Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block patches

# gpg: Signature made Fri 15 Aug 2014 14:07:42 BST using RSA key ID C88F2FD6
# gpg: Good signature from "Kevin Wolf <[email protected]>"

* remotes/kevin/tags/for-upstream: (59 commits)
  block: Catch !bs->drv in bdrv_check()
  iotests: Add test for image header overlap
  qcow2: Catch !*host_offset for data allocation
  qcow2: Return useful error code in refcount_init()
  mirror: Handle failure for potentially large allocations
  vpc: Handle failure for potentially large allocations
  vmdk: Handle failure for potentially large allocations
  vhdx: Handle failure for potentially large allocations
  vdi: Handle failure for potentially large allocations
  rbd: Handle failure for potentially large allocations
  raw-win32: Handle failure for potentially large allocations
  raw-posix: Handle failure for potentially large allocations
  qed: Handle failure for potentially large allocations
  qcow2: Handle failure for potentially large allocations
  qcow1: Handle failure for potentially large allocations
  parallels: Handle failure for potentially large allocations
  nfs: Handle failure for potentially large allocations
  iscsi: Handle failure for potentially large allocations
  dmg: Handle failure for potentially large allocations
  curl: Handle failure for potentially large allocations
  ...

Signed-off-by: Peter Maydell <[email protected]>

block: Catch !bs->drv in bdrv_check()

qemu-img check calls bdrv_check() twice if the first run repaired some
inconsistencies. If the first run however again triggered corruption
prevention (on qcow2) due to very bad inconsistencies, bs->drv may be
NULL afterwards. Thus, bdrv_check() should check whether bs->drv is set.

Signed-off-by: Max Reitz <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

iotests: Add test for image header overlap

Add a test for an image with an unallocated image header; instead of an
assertion, this should result in the image being marked corrupt.

Signed-off-by: Max Reitz <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

qcow2: Catch !*host_offset for data allocation

qcow2_alloc_cluster_offset() uses host_offset == 0 as "no preferred
offset" for the (data) cluster range to be allocated. However, this
offset is actually valid and may be allocated on images with a corrupted
refcount table or first refcount block.

In this case, the corruption prevention should normally catch that
write anyway (because it would overwrite the image header). But since 0
is a special value here, the function assumes that nothing has been
allocated at all which it asserts against.

Because this condition is not qemu's fault but rather that of a broken
image, it shouldn't throw an assertion but rather mark the image corrupt
and show an appropriate message, which this patch does by calling the
corruption check earlier than it would be called normally (before the
assertion).

Signed-off-by: Max Reitz <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

qcow2: Return useful error code in refcount_init()

If bdrv_pread() returns an error, it is very unlikely that it was
ENOMEM. In this case, the return value should be passed along; as
bdrv_pread() will always either return the number of bytes read or a
negative value (the error code), the condition for checking whether
bdrv_pread() failed can be simplified (and clarified) as well.

Signed-off-by: Max Reitz <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

mirror: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the mirror block job.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

vpc: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the vpc block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

vmdk: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the vmdk block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

vhdx: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the vhdx block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

vdi: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the vdi block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

rbd: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the rbd block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>

raw-win32: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the raw-win32 block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>

raw-posix: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the raw-posix block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>

qed: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the qed block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

qcow2: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the qcow2 block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>

qcow1: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the qcow1 block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>

parallels: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the parallels block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

nfs: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the nfs block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

iscsi: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the iscsi block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Acked-by: Paolo Bonzini <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>
Reviewed-by: Eric Blake <[email protected]>

dmg: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the dmg block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

curl: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the curl block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

cloop: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the cloop block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

bochs: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses the allocations in the bochs block driver.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

block: Handle failure for potentially large allocations

Some code in the block layer makes potentially huge allocations. Failure
is not completely unexpected there, so avoid aborting qemu and handle
out-of-memory situations gracefully.

This patch addresses bounce buffer allocations in block.c. While at it,
convert bdrv_commit() from plain g_malloc() to qemu_try_blockalign().

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>

block: Introduce qemu_try_blockalign()

This function returns NULL instead of aborting when an allocation fails.

Signed-off-by: Kevin Wolf <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>

block: iotest - update 084 to test static VDI image creation

This updates the VDI corruption test to also test static VDI image
creation, as well as the default dynamic image creation.

Reviewed-by: Max Reitz <[email protected]>
Signed-off-by: Jeff Cody <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

block: vpc - use block layer ops in vpc_create, instead of posix calls

Use the block layer to create, and write to, the image file in the VPC
.bdrv_create() operation.

This has a couple of benefits: Images can now be created over protocols,
and hacks such as NOCOW are not needed in the image format driver, and
the underlying file protocol appropriate for the host OS can be relied
upon.

Reviewed-by: Max Reitz <[email protected]>
Signed-off-by: Jeff Cody <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

block: use the standard 'ret' instead of 'result'

Most QEMU code uses 'ret' for function return values. The VDI driver
uses a mix of 'result' and 'ret'.  This cleans that up, switching over
to the standard 'ret' usage.

