RDS_TCP_DEFAULT_BUFSIZE has been unused since commit 1edd6a14d24f
("RDS-TCP: Do not bloat sndbuf/rcvbuf in rds_tcp_tune").
Signed-off-by: Sowmini Varadhan <sowmini.varadhan@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net> Reviewed-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Upstream commit c6a58ffed536 ("RDS: TCP: Add sysctl tunables for
sndbuf/rcvbuf on rds-tcp socket")
Add per-net sysctl tunables to set the size of sndbuf and
rcvbuf on the kernel tcp socket.
The tunables are added at /proc/sys/net/rds/tcp/rds_tcp_sndbuf
and /proc/sys/net/rds/tcp/rds_tcp_rcvbuf.
These values must be set before accept() or connect(),
and there may be an arbitrary number of existing rds-tcp
sockets when the tunable is modified. To make sure that all
connections in the netns pick up the same value for the tunable,
we reset existing rds-tcp connections in the netns, so that
they can reconnect with the new parameters.
Signed-off-by: Sowmini Varadhan <sowmini.varadhan@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net> Reviewed-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Reviewed-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Wei Lin Guay [Wed, 27 Jan 2016 12:18:08 +0000 (13:18 +0100)]
RDS: add flow control info to rds_info_rdma_connection
Added per connection flow_ctl_post_credit and
flow_ctl_send_credit to rds-info. These info help
in debugging RDS IB flow control. The newly added
attributes are placed at the bottom of the data
structure to ensure backward compatibility.
Wei Lin Guay [Thu, 17 Dec 2015 08:34:33 +0000 (09:34 +0100)]
RDS: update IB flow control algorithm
The current algorithm that uses 16 as a hard-coded value
in rds_ib_advertise_credits() doesn't serve the purpose, as
post_recvs() are performed in bulk. Thus, the test
condition will always be true.
This patch moves rds_ib_advertise_credits() in to the
post_recvs() loop. Instead of updating the post_recv credits
after all the post_recvs() have completed, the post_recv
credit is being updated in log2 incremental manner.
The proposed exponential quadrupling algorithm serves as a
good compromise between early start of the peer and at the
same time reducing the amount of explicit ACKs. The credit
update explicit ACKs will be generated starting from 16,
256, 4096...etc.
The performance number below shows that this new flow
control algorithm has minimal impact performance even though
it requires additional explicit ACKs.
IB flow control is always disabled regardless of
rds_ib_sysctl_flow_control flag.
The issue is that the initial credit advertisement
annouces zero credits, because ib_recv_refill() has
not yet been called. An initial credit offering
of zero effectively disables flow control.
IB flow control is only enabled if both active and
passive connections have set the rds_ib_sysctl
flow_control flag. E.g,
Conn. A (on), Conn. B (on) = enable
Conn. A (off), Conn. B (on) = disable
Conn. A (on), Conn. B (off) = disable
Conn. A (off), Conn. B (off) = disable
Roger Pau Monné [Tue, 3 Nov 2015 16:40:43 +0000 (16:40 +0000)]
xen-blkback: read from indirect descriptors only once
Since indirect descriptors are in memory shared with the frontend, the
frontend could alter the first_sect and last_sect values after they have
been validated but before they are recorded in the request. This may
result in I/O requests that overflow the foreign page, possibly
overwriting local pages when the I/O request is executed.
When parsing indirect descriptors, only read first_sect and last_sect
once.
This is part of XSA155.
(cherry-pick from 18779149101c0dd43ded43669ae2a92d21b6f9cb) CC: stable@vger.kernel.org Signed-off-by: Roger Pau Monné <roger.pau@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
----
v2: This is against v4.3
mlx4_core: scale_profile should work without params set to 0
The "scale_profile" parameter is to be used to do scaling of
hca params without requiring specific tuning setting for each
one of them individually and yet allowing manual setting of
variables.
In UEK2, the module params were default zero and initialized
later from a "driver default" or a scale_profile dictated value.
In UEK4 (derived from Mellanox OFED 2.4) module params were
pre-initialized to default values and are not zero and have
to be forced to 0 for dynamic scaling to be activated.
This defeats the purpose of having a single parameter to achieve
scaling and not requiring setting individual parameters (while
retaining ability to revert to driver defaults).
The changes here (re)introduce a separate static instance
containing default parameters separate from module parameters
which are pre-initialized to zero for parameters that can scale
dynamically and to default values for others. The zero module
parameters are later initialized to either a scale_profile
governed value or driver defaults.
When RAC tries to scale RDS-TCP, they are hitting bottlenecks
due to inefficiencies in rds_bind_lookup. Each call to
rds_bind_lookup results in an irqsave/irqrestore sequence, and
when the list of RDS sockets is large, we end up having IRQs
suppressed for long intervals. This trigger flow-control assertions
and causes TX queue watchdog hangs in the sender. The current
implementation makes this even worse, by superfluously calling
rds_bind_lookup(). This patch set takes the first step to solving
this problem by avoiding one of the redundant calls to rds_bind_lookup.
When errors such as connection hangs or failures are encountered
over RDS-TCP, the sending RDS, in an attempt at HA, will try to
reconnect, and trip up on all sorts of data structures intended
for ToS support. The ToS feature is currently only supported for
RDS-IB, and unplanned/untested usage of these data
structures by RDS-TCP causes deadlocks and panics.
Until we properly design, support, and test the ToS feature for
RDS-TCP, such paths should not be wandered into. Thus this patchset
adds defensive checks to ignore rs_tos settings in rds_sendmsg() for
TCP transports, and prevents the sending of ToS heartbeat pings
Until we properly design, support, and test the ToS feature for
RDS-TCP, such paths should not be wandered into. Thus this patchset
adds defensive checks to ignore rs_tos settings in rds_sendmsg() for
TCP transports, and prevents the sending of ToS heartbeat pings
in rds_send_hb() for TCP transport.
For reference, the deadlock that can be encountered in the
hb ping path is:
shamir rabinovitch [Wed, 16 Mar 2016 13:57:19 +0000 (09:57 -0400)]
rds: rds-stress show all zeros after few minutes
Issue can be seen on platforms that use 8K and above page size
while rds fragment size is 4K. On those platforms single page is
shared between 2 or more rds fragments. Each fragment has it's own
offeset and rds cong map code need to take this offset to account.
Not taking this offset to account lead to reading the data fragment
as congestion map fragment and hang of the rds transmit due to far
cong map corruption.
Two different threads with different rds sockets may be in
rds_recv_rcvbuf_delta() via receive path. If their ports
both map to the same word in the congestion map, then
using non-atomic ops to update it could cause the map to
be incorrect. Lets use atomics to avoid such an issue.
Full credit to Wengang <wen.gang.wang@oracle.com> for
finding the issue, analysing it and also pointing out
to offending code with spin lock based fix.
Signed-off-by: Wengang Wang <wen.gang.wang@oracle.com> Signed-off-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
rds_fmr_flush workqueue is calling ib_unmap_fmr
to invalidate a list of FMRs. Today, this workqueue
can be scheduled at any CPUs. In a NUMA-aware system,
schedule this workqueue to run on a CPU core closer to
ib_device can improve performance. As for now, we use
"sequential-low" policy. This policy selects two lower
cpu cores closer to HCA. In a non-NUMA aware system,
schedule rds_fmr_flush workqueue in a fixed cpu core
improves performance.
