Ensure direct jump patching in s390 is atomic by:
 * naturally aligning a location of direct jump address;
 * using atomic_read()/atomic_set() for code patching.
Signed-off-by: NSergey Fedorov <serge.fdrv@gmail.com>
Signed-off-by: NSergey Fedorov <sergey.fedorov@linaro.org>
Message-Id: <1461341333-19646-7-git-send-email-sergey.fedorov@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Ensure direct jump patching in i386 is atomic by:
 * naturally aligning a location of direct jump address;
 * using atomic_read()/atomic_set() for code patching.

tcg_out_nopn() implementation:
Suggested-by: Richard Henderson <rth@twiddle.net>.
Signed-off-by: NSergey Fedorov <serge.fdrv@gmail.com>
Signed-off-by: NSergey Fedorov <sergey.fedorov@linaro.org>
Message-Id: <1461341333-19646-6-git-send-email-sergey.fedorov@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Ensure direct jump patching in PPC is atomic by:
 * limiting translation buffer size in 32-bit mode to be addressable by
   Branch I-form instruction;
 * using atomic_read()/atomic_set() for code patching.
Signed-off-by: NSergey Fedorov <serge.fdrv@gmail.com>
Signed-off-by: NSergey Fedorov <sergey.fedorov@linaro.org>
Reviewed-by: NAlex Bennée <alex.bennee@linaro.org>
Message-Id: <1461341333-19646-5-git-send-email-sergey.fedorov@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Ensure direct jump patching in TCI is atomic by:
 * naturally aligning a location of direct jump address;
 * using atomic_read()/atomic_set() to load/store the address.
Signed-off-by: NSergey Fedorov <serge.fdrv@gmail.com>
Signed-off-by: NSergey Fedorov <sergey.fedorov@linaro.org>
Message-Id: <1461341333-19646-4-git-send-email-sergey.fedorov@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

These macros provide a convenient way to n-byte align pointers up and
down and check if a pointer is n-byte aligned.
Signed-off-by: NSergey Fedorov <serge.fdrv@gmail.com>
Signed-off-by: NSergey Fedorov <sergey.fedorov@linaro.org>
Message-Id: <1461341333-19646-3-git-send-email-sergey.fedorov@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Signed-off-by: NSergey Fedorov <serge.fdrv@gmail.com>
Signed-off-by: NSergey Fedorov <sergey.fedorov@linaro.org>
Message-Id: <1461341333-19646-2-git-send-email-sergey.fedorov@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

We are inconsistent with the type of tb->flags: usage varies loosely
between int and uint64_t. Settle to uint32_t everywhere, which is
superior to both: at least one target (aarch64) uses the most significant
bit in the u32, and uint64_t is wasteful.

Compile-tested for all targets.
Suggested-by: NLaurent Desnogues <laurent.desnogues@gmail.com>
Suggested-by: NRichard Henderson <rth@twiddle.net>
Tested-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: NLaurent Desnogues <laurent.desnogues@gmail.com>
Signed-off-by: NEmilio G. Cota <cota@braap.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>
Message-Id: <1460049562-23517-1-git-send-email-cota@braap.org>

Block layer patches

# gpg: Signature made Thu 12 May 2016 14:37:05 BST using RSA key ID C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"

* remotes/kevin/tags/for-upstream: (69 commits)
  qemu-iotests: iotests: fail hard if not run via "check"
  block: enable testing of LUKS driver with block I/O tests
  block: add support for encryption secrets in block I/O tests
  block: add support for --image-opts in block I/O tests
  qemu-io: Add 'write -z -u' to test MAY_UNMAP flag
  qemu-io: Add 'write -f' to test FUA flag
  qemu-io: Allow unaligned access by default
  qemu-io: Use bool for command line flags
  qemu-io: Make 'open' subcommand more like command line
  qemu-io: Add missing option documentation
  qmp: add monitor command to add/remove a child
  quorum: implement bdrv_add_child() and bdrv_del_child()
  Add new block driver interface to add/delete a BDS's child
  qemu-img: check block status of backing file when converting.
  iotests: fix the redirection order in 083
  block: Inactivate all children
  block: Drop superfluous invalidating bs->file from drivers
  block: Invalidate all children
  nbd: Simplify client FUA handling
  block: Honor BDRV_REQ_FUA during write_zeroes
  ...
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

target-arm queue:
 * blizzard, omap_lcdc: code cleanup to remove DEPTH != 32 dead code
 * QOMify various ARM devices
 * bcm2835_property: use cached values when querying framebuffer
 * hw/arm/nseries: don't allocate large sized array on the stack
 * fix LPAE descriptor address masking (only visible for EL2)
 * fix stage 2 exec permission handling for AArch32
 * first part of supporting syndrome info for data aborts to EL2
 * virt: NUMA support
 * work towards i.MX6 support
 * avoid unnecessary TLB flush on TCR_EL2, TCR_EL3 writes