Reviewed-by: Max Reitz <[email protected]>
Signed-off-by: Jeff Cody <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

block: vdi - use block layer ops in vdi_create, instead of posix calls

Use the block layer to create, and write to, the image file in the
VDI .bdrv_create() operation.

This has a couple of benefits: Images can now be created over protocols,
and hacks such as NOCOW are not needed in the image format driver, and
the underlying file protocol appropriate for the host OS can be relied
upon.

Also some minor cleanup for error handling.

Reviewed-by: Max Reitz <[email protected]>
Signed-off-by: Jeff Cody <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

block: allow bdrv_unref() to be passed NULL pointers

If bdrv_unref() is passed a NULL BDS pointer, it is safe to
exit with no operation.  This will allow cleanup code to blindly
call bdrv_unref() on a BDS that has been initialized to NULL.

Reviewed-by: Max Reitz <[email protected]>
Signed-off-by: Jeff Cody <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

test-coroutine: add baseline test that times the cost of function calls

This can be used to compute the cost of coroutine operations.  In the
end the cost of the function call is a few clock cycles, so it's pretty
cheap for now, but it may become more relevant as the coroutine code
is optimized.

For example, here are the results on my machine:

   Function call 100000000 iterations: 0.173884 s
   Yield 100000000 iterations: 8.445064 s
   Lifecycle 1000000 iterations: 0.098445 s
   Nesting 10000 iterations of 1000 depth each: 7.406431 s

One yield takes 83 nanoseconds, one enter takes 97 nanoseconds,
one coroutine allocation takes (roughly, since some of the allocations
in the nesting test do hit the pool) 739 nanoseconds:

   (8.445064 - 0.173884) * 10^9 / 100000000 = 82.7
   (0.098445 * 100 - 0.173884) * 10^9 / 100000000 = 96.7
   (7.406431 * 10 - 0.173884) * 10^9 / 100000000 = 738.9

Signed-off-by: Paolo Bonzini <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

block: VHDX endian fixes

This patch contains several changes for endian conversion fixes for
VHDX, particularly for big-endian machines (multibyte values in VHDX are
all on disk in LE format).

Tests were done with existing qemu-iotests on an IBM POWER7 (8406-71Y).
This includes sample images created by Hyper-V, both with dirty logs and
without.

In addition, VHDX image files created (and written to) on a BE machine
were tested on a LE machine, and vice-versa.

Reported-by: Markus Armburster <[email protected]>
Reported-by: Paolo Bonzini <[email protected]>
Signed-off-by: Jeff Cody <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

block: vhdx - add error check

This add an error check for an invalid descriptor entry signature,
when flushing the log descriptor entries.

Signed-off-by: Jeff Cody <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

thread-pool: avoid deadlock in nested aio_poll() calls

The thread pool has a race condition if two elements complete before
thread_pool_completion_bh() runs:

  If element A's callback waits for element B using aio_poll() it will
  deadlock since pool->completion_bh is not marked scheduled when the
  nested aio_poll() runs.

Fix this by marking the BH scheduled while thread_pool_completion_bh()
is executing.  This way any nested aio_poll() loops will enter
thread_pool_completion_bh() and complete the remaining elements.

Signed-off-by: Stefan Hajnoczi <[email protected]>

thread-pool: avoid per-thread-pool EventNotifier

EventNotifier is implemented using an eventfd or pipe.  It therefore
consumes file descriptors, which can be limited by rlimits and should
therefore be used sparingly.

Switch from EventNotifier to QEMUBH in thread-pool.c.  Originally
EventNotifier was used because qemu_bh_schedule() was not thread-safe
yet.

Reported-by: Christian Borntraeger <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

block: bump coroutine pool size for drives

When a BlockDriverState is associated with a storage controller
DeviceState we expect guest I/O.  Use this opportunity to bump the
coroutine pool size by 64.

This patch ensures that the coroutine pool size scales with the number
of drives attached to the guest.  It should increase coroutine pool
usage (which makes qemu_coroutine_create() fast) without hogging too
much memory when fewer drives are attached.

Signed-off-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Eric Blake <[email protected]>

coroutine: make pool size dynamic

Allow coroutine users to adjust the pool size.  For example, if the
guest has multiple emulated disk drives we should keep around more
coroutines.

Signed-off-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Eric Blake <[email protected]>

qemu-iotests: add support for Archipelago protocol

Signed-off-by: Chrysostomos Nanakos <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

QMP: Add support for Archipelago

Introduce new enum BlockdevOptionsArchipelago.

@volume:              #Name of the Archipelago volume image

@mport:               #'mport' is the port number on which mapperd is
                      listening. This is optional and if not specified,
                      QEMU will make Archipelago to use the default port.

@vport:               #'vport' is the port number on which vlmcd is
                      listening. This is optional and if not specified,
                      QEMU will make Archipelago to use the default port.

@segment:             #optional The name of the shared memory segment
                      Archipelago stack is using. This is optional
                      and if not specified, QEMU will make Archipelago
                      use the default value, 'archipelago'.