The mapping of cpu to the rds_fmr_flush workqueue
can be enabled/disabled via sysctl and it is enable
by default. To disable the feature, use below sysctl.
rds_ib_sysctl_disable_unmap_fmr_cpu = 1
Putting down some of the rds-stress performance number
comparing default and sequential-low policy in a NUMA
system with Oracle M4 QDR and Mellanox CX3.
rds-stress 4 conns, 32 threads, 16 depths, RDMA write
and unidirectional (higher is better).
Santosh Shilimkar [Wed, 21 Oct 2015 23:47:28 +0000 (16:47 -0700)]
RDS: IB: support larger frag size up to 16KB
Infiniband (IB) transport supports larger message size
than RDS_FRAG_SIZE, which is usually in 4KB PAGE_SIZE.
Nevertheless, RDS always fragments each payload into
RDS_FRAG_SIZE before hands it over to the underlying
IB HCA.
One of the important message size required for database
is 8448 (8K + 256B control message) for BCOPY. This RDS
message, even with IB transport, will generate three
IB work requests (WR) with each having its own RDS header.
This series of patches improve RDS performance by allowing
IB transport to send/receive RDS message with a larger
RDS_FRAG_SIZE (Ideally, using a single WR).
In order to maintain the backward compatibility and
interoperability between various RDS versions, and at
the same time to support various FRAG_SIZE, the IB
fragment size is negotiated per connection.
Although IB is capable of supporting 4GB of message size,
currently we limit the IB RDS_FRAG_SIZE up to 16KB due to
two reasons:-
1. This is needed for current 8448 RDS message size usecase.
2. Minizing the size for each receive queue entry in order
to optimal memory usage.
In term of implementation, The 'dp_reserved2' field of
'struct rds_ib_connect_private' now carries information about
supported IB fragment size. Since we are just
using the IB connection private data and a reserved field,
the protocol version is not bumped up. Furthermore, the feature
is enabled only for RDS_PROTOCOL_v4.1 and above (future).
To keep thing simpler for user, a module parameter
'rds_ib_max_frag' is provided. Without module parameter,
the default PAGE_SIZE frag will be used. During the connection
establishment, the smallest fragment size will be
chosen. If the fragment size is 0, it means RDS module
doesn't support large fragment size and the default
RDS_FRAG_SIZE will be used.
Upto ~10+ % improvement seen with Orion and ~9+ % with RDBMS
update queries.
Santosh Shilimkar [Mon, 21 Mar 2016 06:24:32 +0000 (23:24 -0700)]
RDS: IB: purge receive frag cache on connection shutdown
RDS IB connections can be formed with different fragment size across
reconnect and hence the current frag cache needs to be purged to
avoid stale frag usages.
Santosh Shilimkar [Wed, 4 Nov 2015 21:42:39 +0000 (13:42 -0800)]
RDS: IB: scale rds_ib_allocation based on fragment size
The 'rds_ib_sysctl_max_recv_allocation' allocation is used to manage
and allocate the size of IB receive queue entry (RQE) for each IB
connection. However, it relies on the hardcoded RDS_FRAG_SIZE.
Lets make it scalable based on supported fragment sizes for different
IB connection. Each connection can allocate different RQE size
depending on the per connection fragment_size.
Santosh Shilimkar [Wed, 21 Oct 2015 23:47:28 +0000 (16:47 -0700)]
RDS: IB: make fragment size (RDS_FRAG_SIZE) dynamic
IB fabric is capable of fragment 4GB of data payload into
send_first, send_middle and send_last. Nevertheless,
RDS fragments each payload into PAGE_SIZE, which is usually
4KB. This patch makes the RDS_FRAG_SIZE for RDS IB transport
dynamic.
In the preperation for subsequent patch(es), this patch
adds per connection peer negotiation to determine the
supported fragment size for IB transport.
Orabug: 21894138 Reviewed-by: Wei Lin Guay <wei.lin.guay@oracle.com> Signed-off-by: Wei Lin Guay <wei.lin.guay@oracle.com> Signed-off-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Wei Lin Guay [Fri, 20 Nov 2015 23:14:48 +0000 (15:14 -0800)]
RDS: fix the sg allocation based on actual message size
Fix an issue where only PAGE_SIZE bytes are allocated per
scatter-gather entry (SGE) regardless of the actual message
size: Furthermore, use buddy memory allocation technique to
allocate/free memmory (if possible) to reduce SGE.
Santosh Shilimkar [Mon, 16 Nov 2015 21:28:11 +0000 (13:28 -0800)]
RDS: make congestion code independent of PAGE_SIZE
RDS congestion map code is designed with base assumption of
4K page size. The map update as well transport code assumes
it that way. Ofcourse it breaks when transport like IB starts
supporting larger fragments than 4K.
To overcome this limitation without too many changes to the core
congestion map update logic, define indepedent RDS_CONG_PAGE_SIZE
and use it.
While at it we also move rds_message_map_pages() whose sole
purpose it to map congestion pages to congestion code.
Santosh Shilimkar [Tue, 16 Feb 2016 18:24:31 +0000 (10:24 -0800)]
RDS: Back out OoO send status fix since it causes the regression
With the DB build, the crash was observed which was boiled down to
this change on UEK2. Proactively we back this out on UEK4 as well
till the issue gets addresssed.
net/mlx4_core: Modify default value of log_rdmarc_per_qp to be consistent with HW capability
This value is used to calculate max_qp_dest_rdma.
Default value of 4 brings us to 16 while HW supports 128
(max_requester_per_qp)
Although this value can be changed by module param it is best that default
will be optimized
Roman Gushchin [Mon, 12 Oct 2015 13:33:44 +0000 (16:33 +0300)]
fuse: break infinite loop in fuse_fill_write_pages()
I got a report about unkillable task eating CPU. Further
investigation shows, that the problem is in the fuse_fill_write_pages()
function. If iov's first segment has zero length, we get an infinite
loop, because we never reach iov_iter_advance() call.
Fix this by calling iov_iter_advance() before repeating an attempt to
copy data from userspace.
A similar problem is described in 124d3b7041f ("fix writev regression:
pan hanging unkillable and un-straceable"). If zero-length segmend
is followed by segment with invalid address,
iov_iter_fault_in_readable() checks only first segment (zero-length),
iov_iter_copy_from_user_atomic() skips it, fails at second and
returns zero -> goto again without skipping zero-length segment.
Patch calls iov_iter_advance() before goto again: we'll skip zero-length
segment at second iteraction and iov_iter_fault_in_readable() will detect
invalid address.
Special thanks to Konstantin Khlebnikov, who helped a lot with the commit
description.
Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Maxim Patlasov <mpatlasov@parallels.com> Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru> Signed-off-by: Roman Gushchin <klamm@yandex-team.ru> Signed-off-by: Miklos Szeredi <miklos@szeredi.hu> Fixes: ea9b9907b82a ("fuse: implement perform_write") Cc: <stable@vger.kernel.org>
(cherry picked from commit 3ca8138f014a913f98e6ef40e939868e1e9ea876)
Alexey Dobriyan [Thu, 25 Jun 2015 22:00:54 +0000 (15:00 -0700)]
proc: fix PAGE_SIZE limit of /proc/$PID/cmdline
/proc/$PID/cmdline truncates output at PAGE_SIZE. It is easy to see with
$ cat /proc/self/cmdline $(seq 1037) 2>/dev/null
However, command line size was never limited to PAGE_SIZE but to 128 KB
and relatively recently limitation was removed altogether.