# gpg: Signature made Thu 12 May 2016 14:29:14 BST using RSA key ID 14360CDE
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>"
# gpg:                 aka "Peter Maydell <pmaydell@gmail.com>"
# gpg:                 aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>"

* remotes/pmaydell/tags/pull-target-arm-20160512: (43 commits)
  hw/arm: QOM'ify versatilepb.c
  hw/arm: QOM'ify strongarm.c
  hw/arm: QOM'ify stellaris.c
  hw/arm: QOM'ify spitz.c
  hw/arm: QOM'ify pxa2xx_pic.c
  hw/arm: QOM'ify pxa2xx.c
  hw/arm: QOM'ify integratorcp.c
  hw/arm: QOM'ify highbank.c
  hw/arm: QOM'ify armv7m.c
  target-arm: Avoid unnecessary TLB flush on TCR_EL2, TCR_EL3 writes
  hw/display/blizzard: Remove blizzard_template.h
  hw/display/blizzard: Expand out macros
  i.MX: Add sabrelite i.MX6 emulation.
  i.MX: Add i.MX6 SOC implementation.
  i.MX: Add the Freescale SPI Controller
  FIFO: Add a FIFO32 implementation
  i.MX: Add i.MX6 System Reset Controller device.
  ARM: Factor out ARM on/off PSCI control functions
  ACPI: Virt: Generate SRAT table
  ACPI: move acpi_build_srat_memory to common place
  ...
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

QAPI patches for 2016-05-12

# gpg: Signature made Thu 12 May 2016 08:49:04 BST using RSA key ID EB918653
# gpg: Good signature from "Markus Armbruster <armbru@redhat.com>"
# gpg:                 aka "Markus Armbruster <armbru@pond.sub.org>"

* remotes/armbru/tags/pull-qapi-2016-05-12: (23 commits)
  qapi: Change visit_type_FOO() to no longer return partial objects
  qapi: Simplify semantics of visit_next_list()
  qapi: Fix string input visitor handling of invalid list
  tests/string-input-visitor: Add negative integer tests
  qapi: Split visit_end_struct() into pieces
  qmp: Tighten output visitor rules
  qmp: Don't reuse qmp visitor after grabbing output
  spapr_drc: Expose 'null' in qom-get when there is no fdt
  qmp: Support explicit null during visits
  qapi: Add visit_type_null() visitor
  tests: Add check-qnull
  qapi: Document visitor interfaces, add assertions
  qmp-input: Refactor when list is advanced
  qmp-input: Require struct push to visit members of top dict
  qom: Wrap prop visit in visit_start_struct
  qapi-commands: Wrap argument visit in visit_start_struct
  qmp-input: Don't consume input when checking has_member
  qapi: Use strict QMP input visitor in more places
  qapi: Consolidate QMP input visitor creation
  qmp-input: Clean up stack handling
  ...
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Block patches for 2.7

# gpg: Signature made Thu May 12 15:34:13 2016 CEST using RSA key ID E838ACAD
# gpg: Good signature from "Max Reitz <mreitz@redhat.com>"

* mreitz/tags/pull-block-for-kevin-2016-05-12:
  qemu-iotests: iotests: fail hard if not run via "check"
  block: enable testing of LUKS driver with block I/O tests
  block: add support for encryption secrets in block I/O tests
  block: add support for --image-opts in block I/O tests
  qemu-io: Add 'write -z -u' to test MAY_UNMAP flag
  qemu-io: Add 'write -f' to test FUA flag
  qemu-io: Allow unaligned access by default
  qemu-io: Use bool for command line flags
  qemu-io: Make 'open' subcommand more like command line
  qemu-io: Add missing option documentation
  qmp: add monitor command to add/remove a child
  quorum: implement bdrv_add_child() and bdrv_del_child()
  Add new block driver interface to add/delete a BDS's child
  qemu-img: check block status of backing file when converting.
  iotests: fix the redirection order in 083
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

usb: misc fixes

# gpg: Signature made Wed 11 May 2016 12:18:25 BST using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"

* remotes/kraxel/tags/pull-usb-20160511-1:
  usb: Support compilation without poll.h
  usb-mtp: fix usb_mtp_get_device_info so that libmtp on the guest doesn't complain
  usb:xhci: no DMA on HC reset
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Running an iotests-based Python test directly might appear to work,
but may fail in subtle ways and is insecure:

- It creates files with predictable file names in a world-writable
  location (/var/tmp).

