address
parameter.
So an attacker may send a parameter type other than SCTP_PARAM_IPV4_ADDRESS
or SCTP_PARAM_IPV6_ADDRESS, param_type2af() will subsequently return 0
and thus sctp_get_af_specific() returns NULL, too, which we then happily
dereference unconditionally through af->from_addr_param().
A minimal way to address this is to check for NULL as we do on all
other such occasions where we know sctp_get_af_specific() could
possibly return with NULL.
Fixes: d6de3097592b ("[SCTP]: Add the handling of "Set Primary IP Address" parameter to INIT") Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Cc: Vlad Yasevich <vyasevich@gmail.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit e40607cbe270a9e8360907cb1e62ddf0736e4864)
CVE-2014-7841 BugLink: http://bugs.launchpad.net/bugs/1392820 Signed-off-by: Luis Henriques <luis.henriques@canonical.com> Acked-by: Andy Whitcroft <apw@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Signed-off-by: Andy Whitcroft <apw@canonical.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
We did not implement any bound on number of indirect ICBs we follow when
loading inode. Thus corrupted medium could cause kernel to go into an
infinite loop, possibly causing a stack overflow.
Fix the possible stack overflow by removing recursion from
__udf_read_inode() and limit number of indirect ICBs we follow to avoid
infinite loops.
Signed-off-by: Jan Kara <jack@suse.cz>
(back ported from commit c03aa9f6e1f938618e6db2e23afef0574efeeb65)
[ luis: adjusted context and replaced udf_err() by printk() ]
CVE-2014-6410 BugLink: http://bugs.launchpad.net/bugs/1370042 Signed-off-by: Luis Henriques <luis.henriques@canonical.com> Signed-off-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
... where ASCONF_a, ASCONF_b, ..., ASCONF_z are good-formed
ASCONFs and have increasing serial numbers, we process such
ASCONF chunk(s) marked with !end_of_packet and !singleton,
since we have not yet reached the SCTP packet end. SCTP does
only do verification on a chunk by chunk basis, as an SCTP
packet is nothing more than just a container of a stream of
chunks which it eats up one by one.
We could run into the case that we receive a packet with a
malformed tail, above marked as trailing JUNK. All previous
chunks are here goodformed, so the stack will eat up all
previous chunks up to this point. In case JUNK does not fit
into a chunk header and there are no more other chunks in
the input queue, or in case JUNK contains a garbage chunk
header, but the encoded chunk length would exceed the skb
tail, or we came here from an entirely different scenario
and the chunk has pdiscard=1 mark (without having had a flush
point), it will happen, that we will excessively queue up
the association's output queue (a correct final chunk may
then turn it into a response flood when flushing the
queue ;)): I ran a simple script with incremental ASCONF
serial numbers and could see the server side consuming
excessive amount of RAM [before/after: up to 2GB and more].
The issue at heart is that the chunk train basically ends
with !end_of_packet and !singleton markers and since commit 2e3216cd54b1 ("sctp: Follow security requirement of responding
with 1 packet") therefore preventing an output queue flush
point in sctp_do_sm() -> sctp_cmd_interpreter() on the input
chunk (chunk = event_arg) even though local_cork is set,
but its precedence has changed since then. In the normal
case, the last chunk with end_of_packet=1 would trigger the
queue flush to accommodate possible outgoing bundling.
In the input queue, sctp_inq_pop() seems to do the right thing
in terms of discarding invalid chunks. So, above JUNK will
not enter the state machine and instead be released and exit
the sctp_assoc_bh_rcv() chunk processing loop. It's simply
the flush point being missing at loop exit. Adding a try-flush
approach on the output queue might not work as the underlying
infrastructure might be long gone at this point due to the
side-effect interpreter run.
One possibility, albeit a bit of a kludge, would be to defer
invalid chunk freeing into the state machine in order to
possibly trigger packet discards and thus indirectly a queue
flush on error. It would surely be better to discard chunks
as in the current, perhaps better controlled environment, but
going back and forth, it's simply architecturally not possible.