People noticed and ask questions:
http://stackoverflow.com/questions/199130/how-do-i-increase-the-proc-pid-cmdline-4096-byte-limit
seq file interface is not OK, because it kmalloc's for whole output and
open + read(, 1) + sleep will pin arbitrary amounts of kernel memory. To
not do that, limit must be imposed which is incompatible with arbitrary
sized command lines.
I apologize for hairy code, but this it direct consequence of command line
layout in memory and hacks to support things like "init [3]".
The loops are "unrolled" otherwise it is either macros which hide control
flow or functions with 7-8 arguments with equal line count.
There should be real setproctitle(2) or something.
[akpm@linux-foundation.org: fix a billion min() warnings] Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com> Tested-by: Jarod Wilson <jarod@redhat.com> Acked-by: Jarod Wilson <jarod@redhat.com> Cc: Cyrill Gorcunov <gorcunov@openvz.org> Cc: Jan Stancek <jstancek@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Orabug: 23093364
Mainline commit c2c0bb44620dece7ec97e28167e32c343da22867 Acked-by: Chuck Anderson <chuck.anderson@oracle.com>
The problem comes with the fact that many such jobs (for the same device)
are being scheduled simultaneously. While scsi_remove_device() uses
shost->scan_mutex and scsi_device_lookup() will fail for a device in
SDEV_DEL state there is no protection against someone who did
scsi_device_lookup() before we actually entered __scsi_remove_device(). So
the whole scenario looks like that: two callers do simultaneous (or
preemption happens) calls to scsi_device_lookup() ant these calls succeed
for both of them, after that they try doing scsi_remove_device().
shost->scan_mutex only serializes their calls to __scsi_remove_device()
and we end up doing the cleanup path twice.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
(cherry picked from commit be821fd8e62765de43cc4f0e2db363d0e30a7e9b)
Orabug: 23021563 Signed-off-by: Jason Luo <zhangqing.luo@oracle.com>
David S. Miller [Mon, 14 Mar 2016 03:28:00 +0000 (23:28 -0400)]
ipv4: Don't do expensive useless work during inetdev destroy.
When an inetdev is destroyed, every address assigned to the interface
is removed. And in this scenerio we do two pointless things which can
be very expensive if the number of assigned interfaces is large:
1) Address promotion. We are deleting all addresses, so there is no
point in doing this.
2) A full nf conntrack table purge for every address. We only need to
do this once, as is already caught by the existing
masq_dev_notifier so masq_inet_event() can skip this.
Reported-by: Solar Designer <solar@openwall.com> Signed-off-by: David S. Miller <davem@davemloft.net> Tested-by: Cyrill Gorcunov <gorcunov@openvz.org>
(cherry picked from commit fbd40ea0180a2d328c5adc61414dc8bab9335ce2)
Filipe Manana [Sat, 20 Jun 2015 17:20:09 +0000 (18:20 +0100)]
Btrfs: fix shrinking truncate when the no_holes feature is enabled
If the no_holes feature is enabled, we attempt to shrink a file to a size
that ends up in the middle of a hole and we don't have any file extent
items in the fs/subvol tree that go beyond the new file size (or any
ordered extents that will insert such file extent items), we end up not
updating the inode's disk_i_size, we only update the inode's i_size.
This means that after unmounting and mounting the filesystem, or after
the inode is evicted and reloaded, its i_size ends up being incorrect
(an inode's i_size is set to the disk_i_size field when an inode is
loaded). This happens when btrfs_truncate_inode_items() doesn't find
any file extent items to drop - in this case it never makes a call to
btrfs_ordered_update_i_size() in order to update the inode's disk_i_size.
Example reproducer:
$ mkfs.btrfs -O no-holes -f /dev/sdd
$ mount /dev/sdd /mnt
# Create our test file with some data and durably persist it.
$ xfs_io -f -c "pwrite -S 0xaa 0 128K" /mnt/foo
$ sync
# Append some data to the file, increasing its size, and leave a hole
# between the old size and the start offset if the following write. So
# our file gets a hole in the range [128Kb, 256Kb[.
$ xfs_io -c "truncate 160K" /mnt/foo
# We expect to see our file with a size of 160Kb, with the first 128Kb
# of data all having the value 0xaa and the remaining 32Kb of data all
# having the value 0x00.
$ od -t x1 /mnt/foo 0000000 aa aa aa aa aa aa aa aa aa aa aa aa aa aa aa aa
* 0400000 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
* 0500000
# Now cleanly unmount and mount again the filesystem.
$ umount /mnt
$ mount /dev/sdd /mnt
# We expect to get the same result as before, a file with a size of
# 160Kb, with the first 128Kb of data all having the value 0xaa and the
# remaining 32Kb of data all having the value 0x00.
$ od -t x1 /mnt/foo 0000000 aa aa aa aa aa aa aa aa aa aa aa aa aa aa aa aa
* 0400000
In the example above the file size/data do not match what they were before
the remount.
Fix this by always calling btrfs_ordered_update_i_size() with a size
matching the size the file was truncated to if btrfs_truncate_inode_items()
is not called for a log tree and no file extent items were dropped. This
ensures the same behaviour as when the no_holes feature is not enabled.
Even though we delay the rename of directories when they become
descendents of other directories that were also renamed in the send
root to prevent infinite path build loops, we were doing it in cases
where this was not needed and was actually harmful resulting in
infinite path build loops as we ended up with a circular dependency
of delayed directory renames.
Consider the following reproducer:
$ mkfs.btrfs -f /dev/sdb
$ mount /dev/sdb /mnt
$ mkfs.btrfs -f /dev/sdc
$ mount /dev/sdc /mnt2
This reproducer resulted in an infinite path build loop when building the
path for inode 266 because the following circular dependency of delayed
directory renames was created:
Where the notation "X <- Y" means the rename of inode X is delayed by the
rename of inode Y (X will be renamed after Y is renamed). This resulted
in an infinite path build loop of inode 266 because that inode has inode
261 as an ancestor in the send root and inode 261 is in the circular
dependency of delayed renames listed above.
Fix this by not delaying the rename of a directory inode if an ancestor of
the inode in the send root, which has a delayed rename operation, is not
also a descendent of the inode in the parent root.
Thanks to Robbie Ko for sending the reproducer example.
A test case for xfstests follows soon.
Reported-by: Robbie Ko <robbieko@synology.com> Signed-off-by: Filipe Manana <fdmanana@suse.com>
(cherry picked from mainline commit 80aa6027561eef12b49031d46fd6724daf1e7fb6) Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
We have another case where after an fsync log replay we get an inode with
a wrong link count (smaller than it should be) and a number of directory
entries greater than its link count. This happens when we add a new link
hard link to our inode A and then we fsync some other inode B that has
the side effect of logging the parent directory inode too. In this case
at log replay time we add the new hard link to our inode (the item with
key BTRFS_INODE_REF_KEY) when processing the parent directory but we
never adjust the link count of our inode A. As a result we get stale dir
entries for our inode A that can never be deleted and therefore it makes
it impossible to remove the parent directory (as its i_size can never
decrease back to 0).