Signed-off-by: Chrysostomos Nanakos <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

block/archipelago: Add support for creating images

qemu-img archipelago:<volumename>[/mport=<mapperd_port>[:vport=<vlmcd_port>]
 [:segment=<segment_name>]] [size]

Signed-off-by: Chrysostomos Nanakos <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

block/archipelago: Implement bdrv_parse_filename()

VM Image on Archipelago volume can also be specified like this:

file=archipelago:<volumename>[/mport=<mapperd_port>[:vport=<vlmcd_port>][:
segment=<segment_name>]]

Examples:

file=archipelago:my_vm_volume
file=archipelago:my_vm_volume/mport=123
file=archipelago:my_vm_volume/mport=123:vport=1234
file=archipelago:my_vm_volume/mport=123:vport=1234:segment=my_segment

Signed-off-by: Chrysostomos Nanakos <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

block: Support Archipelago as a QEMU block backend

VM Image on Archipelago volume is specified like this:

file.driver=archipelago,file.volume=<volumename>[,file.mport=<mapperd_port>[,
file.vport=<vlmcd_port>][,file.segment=<segment_name>]]

'archipelago' is the protocol.

'mport' is the port number on which mapperd is listening. This is optional
and if not specified, QEMU will make Archipelago to use the default port.

'vport' is the port number on which vlmcd is listening. This is optional
and if not specified, QEMU will make Archipelago to use the default port.

'segment' is the name of the shared memory segment Archipelago stack is using.
This is optional and if not specified, QEMU will make Archipelago to use the
default value, 'archipelago'.

Examples:

file.driver=archipelago,file.volume=my_vm_volume
file.driver=archipelago,file.volume=my_vm_volume,file.mport=123
file.driver=archipelago,file.volume=my_vm_volume,file.mport=123,
file.vport=1234
file.driver=archipelago,file.volume=my_vm_volume,file.mport=123,
file.vport=1234,file.segment=my_segment

Signed-off-by: Chrysostomos Nanakos <[email protected]>
Reviewed-by: Stefan Hajnoczi <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

qemu-img info: show nocow info

Add nocow info in 'qemu-img info' output to show whether the file
currently has NOCOW flag set or not.

Signed-off-by: Chunyan Liu <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

vmdk: Optimize cluster allocation

This drops the unnecessary bdrv_truncate() from, and also improves,
cluster allocation code path.

Before, when we need a new cluster, get_cluster_offset truncates the
image to bdrv_getlength() + cluster_size, and returns the offset of
added area, i.e. the image length before truncating.

This is not efficient, so it's now rewritten as:

  - Save the extent file length when opening.

  - When allocating cluster, use the saved length as cluster offset.

  - Don't truncate image, because we'll anyway write data there: just
    write any data at the EOF position, in descending priority:

    * New user data (cluster allocation happens in a write request).

    * Filling data in the beginning and/or ending of the new cluster, if
      not covered by user data: either backing file content (COW), or
      zero for standalone images.

One major benifit of this change is, on host mounted NFS images, even
over a fast network, ftruncate is slow (see the example below). This
change significantly speeds up cluster allocation. Comparing by
converting a cirros image (296M) to VMDK on an NFS mount point, over
1Gbe LAN:

    $ time qemu-img convert cirros-0.3.1.img /mnt/a.raw -O vmdk

    Before:
        real    0m21.796s
        user    0m0.130s
        sys     0m0.483s

    After:
        real    0m2.017s
        user    0m0.047s
        sys     0m0.190s

We also get rid of unchecked bdrv_getlength() and bdrv_truncate(), and
get a little more documentation in function comments.

Tested that this passes qemu-iotests for all VMDK subformats.

Signed-off-by: Fam Zheng <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

qemu-iotests: Add data pattern in version3 VMDK sample image in 059

It's possible that we diverge from the specification with our
implementation.  Having a reference image in the test cases may detect
such problems when we introduce a bug that can read what it creates, but
can't handle a real VMDK.

Signed-off-by: Fam Zheng <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

qdev-monitor: include QOM properties in -device FOO, help output

Update -device FOO,help to include QOM properties in addition to qdev
properties.  Devices are gradually adding more QOM properties that are
not reflected as qdev properties.

It is important to report all device properties since management tools
like libvirt use this information (and device-list-properties QMP) to
detect the presence of QEMU features.

This patch reuses the device-list-properties QMP machinery to avoid code
duplication.

Reported-by: Cole Robinson <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Tested-by: Cole Robinson <[email protected]>

qmp: hide "hotplugged" device property from device-list-properties

The "hotplugged" device property was not reported before commit
f4eb32b590bf58c1c67570775eb78beb09964fad ("qmp: show QOM properties in
device-list-properties").  Fix this difference.

Signed-off-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Eric Blake <[email protected]>

docs/multiple-iothreads.txt: add documentation on IOThread programming

This document explains how IOThreads and the main loop are related,
especially how to write code that can run in an IOThread.  Currently
only virtio-blk-data-plane uses these techniques.  The next obvious
target is virtio-scsi; there has also been work on virtio-net.