I tried various trailing JUNK attack cases and it seems to
look good now.
Joint work with Vlad Yasevich.
Fixes: 2e3216cd54b1 ("sctp: Follow security requirement of responding with 1 packet") Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit 26b87c7881006311828bb0ab271a551a62dcceb4)
CVE-2014-3688 BugLink: http://bugs.launchpad.net/bugs/1386393 Signed-off-by: Luis Henriques <luis.henriques@canonical.com> Acked-by: Andy Whitcroft <andy.whitcroft@canonical.com> Signed-off-by: Brad Figg <brad.figg@canonical.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
... where ASCONF_a equals ASCONF_b chunk (at least both serials
need to be equal), we panic an SCTP server!
The problem is that good-formed ASCONF chunks that we reply with
ASCONF_ACK chunks are cached per serial. Thus, when we receive a
same ASCONF chunk twice (e.g. through a lost ASCONF_ACK), we do
not need to process them again on the server side (that was the
idea, also proposed in the RFC). Instead, we know it was cached
and we just resend the cached chunk instead. So far, so good.
Where things get nasty is in SCTP's side effect interpreter, that
is, sctp_cmd_interpreter():
While incoming ASCONF_a (chunk = event_arg) is being marked
!end_of_packet and !singleton, and we have an association context,
we do not flush the outqueue the first time after processing the
ASCONF_ACK singleton chunk via SCTP_CMD_REPLY. Instead, we keep it
queued up, although we set local_cork to 1. Commit 2e3216cd54b1
changed the precedence, so that as long as we get bundled, incoming
chunks we try possible bundling on outgoing queue as well. Before
this commit, we would just flush the output queue.
Now, while ASCONF_a's ASCONF_ACK sits in the corked outq, we
continue to process the same ASCONF_b chunk from the packet. As
we have cached the previous ASCONF_ACK, we find it, grab it and
do another SCTP_CMD_REPLY command on it. So, effectively, we rip
the chunk->list pointers and requeue the same ASCONF_ACK chunk
another time. Since we process ASCONF_b, it's correctly marked
with end_of_packet and we enforce an uncork, and thus flush, thus
crashing the kernel.
Fix it by testing if the ASCONF_ACK is currently pending and if
that is the case, do not requeue it. When flushing the output
queue we may relink the chunk for preparing an outgoing packet,
but eventually unlink it when it's copied into the skb right
before transmission.
Joint work with Vlad Yasevich.
Fixes: 2e3216cd54b1 ("sctp: Follow security requirement of responding with 1 packet") Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit b69040d8e39f20d5215a03502a8e8b4c6ab78395)
CVE-2014-3687 BugLink: http://bugs.launchpad.net/bugs/1386392 Signed-off-by: Luis Henriques <luis.henriques@canonical.com> Acked-by: Andy Whitcroft <andy.whitcroft@canonical.com> Signed-off-by: Brad Figg <brad.figg@canonical.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
This patch fixes a potential security issue in the whiteheat USB driver
which might allow a local attacker to cause kernel memory corrpution. This
is due to an unchecked memcpy into a fixed size buffer (of 64 bytes). On
EHCI and XHCI busses it's possible to craft responses greater than 64
bytes leading a buffer overflow.
Signed-off-by: James Forshaw <forshaw@google.com> Cc: stable <stable@vger.kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(backported from commit 6817ae225cd650fb1c3295d769298c38b1eba818)
CVE-2014-3185 BugLink: http://bugs.launchpad.net/bugs/1370036 Signed-off-by: Luis Henriques <luis.henriques@canonical.com> Signed-off-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
There are two issues with the current implementation for replacing user
controls. The first is that the code does not check if the control is actually a
user control and neither does it check if the control is owned by the process
that tries to remove it. That allows userspace applications to remove arbitrary
controls, which can cause a user after free if a for example a driver does not
expect a control to be removed from under its feed.