A simple reproducer for fstests that triggers this issue:
seq=`basename $0`
seqres=$RESULT_DIR/$seq
echo "QA output created by $seq"
tmp=/tmp/$$
status=1 # failure is the default!
trap "_cleanup; exit \$status" 0 1 2 3 15
_cleanup()
{
_cleanup_flakey
rm -f $tmp.*
}
# get standard environment, filters and checks
. ./common/rc
. ./common/filter
. ./common/dmflakey
# real QA test starts here
_need_to_be_root
_supported_fs generic
_supported_os Linux
_require_scratch
_require_dm_flakey
_require_metadata_journaling $SCRATCH_DEV
# Create our test directory and files.
mkdir $SCRATCH_MNT/testdir
touch $SCRATCH_MNT/testdir/foo
touch $SCRATCH_MNT/testdir/bar
# Make sure everything done so far is durably persisted.
sync
# Create one hard link for file foo and another one for file bar. After
# that fsync only the file bar.
ln $SCRATCH_MNT/testdir/bar $SCRATCH_MNT/testdir/bar_link
ln $SCRATCH_MNT/testdir/foo $SCRATCH_MNT/testdir/foo_link
$XFS_IO_PROG -c "fsync" $SCRATCH_MNT/testdir/bar
# Silently drop all writes on scratch device to simulate power failure.
_load_flakey_table $FLAKEY_DROP_WRITES
_unmount_flakey
# Allow writes again and mount the fs to trigger log/journal replay.
_load_flakey_table $FLAKEY_ALLOW_WRITES
_mount_flakey
# Now verify both our files have a link count of 2.
echo "Link count for file foo: $(stat --format=%h $SCRATCH_MNT/testdir/foo)"
echo "Link count for file bar: $(stat --format=%h $SCRATCH_MNT/testdir/bar)"
# We should be able to remove all the links of our files in testdir, and
# after that the parent directory should become empty and therefore
# possible to remove it.
rm -f $SCRATCH_MNT/testdir/*
rmdir $SCRATCH_MNT/testdir
_unmount_flakey
# The fstests framework will call fsck against our filesystem which will verify
# that all metadata is in a consistent state.
status=0
exit
The test fails with:
-Link count for file foo: 2
+Link count for file foo: 1
Link count for file bar: 2
+rm: cannot remove '/home/fdmanana/btrfs-tests/scratch_1/testdir/foo_link': Stale file handle
+rmdir: failed to remove '/home/fdmanana/btrfs-tests/scratch_1/testdir': Directory not empty
(...)
_check_btrfs_filesystem: filesystem on /dev/sdc is inconsistent
And fsck's output:
(...)
checking fs roots
root 5 inode 258 errors 2001, no inode item, link count wrong
unresolved ref dir 257 index 5 namelen 8 name foo_link filetype 1 errors 4, no inode ref
Checking filesystem on /dev/sdc
(...)
So fix this by marking inodes for link count fixup at log replay time
whenever a directory entry is replayed if the entry was created in the
transaction where the fsync was made and if it points to a non-directory
inode.
This isn't a new problem/regression, the issue exists for a long time,
possibly since the log tree feature was added (2008).
Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
(cherry picked from mainline commit bb53eda9029fd52b466fa501ba4aa58e94789b18) Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
When we call pci_scan_root_bus we would pass in 'sd' (sysdata)
pointer which was an 'pcifront_sd' structure. However in the
pci_device_add it expects that the 'sd' is 'struct sysdata' and
sets the dev->node to what is in sd->node (offset 4):
That is an hole - filled with garbage as we used kmalloc instead of
kzalloc (the second problem).
This patch fixes the issue by:
1) Use kzalloc to initialize to a well known state.
2) Put 'struct pci_sysdata' at the start of 'pcifront_sd'. That
way access to the 'node' will access the right offset.
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: <stable@vger.kernel.org> Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 4d8c8bd6f2062c9988817183a91fe2e623c8aa5e) Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
OraBug: 23017418 - Backport Linux v4.4 Xen patches
Konrad Rzeszutek Wilk [Thu, 11 Feb 2016 21:10:24 +0000 (16:10 -0500)]
xen/pciback: Save the number of MSI-X entries to be copied later.
Commit 8135cf8b092723dbfcc611fe6fdcb3a36c9951c5 (xen/pciback: Save
xen_pci_op commands before processing it) broke enabling MSI-X because
it would never copy the resulting vectors into the response. The
number of vectors requested was being overwritten by the return value
(typically zero for success).
Save the number of vectors before processing the op, so the correct
number of vectors are copied afterwards.
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: <stable@vger.kernel.org> Reviewed-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit d159457b84395927b5a52adb72f748dd089ad5e5) Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
OraBug: 23017418 - Backport Linux v4.4 Xen patches
Konrad Rzeszutek Wilk [Thu, 11 Feb 2016 21:10:23 +0000 (16:10 -0500)]
xen/pciback: Check PF instead of VF for PCI_COMMAND_MEMORY
Commit 408fb0e5aa7fda0059db282ff58c3b2a4278baa0 (xen/pciback: Don't
allow MSI-X ops if PCI_COMMAND_MEMORY is not set) prevented enabling
MSI-X on passed-through virtual functions, because it checked the VF
for PCI_COMMAND_MEMORY but this is not a valid bit for VFs.
Instead, check the physical function for PCI_COMMAND_MEMORY.
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: <stable@vger.kernel.org> Reviewed-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 8d47065f7d1980dde52abb874b301054f3013602) Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
OraBug: 23017418 - Backport Linux v4.4 Xen patches
Konrad Rzeszutek Wilk [Tue, 29 Mar 2016 14:55:42 +0000 (10:55 -0400)]
Merge branch 'linux-4.1/4.4-xen-backport' of git://ca-git.us.oracle.com/linux-joaomart-public into uek4/4.4-xen-backport
* 'linux-4.1/4.4-xen-backport' of git://ca-git.us.oracle.com/linux-joaomart-public: (113 commits)
arch/x86/xen/suspend.c: include xen/xen.h
x86/paravirt: Prevent rtc_cmos platform device init on PV guests
xen-pciback: fix up cleanup path when alloc fails
xen/pciback: Don't allow MSI-X ops if PCI_COMMAND_MEMORY is not set.
xen/pciback: For XEN_PCI_OP_disable_msi[|x] only disable if device has MSI(X) enabled.
xen/pciback: Do not install an IRQ handler for MSI interrupts.
xen/pciback: Return error on XEN_PCI_OP_enable_msix when device has MSI or MSI-X enabled
xen/pciback: Return error on XEN_PCI_OP_enable_msi when device has MSI or MSI-X enabled
xen/pciback: Save xen_pci_op commands before processing it
xen-scsiback: safely copy requests
xen-blkback: read from indirect descriptors only once
xen-blkback: only read request operation from shared ring once
xen-netback: use RING_COPY_REQUEST() throughout
xen-netback: don't use last request to determine minimum Tx credit
xen: Add RING_COPY_REQUEST()
xen/x86/pvh: Use HVM's flush_tlb_others op
xen: Resume PMU from non-atomic context
xen/events/fifo: Consume unprocessed events when a CPU dies
xen/evtchn: dynamically grow pending event channel ring
xen/gntdev: Grant maps should not be subject to NUMA balancing
...
Backport from Linux v4.4
OraBug: 23017418 - Backport Linux v4.4 Xen patches
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Andrew Morton [Tue, 29 Dec 2015 22:54:13 +0000 (14:54 -0800)]
arch/x86/xen/suspend.c: include xen/xen.h
Fix the build warning:
arch/x86/xen/suspend.c: In function 'xen_arch_pre_suspend':
arch/x86/xen/suspend.c:70:9: error: implicit declaration of function 'xen_pv_domain' [-Werror=implicit-function-declaration]
if (xen_pv_domain())
^
Reported-by: kbuild test robot <fengguang.wu@intel.com> Cc: Sasha Levin <sasha.levin@oracle.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com> Cc: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit facca61683f937f31f90307cc64851436c8a3e21) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
David Vrabel [Fri, 11 Dec 2015 14:07:53 +0000 (09:07 -0500)]
x86/paravirt: Prevent rtc_cmos platform device init on PV guests
Adding the rtc platform device in non-privileged Xen PV guests causes
an IRQ conflict because these guests do not have legacy PIC and may
allocate irqs in the legacy range.