Signed-off-by: Stefan Hajnoczi <[email protected]>
Reviewed-by: Eric Blake <[email protected]>

xen_disk:  fix possible null-ptr dereference

Signed-off-by: Gonglei <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

configure: explicitly state version requirements to devel packages

Signed-off-by: Hu Tao <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

docs: Make the recommendation for the backing file name position a requirement

The current version of the qcow2 specification recommends to save the backing
file name in the end of the first cluster. It follows that the backing file
name can be saved somewhere in the image, but the first cluster, which
contradicts the current QEMU implementation.

The patch makes the backing file name required to be placed after the header
extensions in the first image cluster.

Signed-off-by: Maria Kustova <[email protected]>
Signed-off-by: Kevin Wolf <[email protected]>

block: Avoid bdrv_get_geometry() where errors should be detected

bdrv_get_geometry() hides errors.  Use bdrv_nb_sectors() or
bdrv_getlength() instead where that's obviously inappropriate.

Signed-off-by: Markus Armbruster <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

qemu-img: Make img_convert() get image size just once per image

Chiefly so I don't have to do the error checking in quadruplicate in
the next commit.  Moreover, replacing the frequently updated
bs_sectors by an array assigned just once makes the code easier to
understand.

Signed-off-by: Markus Armbruster <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

block: Drop superfluous aligning of bdrv_getlength()'s value

It returns a multiple of the sector size.

Signed-off-by: Markus Armbruster <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

block: Use bdrv_nb_sectors() where sectors, not bytes are wanted

Instead of bdrv_getlength().

Aside: a few of these callers don't handle errors.  I didn't
investigate whether they should.

Signed-off-by: Markus Armbruster <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

block: Use bdrv_nb_sectors() in img_convert()

Instead of bdrv_getlength().  Replace variable output_length by
output_sectors.

Signed-off-by: Markus Armbruster <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

block: Use bdrv_nb_sectors() in bdrv_co_get_block_status()

Instead of bdrv_getlength().

Replace variables length, length2 by total_sectors, nb_sectors2.
Bonus: use total_sectors instead of the slightly unclean
bs->total_sectors.

Signed-off-by: Markus Armbruster <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

block: Use bdrv_nb_sectors() in bdrv_aligned_preadv()

Instead of bdrv_getlength().  Eliminate variable len.

Signed-off-by: Markus Armbruster <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

block: Use bdrv_nb_sectors() in bdrv_make_zero()

Instead of bdrv_getlength().

Variable target_size is initially in bytes, then changes meaning to
sectors.  Ugh.  Replace by target_sectors.

Signed-off-by: Markus Armbruster <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

block: New bdrv_nb_sectors()

A call to retrieve the image size converts between bytes and sectors
several times:

* BlockDriver method bdrv_getlength() returns bytes.

* refresh_total_sectors() converts to sectors, rounding up, and stores
  in total_sectors.

* bdrv_getlength() converts total_sectors back to bytes (now rounded
  up to a multiple of the sector size).

* Callers wanting sectors rather bytes convert it right back.
  Example: bdrv_get_geometry().

bdrv_nb_sectors() provides a way to omit the last two conversions.
It's exactly bdrv_getlength() with the conversion to bytes omitted.
It's functionally like bdrv_get_geometry() without its odd error
handling.

Reimplement bdrv_getlength() and bdrv_get_geometry() on top of
bdrv_nb_sectors().

The next patches will convert some users of bdrv_getlength() to
bdrv_nb_sectors().

Signed-off-by: Markus Armbruster <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Reviewed-by: Benoit Canet <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

Merge remote-tracking branch 'remotes/mjt/tags/trivial-patches-2014-08-09' into staging

trivial patches for 2014-08-09

# gpg: Signature made Fri 08 Aug 2014 21:36:44 BST using RSA key ID A4C3D7DB
# gpg: Good signature from "Michael Tokarev <[email protected]>"
# gpg:                 aka "Michael Tokarev <[email protected]>"
# gpg:                 aka "Michael Tokarev <[email protected]>"
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 6EE1 95D1 886E 8FFB 810D  4324 457C E0A0 8044 65C5
#      Subkey fingerprint: 6F67 E18E 7C91 C5B1 5514  66A7 BEE5 9D74 A4C3 D7DB

* remotes/mjt/tags/trivial-patches-2014-08-09:
  build-sys: Move qapi-{types, visit, event}.o into util-obj-y
  po: Add Chinese translation
  qemu-img: Check getchar() return value in read_password() for WIN32
  hw/timer: Move extern declaration from .c to .h file
  virtio: Move extern declaration to header file
  Show length mismatch error is hex
  target-i386/cpu.c: Fix two error output indentation
  l2tpv3 (configure): it is linux-specific
  hw/timer/imx_*: fix TIMER_MAX clash with system symbol

Signed-off-by: Peter Maydell <[email protected]>

build-sys: Move qapi-{types, visit, event}.o into util-obj-y

These three objects are repeated in multiple times in Makefiles. Let's
just add them to libqemuutil.a, and don't list explicitly elsewhere.

Signed-off-by: Fam Zheng <[email protected]>
Signed-off-by: Michael Tokarev <[email protected]>

po: Add Chinese translation

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Amos Kong <[email protected]>
Reviewed-by: Dongsheng Song <[email protected]>
Reviewed-by: Wei Huang <[email protected]>
Signed-off-by: Michael Tokarev <[email protected]>

qemu-img: Check getchar() return value in read_password() for WIN32

getchar() is a standard c library function which may return with failure
(e.g. -1), so like another platforms, also need check it under WIN32.