The second issue is that on one hand when a control is replaced the
user_ctl_count limit is not checked and on the other hand the user_ctl_count is
increased (even though the number of user controls does not change). This allows
userspace, once the user_ctl_count limit as been reached, to repeatedly replace
a control until user_ctl_count overflows. Once that happens new controls can be
added effectively bypassing the user_ctl_count limit.
Both issues can be fixed by instead of open-coding the removal of the control
that is to be replaced to use snd_ctl_remove_user_ctl(). This function does
proper permission checks as well as decrements user_ctl_count after the control
has been removed.
Note that by using snd_ctl_remove_user_ctl() the check which returns -EBUSY at
beginning of the function if the control already exists is removed. This is not
a problem though since the check is quite useless, because the lock that is
protecting the control list is released between the check and before adding the
new control to the list, which means that it is possible that a different
control with the same settings is added to the list after the check. Luckily
there is another check that is done while holding the lock in snd_ctl_add(), so
we'll rely on that to make sure that the same control is not added twice.
A control that is visible on the card->controls list can be freed at any time.
This means we must not access any of its memory while not holding the
controls_rw_lock. Otherwise we risk a use after free access.
The l2tp [get|set]sockopt() code has fallen back to the UDP functions
for socket option levels != SOL_PPPOL2TP since day one, but that has
never actually worked, since the l2tp socket isn't an inet socket.
As David Miller points out:
"If we wanted this to work, it'd have to look up the tunnel and then
use tunnel->sk, but I wonder how useful that would be"
Since this can never have worked so nobody could possibly have depended
on that functionality, just remove the broken code and return -EINVAL.
Reported-by: Sasha Levin <sasha.levin@oracle.com> Acked-by: James Chapman <jchapman@katalix.com> Acked-by: David Miller <davem@davemloft.net> Cc: Phil Turnbull <phil.turnbull@oracle.com> Cc: Vegard Nossum <vegard.nossum@oracle.com> Cc: Willy Tarreau <w@1wt.eu> Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
[geissert: adjust file paths and context for 2.6.32]
[wt: fixes CVE-2014-4943] Signed-off-by: Willy Tarreau <w@1wt.eu>
Andy Lutomirski [Sun, 23 Nov 2014 02:00:33 +0000 (18:00 -0800)]
x86_64, traps: Rework bad_iret
It's possible for iretq to userspace to fail. This can happen because
of a bad CS, SS, or RIP.
Historically, we've handled it by fixing up an exception from iretq to
land at bad_iret, which pretends that the failed iret frame was really
the hardware part of #GP(0) from userspace. To make this work, there's
an extra fixup to fudge the gs base into a usable state.
This is suboptimal because it loses the original exception. It's also
buggy because there's no guarantee that we were on the kernel stack to
begin with. For example, if the failing iret happened on return from an
NMI, then we'll end up executing general_protection on the NMI stack.
This is bad for several reasons, the most immediate of which is that
general_protection, as a non-paranoid idtentry, will try to deliver
signals and/or schedule from the wrong stack.
This patch throws out bad_iret entirely. As a replacement, it augments
the existing swapgs fudge into a full-blown iret fixup, mostly written
in C. It's should be clearer and more correct.
Signed-off-by: Andy Lutomirski <luto@amacapital.net> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit b645af2d5905c4e32399005b867987919cbfc3ae)
[wt: notes for backport to 2.6.32:
- _ASM_EXTABLE was open-coded.
- removed unneeded CFI_ENDPROC
- removed __visible (introduced in 2.6.37-rc1, not needed here)
/wt] Signed-off-by: Willy Tarreau <w@1wt.eu>
Andy Lutomirski [Sun, 23 Nov 2014 02:00:31 +0000 (18:00 -0800)]
x86_64, traps: Fix the espfix64 #DF fixup and rewrite it in C
There's nothing special enough about the espfix64 double fault fixup to
justify writing it in assembly. Move it to C.
This also fixes a bug: if the double fault came from an IST stack, the
old asm code would return to a partially uninitialized stack frame.