But hvc_console cannot get its interrupt because it is already in use
by rtc0 and the console does not work.
genirq: Flags mismatch irq 8. 00000000 (hvc_console) vs. 00000000 (rtc0)
We can avoid this problem by realizing that unprivileged PV guests (both
Xen and lguests) are not supposed to have rtc_cmos device and so
adding it is not necessary.
Privileged guests (i.e. Xen's dom0) do use it but they should not have
irq conflicts since they allocate irqs above legacy range (above
gsi_top, in fact).
Instead of explicitly testing whether the guest is privileged we can
extend pv_info structure to include information about guest's RTC
support.
Reported-and-tested-by: Sander Eikelenboom <linux@eikelenboom.it> Signed-off-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Cc: vkuznets@redhat.com Cc: xen-devel@lists.xenproject.org Cc: konrad.wilk@oracle.com Cc: stable@vger.kernel.org # 4.2+ Link: http://lkml.kernel.org/r/1449842873-2613-1-git-send-email-boris.ostrovsky@oracle.com Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
(cherry picked from commit d8c98a1d1488747625ad6044d423406e17e99b7a) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Doug Goldstein [Thu, 26 Nov 2015 20:32:39 +0000 (14:32 -0600)]
xen-pciback: fix up cleanup path when alloc fails
When allocating a pciback device fails, clear the private
field. This could lead to an use-after free, however
the 'really_probe' takes care of setting
dev_set_drvdata(dev, NULL) in its failure path (which we would
exercise if the ->probe function failed), so we we
are OK. However lets be defensive as the code can change.
Going forward we should clean up the pci_set_drvdata(dev, NULL)
in the various code-base. That will be for another day.
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Reported-by: Jonathan Creekmore <jonathan.creekmore@gmail.com> Signed-off-by: Doug Goldstein <cardoe@cardoe.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
(cherry picked from commit 584a561a6fee0d258f9ca644f58b73d9a41b8a46) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Konrad Rzeszutek Wilk [Mon, 2 Nov 2015 23:13:27 +0000 (18:13 -0500)]
xen/pciback: Don't allow MSI-X ops if PCI_COMMAND_MEMORY is not set.
commit f598282f51 ("PCI: Fix the NIU MSI-X problem in a better way")
teaches us that dealing with MSI-X can be troublesome.
Further checks in the MSI-X architecture shows that if the
PCI_COMMAND_MEMORY bit is turned of in the PCI_COMMAND we
may not be able to access the BAR (since they are memory regions).
Since the MSI-X tables are located in there.. that can lead
to us causing PCIe errors. Inhibit us performing any
operation on the MSI-X unless the MEMORY bit is set.
Note that Xen hypervisor with:
"x86/MSI-X: access MSI-X table only after having enabled MSI-X"
will return:
xen_pciback: 0000:0a:00.1: error -6 enabling MSI-X for guest 3!
When the generic MSI code tries to setup the PIRQ without
MEMORY bit set. Which means with later versions of Xen
(4.6) this patch is not neccessary.
This is part of XSA-157
CC: stable@vger.kernel.org Reviewed-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
(cherry picked from commit 408fb0e5aa7fda0059db282ff58c3b2a4278baa0) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
xen/pciback: For XEN_PCI_OP_disable_msi[|x] only disable if device has MSI(X) enabled.
Otherwise just continue on, returning the same values as
previously (return of 0, and op->result has the PIRQ value).
This does not change the behavior of XEN_PCI_OP_disable_msi[|x].
The pci_disable_msi or pci_disable_msix have the checks for
msi_enabled or msix_enabled so they will error out immediately.
However the guest can still call these operations and cause
us to disable the 'ack_intr'. That means the backend IRQ handler
for the legacy interrupt will not respond to interrupts anymore.
This will lead to (if the device is causing an interrupt storm)
for the Linux generic code to disable the interrupt line.
Naturally this will only happen if the device in question
is plugged in on the motherboard on shared level interrupt GSI.
This is part of XSA-157
CC: stable@vger.kernel.org Reviewed-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
(cherry picked from commit 7cfb905b9638982862f0331b36ccaaca5d383b49) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Konrad Rzeszutek Wilk [Mon, 2 Nov 2015 22:24:08 +0000 (17:24 -0500)]
xen/pciback: Do not install an IRQ handler for MSI interrupts.
Otherwise an guest can subvert the generic MSI code to trigger
an BUG_ON condition during MSI interrupt freeing:
for (i = 0; i < entry->nvec_used; i++)
BUG_ON(irq_has_action(entry->irq + i));
Xen PCI backed installs an IRQ handler (request_irq) for
the dev->irq whenever the guest writes PCI_COMMAND_MEMORY
(or PCI_COMMAND_IO) to the PCI_COMMAND register. This is
done in case the device has legacy interrupts the GSI line
is shared by the backend devices.
To subvert the backend the guest needs to make the backend
to change the dev->irq from the GSI to the MSI interrupt line,
make the backend allocate an interrupt handler, and then command
the backend to free the MSI interrupt and hit the BUG_ON.
Since the backend only calls 'request_irq' when the guest
writes to the PCI_COMMAND register the guest needs to call
XEN_PCI_OP_enable_msi before any other operation. This will
cause the generic MSI code to setup an MSI entry and
populate dev->irq with the new PIRQ value.
Then the guest can write to PCI_COMMAND PCI_COMMAND_MEMORY
and cause the backend to setup an IRQ handler for dev->irq
(which instead of the GSI value has the MSI pirq). See
'xen_pcibk_control_isr'.
Then the guest disables the MSI: XEN_PCI_OP_disable_msi
which ends up triggering the BUG_ON condition in 'free_msi_irqs'
as there is an IRQ handler for the entry->irq (dev->irq).
Note that this cannot be done using MSI-X as the generic
code does not over-write dev->irq with the MSI-X PIRQ values.
The patch inhibits setting up the IRQ handler if MSI or
MSI-X (for symmetry reasons) code had been called successfully.
P.S.
Xen PCIBack when it sets up the device for the guest consumption
ends up writting 0 to the PCI_COMMAND (see xen_pcibk_reset_device).
XSA-120 addendum patch removed that - however when upstreaming said
addendum we found that it caused issues with qemu upstream. That
has now been fixed in qemu upstream.
This is part of XSA-157
CC: stable@vger.kernel.org Reviewed-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
(cherry picked from commit a396f3a210c3a61e94d6b87ec05a75d0be2a60d0) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Konrad Rzeszutek Wilk [Mon, 2 Nov 2015 23:07:44 +0000 (18:07 -0500)]
xen/pciback: Return error on XEN_PCI_OP_enable_msix when device has MSI or MSI-X enabled
The guest sequence of:
a) XEN_PCI_OP_enable_msix
b) XEN_PCI_OP_enable_msix
results in hitting an NULL pointer due to using freed pointers.
The device passed in the guest MUST have MSI-X capability.
The a) constructs and SysFS representation of MSI and MSI groups.
The b) adds a second set of them but adding in to SysFS fails (duplicate entry).
'populate_msi_sysfs' frees the newly allocated msi_irq_groups (note that
in a) pdev->msi_irq_groups is still set) and also free's ALL of the
MSI-X entries of the device (the ones allocated in step a) and b)).