And make the related code match current qemu code styles, too.

Signed-off-by: Chen Gang <[email protected]>
Signed-off-by: Michael Tokarev <[email protected]>

hw/timer: Move extern declaration from .c to .h file

This fixes a warning from smatch (static code analyser).

Fix also the comment with the renamed source file name.

Signed-off-by: Stefan Weil <[email protected]>
 hw/timer/tusb6010.c |    3 ---
 include/hw/usb.h    |    7 ++++++-
 2 files changed, 6 insertions(+), 4 deletions(-)
Signed-off-by: Michael Tokarev <[email protected]>

virtio: Move extern declaration to header file

This fixes a warning from smatch (static code analyser).

Signed-off-by: Stefan Weil <[email protected]>
Signed-off-by: Michael Tokarev <[email protected]>

Show length mismatch error is hex

When live migrate fails due to a section length mismatch we currently
see an error message like:

Length mismatch: 0000:00:03.0/virtio-net-pci.rom: 10000 in != 20000

The section lengths are in fact in hex, so this should read

Length mismatch: 0000:00:03.0/virtio-net-pci.rom: 0x10000 in != 0x20000

Correct the error string to reflect this.

Signed-off-by: Alex Bligh <[email protected]>
Signed-off-by: Michael Tokarev <[email protected]>

target-i386/cpu.c: Fix two error output indentation

Signed-off-by: Chen Fan <[email protected]>
Reviewed-by: Igor Mammedov <[email protected]>
Signed-off-by: Michael Tokarev <[email protected]>

l2tpv3 (configure): it is linux-specific

Some non-linux systems, for example a system with
FreeBSD kernel and glibc, may declare struct mmsghdr
(in glibc) but may not have linux-specific header
file linux/ip.h.  The actual implementation in qemu
includes this linux-specific header file unconditionally,
so compilation fails if it is not present.  Include
this header in the configure test too.

Signed-off-by: Michael Tokarev <[email protected]>

hw/timer/imx_*: fix TIMER_MAX clash with system symbol

The symbol TIMER_MAX used in imx_epit.c and imx_gpt.c
clashes with system symbol with the same name.  Because
all qemu source files includes qemu-common.h which, in
turn, includes limits.h, which is not unusual to define
it.  Rename local symbol to have a reasonable prefix.

Signed-off-by: Michael Tokarev <[email protected]>

Merge remote-tracking branch 'remotes/mdroth/qga-pull-2014-08-08' into staging

* remotes/mdroth/qga-pull-2014-08-08:
  qga: Disable unsupported commands by default
  qga: Add guest-get-fsinfo command
  qga: Add guest-fsfreeze-freeze-list command

Signed-off-by: Peter Maydell <[email protected]>

qga: Disable unsupported commands by default

Currently management softwares cannot know whether a qemu-ga command is
supported or not on the running platform until they actually execute it.
This patch disables unsupported commands at launch time of qemu-ga, so that
management softwares can check whether they are supported from 'enabled'
property of the result from 'guest-info' command.

Signed-off-by: Tomoki Sekiyama <[email protected]>
Signed-off-by: Michael Roth <[email protected]>

qga: Add guest-get-fsinfo command

Add command to get mounted filesystems information in the guest.
The returned value contains a list of mountpoint paths and
corresponding disks info such as disk bus type, drive address,
and the disk controllers' PCI addresses, so that management layer
such as libvirt can resolve the disk backends.

For example, when `lsblk' result is:

    NAME           MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
    sdb              8:16   0    1G  0 disk
    `-sdb1           8:17   0 1024M  0 part
      `-vg0-lv0    253:1    0  1.4G  0 lvm  /mnt/test
    sdc              8:32   0    1G  0 disk
    `-sdc1           8:33   0  512M  0 part
      `-vg0-lv0    253:1    0  1.4G  0 lvm  /mnt/test
    vda            252:0    0   25G  0 disk
    `-vda1         252:1    0   25G  0 part /

where sdb is a SCSI disk with PCI controller 0000:00:0a.0 and ID=1,
      sdc is an IDE disk with PCI controller 0000:00:01.1, and
      vda is a virtio-blk disk with PCI device 0000:00:06.0,

guest-get-fsinfo command will return the following result:

    {"return":
     [{"name":"dm-1",
       "mountpoint":"/mnt/test",
       "disk":[
        {"bus-type":"scsi","bus":0,"unit":1,"target":0,
         "pci-controller":{"bus":0,"slot":10,"domain":0,"function":0}},
        {"bus-type":"ide","bus":0,"unit":0,"target":0,
         "pci-controller":{"bus":0,"slot":1,"domain":0,"function":1}}],
       "type":"xfs"},
      {"name":"vda1", "mountpoint":"/",
       "disk":[
        {"bus-type":"virtio","bus":0,"unit":0,"target":0,
         "pci-controller":{"bus":0,"slot":6,"domain":0,"function":0}}],
       "type":"ext4"}]}

In Linux guest, the disk information is resolved from sysfs. So far,
it only supports virtio-blk, virtio-scsi, IDE, SATA, SCSI disks on x86
hosts, and "disk" parameter may be empty for unsupported disk types.