Fixes: 3891a04aafd668686239349ea58f3314ea2af86b Signed-off-by: Andy Lutomirski <luto@amacapital.net> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit af726f21ed8af2cdaa4e93098dc211521218ae65)
[wt: backport notes for 2.6.32 :
- Adaptations to entry_64.S in declaration of do_double_fault.
- no exception_enter() in 2.6.32. Seems to be only for context tracking.
/wt] Signed-off-by: Willy Tarreau <w@1wt.eu>
Andy Lutomirski [Sun, 23 Nov 2014 02:00:32 +0000 (18:00 -0800)]
x86_64, traps: Stop using IST for #SS
On a 32-bit kernel, this has no effect, since there are no IST stacks.
On a 64-bit kernel, #SS can only happen in user code, on a failed iret
to user space, a canonical violation on access via RSP or RBP, or a
genuine stack segment violation in 32-bit kernel code. The first two
cases don't need IST, and the latter two cases are unlikely fatal bugs,
and promoting them to double faults would be fine.
This fixes a bug in which the espfix64 code mishandles a stack segment
violation.
This saves 4k of memory per CPU and a tiny bit of code.
Signed-off-by: Andy Lutomirski <luto@amacapital.net> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit 6f442be2fb22be02cafa606f1769fa1e6f894441)
[wt: no CONFIG_TRACING on 2.6.32, Fixes CVE-2014-9090] Signed-off-by: Willy Tarreau <w@1wt.eu>
init_espfix_ap() is currently off by one level when informing hypervisor
that allocated pages will be used for ministacks' page tables.
The most immediate effect of this on a PV guest is that if
'stack_page = __get_free_page()' returns a non-zeroed-out page the hypervisor
will refuse to use it for a page table (which it shouldn't be anyway). This will
result in warnings by both Xen and Linux.
More importantly, a subsequent write to that page (again, by a PV guest) is
likely to result in fatal page fault.
This moves the espfix64 logic into native_iret. To make this work,
it gets rid of the native patch for INTERRUPT_RETURN:
INTERRUPT_RETURN on native kernels is now 'jmp native_iret'.
This changes the 16-bit SS behavior on Xen from OOPSing to leaking
some bits of the Xen hypervisor's RSP (I think).
[ hpa: this is a nonzero cost on native, but probably not enough to
measure. Xen needs to fix this in their own code, probably doing
something equivalent to espfix64. ]
Embedded systems, which may be very memory-size-sensitive, are
extremely unlikely to ever encounter any 16-bit software, so make it
a CONFIG_EXPERT option to turn off support for any 16-bit software
whatsoever.
Signed-off-by: H. Peter Anvin <hpa@zytor.com> Link: http://lkml.kernel.org/r/1398816946-3351-1-git-send-email-hpa@linux.intel.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from 3.2 commit 70d87cbbd92a3611655b39003176ee1033796bf7)
[wt: backport notes for 2.6.32 :
- Fixed arch/x86/kernel/ldt.c (no IS_ENABLED on 2.6.32).
- No CONFIG_EXPERT condition in 2.6.32.
/wt] Signed-off-by: Willy Tarreau <w@1wt.eu>
Make espfix64 a hidden Kconfig option. This fixes the x86-64 UML
build which had broken due to the non-existence of init_espfix_bsp()
in UML: since UML uses its own Kconfig, this option does not appear in
the UML build.
This also makes it possible to make support for 16-bit segments a
configuration option, for the people who want to minimize the size of
the kernel.
Sparse warns that the percpu variables aren't declared before they are
defined. Rather than hacking around it, move espfix definitions into
a proper header file.
Reported-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: H. Peter Anvin <hpa@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from 3.2 commit 62358ee6bb9d3d9dd4761d39ac0e1ede9ba70b0e)
[wt: using DECLARE_PER_CPU instead of DECLARE_PER_CPU_READ_MOSTLY] Signed-off-by: Willy Tarreau <w@1wt.eu>
projects / thirdparty / kernel / stable.git / log