The unwind code: 'free_msi_irqs' deletes all the entries and tries to
delete the pdev->msi_irq_groups (which hasn't been set to NULL).
However the pointers in the SysFS are already freed and we hit an
NULL pointer further on when 'strlen' is attempted on a freed pointer.
The patch adds a simple check in the XEN_PCI_OP_enable_msix to guard
against that. The check for msi_enabled is not stricly neccessary.
This is part of XSA-157
CC: stable@vger.kernel.org Reviewed-by: David Vrabel <david.vrabel@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
(cherry picked from commit 5e0ce1455c09dd61d029b8ad45d82e1ac0b6c4c9) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
xen/pciback: Return error on XEN_PCI_OP_enable_msi when device has MSI or MSI-X enabled
The guest sequence of:
a) XEN_PCI_OP_enable_msi
b) XEN_PCI_OP_enable_msi
c) XEN_PCI_OP_disable_msi
results in hitting an BUG_ON condition in the msi.c code.
The MSI code uses an dev->msi_list to which it adds MSI entries.
Under the above conditions an BUG_ON() can be hit. The device
passed in the guest MUST have MSI capability.
The a) adds the entry to the dev->msi_list and sets msi_enabled.
The b) adds a second entry but adding in to SysFS fails (duplicate entry)
and deletes all of the entries from msi_list and returns (with msi_enabled
is still set). c) pci_disable_msi passes the msi_enabled checks and hits:
BUG_ON(list_empty(dev_to_msi_list(&dev->dev)));
and blows up.
The patch adds a simple check in the XEN_PCI_OP_enable_msi to guard
against that. The check for msix_enabled is not stricly neccessary.
This is part of XSA-157.
CC: stable@vger.kernel.org Reviewed-by: David Vrabel <david.vrabel@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
(cherry picked from commit 56441f3c8e5bd45aab10dd9f8c505dd4bec03b0d) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Konrad Rzeszutek Wilk [Mon, 16 Nov 2015 17:40:48 +0000 (12:40 -0500)]
xen/pciback: Save xen_pci_op commands before processing it
Double fetch vulnerabilities that happen when a variable is
fetched twice from shared memory but a security check is only
performed the first time.
The xen_pcibk_do_op function performs a switch statements on the op->cmd
value which is stored in shared memory. Interestingly this can result
in a double fetch vulnerability depending on the performed compiler
optimization.
This patch fixes it by saving the xen_pci_op command before
processing it. We also use 'barrier' to make sure that the
compiler does not perform any optimization.
This is part of XSA155.
CC: stable@vger.kernel.org Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Jan Beulich <JBeulich@suse.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
(cherry picked from commit 8135cf8b092723dbfcc611fe6fdcb3a36c9951c5) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Roger Pau Monné [Tue, 3 Nov 2015 16:40:43 +0000 (16:40 +0000)]
xen-blkback: read from indirect descriptors only once
Since indirect descriptors are in memory shared with the frontend, the
frontend could alter the first_sect and last_sect values after they have
been validated but before they are recorded in the request. This may
result in I/O requests that overflow the foreign page, possibly
overwriting local pages when the I/O request is executed.
When parsing indirect descriptors, only read first_sect and last_sect
once.
This is part of XSA155.
CC: stable@vger.kernel.org Signed-off-by: Roger Pau Monné <roger.pau@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
(cherry picked from commit 18779149101c0dd43ded43669ae2a92d21b6f9cb) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
David Vrabel [Fri, 30 Oct 2015 15:17:06 +0000 (15:17 +0000)]
xen-netback: use RING_COPY_REQUEST() throughout
Instead of open-coding memcpy()s and directly accessing Tx and Rx
requests, use the new RING_COPY_REQUEST() that ensures the local copy
is correct.
This is more than is strictly necessary for guest Rx requests since
only the id and gref fields are used and it is harmless if the
frontend modifies these.
This is part of XSA155.
CC: stable@vger.kernel.org Reviewed-by: Wei Liu <wei.liu2@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
(cherry picked from commit 68a33bfd8403e4e22847165d149823a2e0e67c9c) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
David Vrabel [Fri, 30 Oct 2015 15:16:01 +0000 (15:16 +0000)]
xen-netback: don't use last request to determine minimum Tx credit
The last from guest transmitted request gives no indication about the
minimum amount of credit that the guest might need to send a packet
since the last packet might have been a small one.
Instead allow for the worst case 128 KiB packet.
This is part of XSA155.
CC: stable@vger.kernel.org Reviewed-by: Wei Liu <wei.liu2@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
(cherry picked from commit 0f589967a73f1f30ab4ac4dd9ce0bb399b4d6357) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
David Vrabel [Fri, 30 Oct 2015 14:58:08 +0000 (14:58 +0000)]
xen: Add RING_COPY_REQUEST()
Using RING_GET_REQUEST() on a shared ring is easy to use incorrectly
(i.e., by not considering that the other end may alter the data in the
shared ring while it is being inspected). Safe usage of a request
generally requires taking a local copy.
Provide a RING_COPY_REQUEST() macro to use instead of
RING_GET_REQUEST() and an open-coded memcpy(). This takes care of
ensuring that the copy is done correctly regardless of any possible
compiler optimizations.
Use a volatile source to prevent the compiler from reordering or
omitting the copy.
This is part of XSA155.
CC: stable@vger.kernel.org Signed-off-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
(cherry picked from commit 454d5d882c7e412b840e3c99010fe81a9862f6fb) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Boris Ostrovsky [Sun, 13 Dec 2015 00:25:55 +0000 (19:25 -0500)]
xen/x86/pvh: Use HVM's flush_tlb_others op
Using MMUEXT_TLB_FLUSH_MULTI doesn't buy us much since the hypervisor
will likely perform same IPIs as would have the guest.
More importantly, using MMUEXT_INVLPG_MULTI may not to invalidate the
guest's address on remote CPU (when, for example, VCPU from another guest
is running there).
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Suggested-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 20f36e0380a7e871a711d5e4e59d04d4948326b4) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Boris Ostrovsky [Wed, 2 Dec 2015 17:10:48 +0000 (12:10 -0500)]
xen: Resume PMU from non-atomic context
Resuming PMU currently triggers a warning from ___might_sleep() (assuming
CONFIG_DEBUG_ATOMIC_SLEEP is set) when xen_pmu_init() allocates GFP_KERNEL
page because we are in state resembling atomic context.
Move resuming PMU to xen_arch_resume() which is called in regular context.
For symmetry move suspending PMU to xen_arch_suspend() as well.
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Reported-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: <stable@vger.kernel.org> # 4.3 Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit de0afc9bdeeadaa998797d2333c754bf9f4d5dcf) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Ross Lagerwall [Fri, 19 Jun 2015 15:15:57 +0000 (16:15 +0100)]
xen/events/fifo: Consume unprocessed events when a CPU dies
When a CPU is offlined, there may be unprocessed events on a port for
that CPU. If the port is subsequently reused on a different CPU, it
could be in an unexpected state with the link bit set, resulting in
interrupts being missed. Fix this by consuming any unprocessed events
for a particular CPU when that CPU dies.
Signed-off-by: Ross Lagerwall <ross.lagerwall@citrix.com> Cc: <stable@vger.kernel.org> # 3.14+ Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 3de88d622fd68bd4dbee0f80168218b23f798fd0) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
David Vrabel [Thu, 26 Nov 2015 16:14:35 +0000 (16:14 +0000)]
xen/evtchn: dynamically grow pending event channel ring
If more than 1024 event channels are bound to a evtchn device then it
possible (even with well behaved applications) for the ring to
overflow and events to be lost (reported as an -EFBIG error).