Signed-off-by: Tomoki Sekiyama <[email protected]>
*updated schema to report 2.2 as initial supported version

Signed-off-by: Michael Roth <[email protected]>

qga: Add guest-fsfreeze-freeze-list command

If an array of mount point paths is specified as 'mountpoints' argument
of guest-fsfreeze-freeze-list, qemu-ga will only freeze the file systems
mounted on specified paths in Linux guests. Otherwise, it works as the
same way as guest-fsfreeze-freeze.
This would be useful when the host wants to create partial disk snapshots.

Signed-off-by: Tomoki Sekiyama <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
*updated schema to report 2.2 as initial supported version

Signed-off-by: Michael Roth <[email protected]>

Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging

KVM changes include a MIPS patch and the testdev backend used by the
ARM kvm-unit-tests.  icount include the first part of reverse execution
and Sebastian Tanase's patches to slow down -icount execution to the
desired speed of the target.

v1->v2: fix dump_drift_info to print nothing outside icount mode,
        and to compile on 32-bit architectures

# gpg: Signature made Thu 07 Aug 2014 14:09:58 BST using RSA key ID 9B4D86F2
# gpg: Good signature from "Paolo Bonzini <[email protected]>"
# gpg:                 aka "Paolo Bonzini <[email protected]>"

* remotes/bonzini/tags/for-upstream:
  target-mips: Ignore unassigned accesses with KVM
  monitor: Add drift info to 'info jit'
  cpu-exec: Print to console if the guest is late
  cpu-exec: Add sleeping algorithm
  icount: Add align option to icount
  icount: Add QemuOpts for icount
  icount: Fix virtual clock start value on ARM
  timer: add cpu_icount_to_ns function.
  migration: migrate icount fields.
  icount: put icount variables into TimerState.
  backends: Introduce chr-testdev

Signed-off-by: Peter Maydell <[email protected]>

target-mips: Ignore unassigned accesses with KVM

MIPS registers an unassigned access handler which raises a guest bus
error exception. However this causes QEMU to crash when KVM is enabled
as it isn't called from the main execution loop so longjmp() gets called
without a corresponding setjmp().

Until the KVM API can be updated to trigger a guest exception in
response to an MMIO exit, prevent the bus error exception being raised
from mips_cpu_unassigned_access() if KVM is enabled.

The check is at run time since the do_unassigned_access callback is
initialised before it is known whether KVM will be enabled.

The problem can be triggered with Malta emulation by making the guest
write to the reset region at physical address 0x1bf00000, since it is
marked read-only which is treated as unassigned for writes.

Signed-off-by: James Hogan <[email protected]>
Reviewed-by: Aurelien Jarno <[email protected]>
Cc: Peter Maydell <[email protected]>
Cc: Paolo Bonzini <[email protected]>
Cc: Gleb Natapov <[email protected]>
Cc: Christoffer Dall <[email protected]>
Cc: Sanjay Lal <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>

monitor: Add drift info to 'info jit'

Show in 'info jit' the current delay between the host clock
and the guest clock. In addition, print the maximum advance
and delay of the guest compared to the host.

Signed-off-by: Sebastian Tanase <[email protected]>
Tested-by: Camille Bégué <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>

Merge remote-tracking branch 'remotes/awilliam/tags/vfio-pci-for-qemu-20140805.0' into staging

VFIO patches: Fix MSI-X vector expansion, remove MSI/X message caching

# gpg: Signature made Tue 05 Aug 2014 20:25:57 BST using RSA key ID 3BB08B22
# gpg: Can't check signature: public key not found

* remotes/awilliam/tags/vfio-pci-for-qemu-20140805.0:
  vfio: Don't cache MSIMessage
  vfio: Fix MSI-X vector expansion

Signed-off-by: Peter Maydell <[email protected]>

cpu-exec: Print to console if the guest is late

If the align option is enabled, we print to the user whenever
the guest clock is behind the host clock in order for he/she
to have a hint about the actual performance. The maximum
print interval is 2s and we limit the number of messages to 100.
If desired, this can be changed in cpu-exec.c

Signed-off-by: Sebastian Tanase <[email protected]>
Tested-by: Camille Bégué <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>

cpu-exec: Add sleeping algorithm

The goal is to sleep qemu whenever the guest clock
is in advance compared to the host clock (we use
the monotonic clocks). The amount of time to sleep
is calculated in the execution loop in cpu_exec.

At first, we tried to approximate at each for loop the real time elapsed
while searching for a TB (generating or retrieving from cache) and
executing it. We would then approximate the virtual time corresponding
to the number of virtual instructions executed. The difference between
these 2 values would allow us to know if the guest is in advance or delayed.
However, the function used for measuring the real time
(qemu_clock_get_ns(QEMU_CLOCK_REALTIME)) proved to be very expensive.
We had an added overhead of 13% of the total run time.