- Tests expect the environment to be set up by check. E.g. 041 and 055
  may take the wrong code paths if QEMU_DEFAULT_MACHINE is not
  set. This can lead to false negatives.

Instead fail hard and tell the user we want to be run via "check".

The actual environment expected by the tests is currently only defined
by the implementation of "check". We use two of the environment
variables set by "check" as indication of whether we're being run via
"check". Anyone writing their own test runner (replacing "check") will
need to replicate the full environment (in a broader sense, not just
environment variables) provided by "check" anyway, including setting
the two environment variables we check. Whereas a regular developer
just trying to invoke the tests usually won't have both of these
defined in their environment so we can catch their mistake and give
out useful advice.
Signed-off-by: NSascha Silbe <silbe@linux.vnet.ibm.com>
Reviewed-by: NBo Tu <tubo@linux.vnet.ibm.com>
Message-id: 1461094442-16014-1-git-send-email-silbe@linux.vnet.ibm.com
Signed-off-by: NMax Reitz <mreitz@redhat.com>

This adds support for testing the LUKS driver with the block
I/O test framework.

   cd tests/qemu-io-tests
   ./check -luks

A handful of test cases are modified to work with luks

 - 004 - whitelist luks format
 - 012 - use TEST_IMG_FILE instead of TEST_IMG for file ops
 - 048 - use TEST_IMG_FILE instead of TEST_IMG for file ops.
         don't assume extended image contents is all zeros,
         explicitly initialize with zeros
         Make file size smaller to avoid having to decrypt
         1 GB of data.
 - 052 - don't assume initial image contents is all zeros,
         explicitly initialize with zeros
 - 100 - don't assume initial image contents is all zeros,
         explicitly initialize with zeros

With this patch applied, the results are as follows:

  Passed: 001 002 003 004 005 008 009 010 011 012 021 032 043
          047 048 049 052 087 100 134 143
  Failed: 033 120 140 145
 Skipped: 007 013 014 015 017 018 019 020 022 023 024 025 026
          027 028 029 030 031 034 035 036 037 038 039 040 041
          042 043 044 045 046 047 049 050 051 053 054 055 056
          057 058 059 060 061 062 063 064 065 066 067 068 069
          070 071 072 073 074 075 076 077 078 079 080 081 082
          083 084 085 086 087 088 089 090 091 092 093 094 095
          096 097 098 099 101 102 103 104 105 107 108 109 110
          111 112 113 114 115 116 117 118 119 121 122 123 124
          128 129 130 131 132 133 134 135 136 137 138 139 141
          142 144 146 148 150 152

The reasons for the failed tests are:

 - 033 - needs adapting to use image opts syntax with blkdebug
         and test image in order to correctly set align property
 - 120 - needs adapting to use correct -drive syntax for luks
 - 140 - needs adapting to use correct -drive syntax for luks
 - 145 - needs adapting to use correct -drive syntax for luks

The vast majority of skipped tests are exercising code that is
qcow2 specific, though a couple could probably be usefully
enabled for luks too.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Message-id: 1462896689-18450-4-git-send-email-berrange@redhat.com
Signed-off-by: NMax Reitz <mreitz@redhat.com>

The LUKS block driver tests will require the ability to specify
encryption secrets with block devices. This requires using the
--object argument to qemu-img/qemu-io to create a 'secret'
object.

When the IMGKEYSECRET env variable is set, it provides the
password to be associated with a secret called 'keysec0'

The _qemu_img_wrapper function isn't modified as that needs
to cope with differing syntax for subcommands, so can't be
made to use the image opts syntax unconditionally.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Message-id: 1462896689-18450-3-git-send-email-berrange@redhat.com
Signed-off-by: NMax Reitz <mreitz@redhat.com>

Currently all block tests use the traditional syntax for images
just specifying a filename. To support the LUKS driver without
resorting to JSON, the tests need to be able to use the new
--image-opts argument to qemu-img and qemu-io.