Dynamically increase the size of the ring so there is always enough
space for all bound events. Well behaved applicables that only unmask
events after draining them from the ring can thus no longer lose
events.
However, an application could unmask an event before draining it,
allowing multiple entries per port to accumulate in the ring, and a
overflow could still occur. So the overflow detection and reporting
is retained.
The ring size is initially only 64 entries so the common use case of
an application only binding a few events will use less memory than
before. The ring size may grow to 512 KiB (enough for all 2^17
possible channels). This order 7 kmalloc() may fail due to memory
fragmentation, so we fall back to trying vmalloc().
Signed-off-by: David Vrabel <david.vrabel@citrix.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
(cherry picked from commit 8620015499101090ae275bf11e9bc2f9febfdf08) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Boris Ostrovsky [Tue, 10 Nov 2015 20:10:33 +0000 (15:10 -0500)]
xen/gntdev: Grant maps should not be subject to NUMA balancing
Doing so will cause the grant to be unmapped and then, during
fault handling, the fault to be mistakenly treated as NUMA hint
fault.
In addition, even if those maps could partcipate in NUMA
balancing, it wouldn't provide any benefit since we are unable
to determine physical page's node (even if/when VNUMA is
implemented).
Marking grant maps' VMAs as VM_IO will exclude them from being
part of NUMA balancing.
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Cc: stable@vger.kernel.org Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 9c17d96500f78d7ecdb71ca6942830158bc75a2b) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Zhenzhong Duan [Tue, 27 Oct 2015 19:19:52 +0000 (15:19 -0400)]
xen: fix the check of e_pfn in xen_find_pfn_range
On some NUMA system, after dom0 up, we see below warning even if there are
enough pfn ranges that could be used for remapping:
"Unable to find available pfn range, not remapping identity pages"
Fix it to avoid getting a memory region of zero size in xen_find_pfn_range.
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@oracle.com> Reviewed-by: Juergen Gross <jgross@suse.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit abed7d0710e8f892c267932a9492ccf447674fb8) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Stefano Stabellini [Thu, 22 Oct 2015 16:21:46 +0000 (16:21 +0000)]
xen, cpu_hotplug: call device_offline instead of cpu_down
When offlining a cpu, instead of cpu_down, call device_offline, which
also takes care of updating the cpu.dev.offline field. This keeps the
sysfs file /sys/devices/system/cpu/cpuN/online, up to date. Also move
the call to disable_hotplug_cpu, because it makes more sense to have it
there.
We don't call device_online at cpu-hotplug time, because that would
immediately take the cpu online, while we want to retain the current
behaviour: the user needs to explicitly enable the cpu after it has
been hotplugged.
Stefano Stabellini [Thu, 22 Oct 2015 16:20:46 +0000 (16:20 +0000)]
xen/arm: Enable cpu_hotplug.c
Build cpu_hotplug for ARM and ARM64 guests.
Rename arch_(un)register_cpu to xen_(un)register_cpu and provide an
empty implementation on ARM and ARM64. On x86 just call
arch_(un)register_cpu as we are already doing.
Julien Grall [Tue, 13 Oct 2015 16:50:13 +0000 (17:50 +0100)]
xenbus: Support multiple grants ring with 64KB
The PV ring may use multiple grants and expect them to be mapped
contiguously in the virtual memory.
Although, the current code is relying on a Linux page will be mapped to
a single grant. On build where Linux is using a different page size than
the grant (i.e other than 4KB), the grant will always be mapped on the
first 4KB of each Linux page which make the final ring not contiguous in
the memory.
This can be fixed by mapping multiple grant in a same Linux page.
Signed-off-by: Julien Grall <julien.grall@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 89bf4b4e4a8d9ab219cd03aada24e782cf0ac359) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Julien Grall [Tue, 13 Oct 2015 16:50:12 +0000 (17:50 +0100)]
xen/grant-table: Add an helper to iterate over a specific number of grants
With the 64KB page granularity support on ARM64, a Linux page may be
split accross multiple grant.
Currently we have the helper gnttab_foreach_grant_in_grant to break a
Linux page based on an offset and a len, but it doesn't fit when we only
have a number of grants in hand.
Introduce a new helper which take an array of Linux page and a number of
grant and will figure out the address of each grant.
Signed-off-by: Julien Grall <julien.grall@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit f73314b28148f9ee9f89a0ae961c8fb36e3269fa) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Julien Grall [Wed, 7 Oct 2015 13:04:33 +0000 (14:04 +0100)]
xen/balloon: Use the correct sizeof when declaring frame_list
The type of the item in frame_list is xen_pfn_t which is not an unsigned
long on ARM but an uint64_t.
With the current computation, the size of frame_list will be 2 *
PAGE_SIZE rather than PAGE_SIZE.
I bet it's just mistake when the type has been switched from "unsigned
long" to "xen_pfn_t" in commit 965c0aaafe3e75d4e65cd4ec862915869bde3abd
"xen: balloon: use correct type for frame_list".
Signed-off-by: Julien Grall <julien.grall@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 3990dd27034606312429a09c807ea74a6ec32dde) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
xen/swiotlb: Add support for 64KB page granularity
Swiotlb is used on ARM64 to support DMA on platform where devices are
not protected by an SMMU. Furthermore it's only enabled for DOM0.
While Xen is always using 4KB page granularity in the stage-2 page table,
Linux ARM64 may either use 4KB or 64KB. This means that a Linux page
can be spanned accross multiple Xen page.
The Swiotlb code has to validate that the buffer used for DMA is
physically contiguous in the memory. As a Linux page can't be shared
between local memory and foreign page by design (the balloon code always
removing entirely a Linux page), the changes in the code are very
minimal because we only need to check the first Xen PFN.
Note that it may be possible to optimize the function
check_page_physically_contiguous to avoid looping over every Xen PFN
for local memory. Although I will let this optimization for a follow-up.
Julien Grall [Tue, 5 May 2015 15:36:56 +0000 (16:36 +0100)]
arm/xen: Add support for 64KB page granularity
The hypercall interface is always using 4KB page granularity. This is
requiring to use xen page definition macro when we deal with hypercall.
Note that pfn_to_gfn is working with a Xen pfn (i.e 4KB). We may want to
rename pfn_gfn to make this explicit.
We also allocate a 64KB page for the shared page even though only the
first 4KB is used. I don't think this is really important for now as it
helps to have the pointer 4KB aligned (XENMEM_add_to_physmap is taking a
Xen PFN).
Julien Grall [Tue, 5 May 2015 15:54:12 +0000 (16:54 +0100)]
xen/privcmd: Add support for Linux 64KB page granularity
The hypercall interface (as well as the toolstack) is always using 4KB
page granularity. When the toolstack is asking for mapping a series of
guest PFN in a batch, it expects to have the page map contiguously in
its virtual memory.
When Linux is using 64KB page granularity, the privcmd driver will have
to map multiple Xen PFN in a single Linux page.
Note that this solution works on page granularity which is a multiple of
4KB.
Signed-off-by: Julien Grall <julien.grall@citrix.com> Reviewed-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 5995a68a6272e4e8f4fe4de82cdc877e650fe8be) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Julien Grall [Tue, 5 May 2015 12:15:29 +0000 (13:15 +0100)]
net/xen-netback: Make it running on 64KB page granularity
The PV network protocol is using 4KB page granularity. The goal of this
patch is to allow a Linux using 64KB page granularity working as a
network backend on a non-modified Xen.