Therefore, we modified the algorithm and only take into account the
difference between the 2 clocks at the begining of the cpu_exec function.
During the for loop we try to reduce the advance of the guest only by
computing the virtual time elapsed and sleeping if necessary. The overhead
is thus reduced to 3%. Even though this method still has a noticeable
overhead, it no longer is a bottleneck in trying to achieve a better
guest frequency for which the guest clock is faster than the host one.

As for the the alignement of the 2 clocks, with the first algorithm
the guest clock was oscillating between -1 and 1ms compared to the host clock.
Using the second algorithm we notice that the guest is 5ms behind the host, which
is still acceptable for our use case.

The tests where conducted using fio and stress. The host machine in an i5 CPU at
3.10GHz running Debian Jessie (kernel 3.12). The guest machine is an arm versatile-pb
built with buildroot.

Currently, on our test machine, the lowest icount we can achieve that is suitable for
aligning the 2 clocks is 6. However, we observe that the IO tests (using fio) are
slower than the cpu tests (using stress).

Signed-off-by: Sebastian Tanase <[email protected]>
Tested-by: Camille Bégué <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>

icount: Add align option to icount

The align option is used for activating the align algorithm
in order to synchronise the host clock and the guest clock.

Signed-off-by: Sebastian Tanase <[email protected]>
Tested-by: Camille Bégué <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>

icount: Add QemuOpts for icount

Make icount parameter use QemuOpts style options in order
to easily add other suboptions.

Signed-off-by: Sebastian Tanase <[email protected]>
Tested-by: Camille Bégué <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>

icount: Fix virtual clock start value on ARM

When using the icount option on ARM, the virtual
clock starts counting at realtime clock but it
should start at 0.

The reason why the virtual clock starts at realtime clock
is because the first time we call qemu_clock_warp (which
calls icount_warp_rt) in tcg_exec_all, qemu_icount_bias
(which is part of the virtual time computation mechanism)
will increment by realtime - vm_clock_warp_start, with
vm_clock_warp_start being 0 (see icount_warp_rt in cpus.c).

By changing the value of vm_clock_warp_start from 0 to -1,
the first time we call qemu_clock_warp which calls
icount_warp_rt, we will return immediatly because
icount_warp_rt first checks if vm_clock_warp_start is -1
and if it's the case it returns. Therefore, qemu_icount_bias
will first be incremented by the value of a virtual timer
deadline when the virtual cpu goes from active to inactive.

The virtual time will start at 0 and increment based
on the instruction counter when the vcpu is active or
the qemu_icount_bias value when inactive.

Signed-off-by: Sebastian Tanase <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>

timer: add cpu_icount_to_ns function.

This adds cpu_icount_to_ns function which is needed for reverse execution.

It returns the time for a specific instruction.

Signed-off-by: KONRAD Frederic <[email protected]>
Reviewed-by: Paolo Bonzini <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>

migration: migrate icount fields.

This fixes a bug where qemu_icount and qemu_icount_bias are not migrated.
It adds a subsection "timer/icount" to vmstate_timers so icount is migrated only
when needed.

Signed-off-by: KONRAD Frederic <[email protected]>
Reviewed-by: Amit Shah <[email protected]>
Reviewed-by: Juan Quintela <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>

icount: put icount variables into TimerState.

This puts qemu_icount and qemu_icount_bias into TimerState structure to allow
them to be migrated.

Signed-off-by: KONRAD Frederic <[email protected]>
Reviewed-by: Paolo Bonzini <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>

backends: Introduce chr-testdev

From: Paolo Bonzini <[email protected]>

chr-testdev enables a virtio serial channel to be used for guest
initiated qemu exits. hw/misc/debugexit already enables guest
initiated qemu exits, but only for PC targets. chr-testdev supports
any virtio-capable target. kvm-unit-tests/arm is already making use
of this backend.

Currently there is a single command implemented, "q".  It takes a
(prefix) argument for the exit code, thus an exit is implemented by
writing, e.g. "1q", to the virtio-serial port.

It can be used as:
   $QEMU ... \
     -device virtio-serial-device \
     -device virtserialport,chardev=ctd -chardev testdev,id=ctd

or, use:
   $QEMU ... \
     -device virtio-serial-device \
     -device virtconsole,chardev=ctd -chardev testdev,id=ctd

to bind it to virtio-serial port0.

Signed-off-by: Paolo Bonzini <[email protected]>
Signed-off-by: Andrew Jones <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>

vfio: Don't cache MSIMessage

Commit 40509f7f added a test to avoid updating KVM MSI routes when the
MSIMessage is unchanged and f4d45d47 switched to relying on this
rather than doing our own comparison.  Our cached msg is effectively
unused now.  Remove it.

Signed-off-by: Alex Williamson <[email protected]>

vfio: Fix MSI-X vector expansion

When new MSI-X vectors are enabled we need to disable MSI-X and
re-enable it with the correct number of vectors.  That means we need
to reprogram the eventfd triggers for each vector.  Prior to f4d45d47
vector->use tracked whether a vector was masked or unmasked and we
could always pick the KVM path when available for unmasked vectors.
Now vfio doesn't track mask state itself and vector->use and virq
remains configured even for masked vectors.  Therefore we need to ask
the MSI-X code whether a vector is masked in order to select the
correct signaling path.  As noted in the comment, MSI relies on
hardware to handle masking.