This introduces a new env variable IMGOPTSSYNTAX. If this is
set to 'true', then qemu-img/qemu-io should use --image-opts.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Message-id: 1462896689-18450-2-git-send-email-berrange@redhat.com
Signed-off-by: NMax Reitz <mreitz@redhat.com>

Make it easier to control whether the BDRV_REQ_MAY_UNMAP flag
can be passed through a write_zeroes command, by adding the '-u'
flag to qemu-io 'write -z' and 'aio_write -z'.  To be useful,
the device has to be opened with BDRV_O_UNMAP (done by default
in qemu-io, but can be made explicit with '-d unmap').
Signed-off-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NMax Reitz <mreitz@redhat.com>
Message-id: 1462677405-4752-7-git-send-email-eblake@redhat.com
Signed-off-by: NMax Reitz <mreitz@redhat.com>

Make it easier to test block drivers with BDRV_REQ_FUA in
.supported_write_flags, by adding the '-f' flag to qemu-io to
conditionally pass the flag through to specific writes ('write',
'write -z', 'writev', 'aio_write', 'aio_write -z'). You'll want
to use 'qemu-io -t none' to actually make -f useful (as
otherwise, the default writethrough mode automatically sets the
FUA bit on every write).
Signed-off-by: NEric Blake <eblake@redhat.com>
Message-id: 1462677405-4752-6-git-send-email-eblake@redhat.com
Reviewed-by: NMax Reitz <mreitz@redhat.com>
Signed-off-by: NMax Reitz <mreitz@redhat.com>

There's no reason to require the user to specify a flag just so
they can pass in unaligned numbers.  Keep 'read -p' and 'write -p'
as no-ops so that I don't have to hunt down and update all users
of qemu-io, but otherwise make their behavior default as 'read' and
'write'.  Also fix 'write -z', 'readv', 'writev', 'writev',
'aio_read', 'aio_write', and 'aio_write -z'.  For now, 'read -b',
'write -b', and 'write -c' still require alignment (and 'multiwrite',
but that's slated to die soon).

qemu-iotest 23 is updated to match, as the only test that was
previously explicitly expecting an error on an unaligned request.
Signed-off-by: NEric Blake <eblake@redhat.com>
Message-id: 1462677405-4752-5-git-send-email-eblake@redhat.com
Reviewed-by: NMax Reitz <mreitz@redhat.com>
Signed-off-by: NMax Reitz <mreitz@redhat.com>

We require a C99 compiler; let's use it to express what we
really mean.

(Yes, we now have an instance of 'if (bool + bool + bool > 1)',
which, although semantically valid C, looks ugly; it gets
cleaned up later.)
Signed-off-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NMax Reitz <mreitz@redhat.com>
Message-id: 1462677405-4752-4-git-send-email-eblake@redhat.com
Signed-off-by: NMax Reitz <mreitz@redhat.com>

The command line defaults to BDRV_O_UNMAP, but can use
-d to reset it.  Meanwhile, the 'open' subcommand was
defaulting to no discards, with no way to set it.

The command line has both -n and -tMODE to set a variety
of cache modes, but the 'open' subcommand had only -n.

The 'open' subcommand had no way to set BDRV_O_NATIVE_AIO.

Note that the 'reopen' subcommand uses '-c' where the
command line and 'open' use -t.  Making that consistent
would be a separate patch.
Signed-off-by: NEric Blake <eblake@redhat.com>
Message-id: 1462677405-4752-3-git-send-email-eblake@redhat.com
Reviewed-by: NMax Reitz <mreitz@redhat.com>
Signed-off-by: NMax Reitz <mreitz@redhat.com>

The Usage: summary is missing several options, but rather than
having to maintain it, it's simpler to just state [OPTIONS],
since the options are spelled out below.

Commit 499afa25 added --image-opts, but forgot to document it in
--help.  Likewise for commit 9e8f1835 and -d/--discard.

Commit e3aff4f6 put "-o/--offset" in the long opts, but it has
never been honored.

Add a note that '-n' is short for '-t none'.

Commit 9a2d77ad killed the -C option, but forgot to undocument
it for the 'open' subcommand.