It's only necessary to adapt the ring size and break skb data in small
chunk of 4KB. The rest of the code is relying on the grant table code.
Signed-off-by: Julien Grall <julien.grall@citrix.com> Reviewed-by: Wei Liu <wei.liu2@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit d0089e8a0e4c9723d85b01713671358e3d6960df) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
net/xen-netfront: Make it running on 64KB page granularity
The PV network protocol is using 4KB page granularity. The goal of this
patch is to allow a Linux using 64KB page granularity using network
device on a non-modified Xen.
It's only necessary to adapt the ring size and break skb data in small
chunk of 4KB. The rest of the code is relying on the grant table code.
Note that we allocate a Linux page for each rx skb but only the first
4KB is used. We may improve the memory usage by extending the size of
the rx skb.
Signed-off-by: Julien Grall <julien.grall@citrix.com> Reviewed-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 30c5d7f0da82f55c86c0a09bf21c0623474bb17f) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Julien Grall [Tue, 5 May 2015 15:25:56 +0000 (16:25 +0100)]
block/xen-blkback: Make it running on 64KB page granularity
The PV block protocol is using 4KB page granularity. The goal of this
patch is to allow a Linux using 64KB page granularity behaving as a
block backend on a non-modified Xen.
It's only necessary to adapt the ring size and the number of request per
indirect frames. The rest of the code is relying on the grant table
code.
Note that the grant table code is allocating a Linux page per grant
which will result to waste 6OKB for every grant when Linux is using 64KB
page granularity. This could be improved by sharing the page between
multiple grants.
block/xen-blkfront: Make it running on 64KB page granularity
The PV block protocol is using 4KB page granularity. The goal of this
patch is to allow a Linux using 64KB page granularity using block
device on a non-modified Xen.
The block API is using segment which should at least be the size of a
Linux page. Therefore, the driver will have to break the page in chunk
of 4K before giving the page to the backend.
When breaking a 64KB segment in 4KB chunks, it is possible that some
chunks are empty. As the PV protocol always require to have data in the
chunk, we have to count the number of Xen page which will be in use and
avoid sending empty chunks.
Note that, a pre-defined number of grants are reserved before preparing
the request. This pre-defined number is based on the number and the
maximum size of the segments. If each segment contains a very small
amount of data, the driver may reserve too many grants (16 grants is
reserved per segment with 64KB page granularity).
Furthermore, in the case of persistent grants we allocate one Linux page
per grant although only the first 4KB of the page will be effectively
in use. This could be improved by sharing the page with multiple grants.
Signed-off-by: Julien Grall <julien.grall@citrix.com> Acked-by: Roger Pau Monné <roger.pau@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit c004a6fe0c405e2aa91b2a88aa1428724e6d06f6) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Julien Grall [Tue, 5 May 2015 15:37:30 +0000 (16:37 +0100)]
xen/events: fifo: Make it running on 64KB granularity
Only use the first 4KB of the page to store the events channel info. It
means that we will waste 60KB every time we allocate page for:
* control block: a page is allocating per CPU
* event array: a page is allocating everytime we need to expand it
I think we can reduce the memory waste for the 2 areas by:
* control block: sharing between multiple vCPUs. Although it will
require some bookkeeping in order to not free the page when the CPU
goes offline and the other CPUs sharing the page still there
* event array: always extend the array event by 64K (i.e 16 4K
chunk). That would require more care when we fail to expand the
event channel.
Signed-off-by: Julien Grall <julien.grall@citrix.com> Reviewed-by: David Vrabel <david.vrabel@citrix.com> Reviewed-by: Stefano Stabellini <stefano.stabellini@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit a001c9d95c4ea96589461d58e77c96416a303e2c) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Julien Grall [Mon, 4 May 2015 14:39:08 +0000 (15:39 +0100)]
xen/balloon: Don't rely on the page granularity is the same for Xen and Linux
For ARM64 guests, Linux is able to support either 64K or 4K page
granularity. Although, the hypercall interface is always based on 4K
page granularity.
With 64K page granularity, a single page will be spread over multiple
Xen frame.
To avoid splitting the page into 4K frame, take advantage of the
extent_order field to directly allocate/free chunk of the Linux page
size.
Note that PVMMU is only used for PV guest (which is x86) and the page
granularity is always 4KB. Some BUILD_BUG_ON has been added to ensure
that because the code has not been modified.
Julien Grall [Mon, 11 May 2015 12:44:21 +0000 (13:44 +0100)]
xen/biomerge: Don't allow biovec's to be merged when Linux is not using 4KB pages
On ARM all dma-capable devices on a same platform may not be protected
by an IOMMU. The DMA requests have to use the BFN (i.e MFN on ARM) in
order to use correctly the device.
While the DOM0 memory is allocated in a 1:1 fashion (PFN == MFN), grant
mapping will screw this contiguous mapping.
When Linux is using 64KB page granularitary, the page may be split
accross multiple non-contiguous MFN (Xen is using 4KB page
granularity). Therefore a DMA request will likely fail.
Checking that a 64KB page is using contiguous MFN is tedious. For
now, always says that biovec are not mergeable.
Prepare the code to support 64KB page granularity. The first
implementation will use a full Linux page per indirect and persistent
grant. When non-persistent grant is used, each page of a bio request
may be split in multiple grant.
Furthermore, the field page of the grant structure is only used to copy
data from persistent grant or indirect grant. Avoid to set it for other
use case as it will have no meaning given the page will be split in
multiple grant.
Provide 2 functions, to setup indirect grant, the other for bio page.
Signed-off-by: Julien Grall <julien.grall@citrix.com> Acked-by: Roger Pau Monné <roger.pau@citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 4f503fbdf319e4411aa48852b8922c93a9cc0c5d) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
pfn = page_to_gfn(...) /* Or similar */
gnttab_grant_foreign_access_ref
Replace it by a new helper. Note that when Linux is using a different
page granularity than Xen, the helper only gives access to the first 4KB
grant.
This is useful where drivers are allocating a full Linux page for each
grant.
Also include xen/interface/grant_table.h rather than xen/grant_table.h in
asm/page.h for x86 to fix a compilation issue [1]. Only the former is
useful in order to get the structure definition.
[1] Interdependency between asm/page.h and xen/grant_table.h which result
to page_mfn not being defined when necessary.
Signed-off-by: Julien Grall <julien.grall@citrix.com> Reviewed-by: David Vrabel <david.vrabel@citrix.com> Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 3922f32c1e6db2e096ff095a5b8af0b940b97508) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Julien Grall [Fri, 19 Jun 2015 16:49:03 +0000 (17:49 +0100)]
xen/grant: Introduce helpers to split a page into grant
Currently, a grant is always based on the Xen page granularity (i.e
4KB). When Linux is using a different page granularity, a single page
will be split between multiple grants.
The new helpers will be in charge of splitting the Linux page into grants
and call a function given by the caller on each grant.
Also provide an helper to count the number of grants within a given
contiguous region.
Note that the x86/include/asm/xen/page.h is now including
xen/interface/grant_table.h rather than xen/grant_table.h. It's
necessary because xen/grant_table.h depends on asm/xen/page.h and will
break the compilation. Furthermore, only definition in
interface/grant_table.h is required.
Signed-off-by: Julien Grall <julien.grall@citrix.com> Reviewed-by: David Vrabel <david.vrabel@citrix.com> Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
(cherry picked from commit 008c320a96d218712043f8db0111d5472697785c) Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
projects / users / jedix / linux-maple.git / log