Signed-off-by: Alex Williamson <[email protected]>
Cc: [email protected] # QEMU 2.1

Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20140804' into staging

target-arm queue:
 * Set PC correctly when loading AArch64 ELF files
 * sdhci: Fix ADMA dma_memory_read access
 * some more foundational work for EL2/EL3 support
 * fix bugs which reveal themselves if the TARGET_PAGE_SIZE
   is not set to 1K

# gpg: Signature made Mon 04 Aug 2014 14:51:34 BST using RSA key ID 14360CDE
# gpg: Good signature from "Peter Maydell <[email protected]>"

* remotes/pmaydell/tags/pull-target-arm-20140804:
  target-arm: A64: fix TLB flush instructions
  target-arm: don't hardcode mask values in arm_cpu_handle_mmu_fault
  target-arm: Fix bit test in sp_el0_access
  target-arm: Add FAR_EL2 and 3
  target-arm: Add ESR_EL2 and 3
  target-arm: Make far_el1 an array
  target-arm: A64: Respect SPSEL when taking exceptions
  target-arm: A64: Respect SPSEL in ERET SP restore
  target-arm: A64: Break out aarch64_save/restore_sp
  sd: sdhci: Fix ADMA dma_memory_read access
  hw/arm/virt: formatting: memory map
  hw/arm/boot: Set PC correctly when loading AArch64 ELF files

Signed-off-by: Peter Maydell <[email protected]>

target-arm: A64: fix TLB flush instructions

According to the ARM ARM we weren't correctly flushing the TLB entries
where bits 63:56 didn't match bit 55 of the virtual address. This
exposed a problem when we switched QEMU's internal TARGET_PAGE_BITS to
12 for aarch64.

Signed-off-by: Alex Bennée <[email protected]>
Reviewed-by: Peter Maydell <[email protected]>
Message-id: 1406733627[email protected]
Signed-off-by: Peter Maydell <[email protected]>

target-arm: don't hardcode mask values in arm_cpu_handle_mmu_fault

Otherwise we break quickly when we change TARGET_PAGE_SIZE.

Signed-off-by: Alex Bennée <[email protected]>
Message-id: 1406733627[email protected]
Signed-off-by: Peter Maydell <[email protected]>

target-arm: Fix bit test in sp_el0_access

Static code analyzers complain about a dubious & operation used for a
boolean value. The code does not test the PSTATE_SP bit as it should.

Cc: Peter Maydell <[email protected]>
Signed-off-by: Stefan Weil <[email protected]>
Message-id: 1406359601[email protected]
Signed-off-by: Peter Maydell <[email protected]>

target-arm: Add FAR_EL2 and 3

Reviewed-by: Greg Bellows <[email protected]>
Signed-off-by: Edgar E. Iglesias <[email protected]>
Reviewed-by: Alex Bennée <[email protected]>
Message-id: 1402994746[email protected]
Signed-off-by: Peter Maydell <[email protected]>

target-arm: Add ESR_EL2 and 3

Reviewed-by: Greg Bellows <[email protected]>
Signed-off-by: Edgar E. Iglesias <[email protected]>
Reviewed-by: Alex Bennée <[email protected]>
Message-id: 1402994746[email protected]
Signed-off-by: Peter Maydell <[email protected]>

target-arm: Make far_el1 an array

No functional change.
Prepares for future additions of the EL2 and 3 versions of this reg.

Reviewed-by: Greg Bellows <[email protected]>
Signed-off-by: Edgar E. Iglesias <[email protected]>
Reviewed-by: Alex Bennée <[email protected]>
Message-id: 1402994746[email protected]
Signed-off-by: Peter Maydell <[email protected]>

target-arm: A64: Respect SPSEL when taking exceptions

Reviewed-by: Alex Bennée <[email protected]>
Signed-off-by: Edgar E. Iglesias <[email protected]>
Reviewed-by: Greg Bellows <[email protected]>
Message-id: 1402994746[email protected]
Signed-off-by: Peter Maydell <[email protected]>

target-arm: A64: Respect SPSEL in ERET SP restore

Reviewed-by: Alex Bennée <[email protected]>
Signed-off-by: Edgar E. Iglesias <[email protected]>
Reviewed-by: Greg Bellows <[email protected]>
Message-id: 1402994746[email protected]
Signed-off-by: Peter Maydell <[email protected]>

target-arm: A64: Break out aarch64_save/restore_sp

Break out code to save/restore AArch64 SP into functions.

Reviewed-by: Alex Bennée <[email protected]>
Signed-off-by: Edgar E. Iglesias <[email protected]>
Reviewed-by: Greg Bellows <[email protected]>
Message-id: 1402994746[email protected]
Signed-off-by: Peter Maydell <[email protected]>

sd: sdhci: Fix ADMA dma_memory_read access

This dma_memory_read was giving too big a size when begin was non-zero.
This could cause segfaults in some circumstances. Fix.

Signed-off-by: Peter Crosthwaite <[email protected]>
Signed-off-by: Peter Maydell <[email protected]>