Finally, commit 10d9d75c removed -g/--growable, but forgot to
cull it from the valid short options.
Signed-off-by: NEric Blake <eblake@redhat.com>
Message-id: 1462677405-4752-2-git-send-email-eblake@redhat.com
Reviewed-by: NMax Reitz <mreitz@redhat.com>
Signed-off-by: NMax Reitz <mreitz@redhat.com>

The new QMP command name is x-blockdev-change. It's just for adding/removing
quorum's child now, and doesn't support all kinds of children, all kinds of
operations, nor all block drivers. So it is experimental now.
Signed-off-by: NWen Congyang <wency@cn.fujitsu.com>
Signed-off-by: Nzhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: NGonglei <arei.gonglei@huawei.com>
Signed-off-by: NChanglong Xie <xiecl.fnst@cn.fujitsu.com>
Reviewed-by: NMax Reitz <mreitz@redhat.com>
Reviewed-by: NAlberto Garcia <berto@igalia.com>
Message-id: 1462865799-19402-4-git-send-email-xiecl.fnst@cn.fujitsu.com
Signed-off-by: NMax Reitz <mreitz@redhat.com>

Signed-off-by: NWen Congyang <wency@cn.fujitsu.com>
Signed-off-by: Nzhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: NGonglei <arei.gonglei@huawei.com>
Signed-off-by: NChanglong Xie <xiecl.fnst@cn.fujitsu.com>
Message-id: 1462865799-19402-3-git-send-email-xiecl.fnst@cn.fujitsu.com
Reviewed-by: NAlberto Garcia <berto@igalia.com>
Reviewed-by: NMax Reitz <mreitz@redhat.com>
Signed-off-by: NMax Reitz <mreitz@redhat.com>

In some cases, we want to take a quorum child offline, and take
another child online.
Signed-off-by: NWen Congyang <wency@cn.fujitsu.com>
Signed-off-by: Nzhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: NGonglei <arei.gonglei@huawei.com>
Signed-off-by: NChanglong Xie <xiecl.fnst@cn.fujitsu.com>
Reviewed-by: NMax Reitz <mreitz@redhat.com>
Reviewed-by: NAlberto Garcia <berto@igalia.com>
Message-id: 1462865799-19402-2-git-send-email-xiecl.fnst@cn.fujitsu.com
Signed-off-by: NMax Reitz <mreitz@redhat.com>

When converting images, check the block status of its backing file chain
to avoid needlessly reading zeros.
Signed-off-by: NRen Kimura <rkx1209dev@gmail.com>
Message-id: 1461773098-20356-1-git-send-email-rkx1209dev@gmail.com
Signed-off-by: NMax Reitz <mreitz@redhat.com>

It should redirect stdout to /dev/null first,
then redirect stderr to whatever stdout currently points at.
Signed-off-by: NWei Jiangang <weijg.fnst@cn.fujitsu.com>
Message-id: 1461665601-14908-1-git-send-email-weijg.fnst@cn.fujitsu.com
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NMax Reitz <mreitz@redhat.com>

Currently we only inactivate the top BDS. Actually bdrv_inactivate
should be the opposite of bdrv_invalidate_cache.

Recurse into the whole subtree instead.

Because a node may have multiple parents, and because once
BDRV_O_INACTIVE is set for a node, further writes are not allowed, we
cannot interleave flag settings and .bdrv_inactivate calls (that may
submit write to other nodes in a graph) within a single pass. Therefore
two passes are used here.
Signed-off-by: NFam Zheng <famz@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

Now they are invalidated by the block layer, so it's not necessary to
do this in block drivers' implementations of .bdrv_invalidate_cache.
Signed-off-by: NFam Zheng <famz@redhat.com>
Reviewed-by: NAlberto Garcia <berto@igalia.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

Currently we only recurse to bs->file, which will miss the children in quorum
and VMDK.

Recurse into the whole subtree to avoid that.
Signed-off-by: NFam Zheng <famz@redhat.com>
Reviewed-by: NAlberto Garcia <berto@igalia.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

Now that the block layer honors per-bds FUA support, we don't
have to duplicate the fallback flush at the NBD layer.  The
static function nbd_co_writev_flags() is no longer needed, and
the driver can just directly use nbd_client_co_writev().
Signed-off-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NFam Zheng <famz@redhat.com>
Acked-by: NStefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

The block layer has a couple of cases where it can lose
Force Unit Access semantics when writing a large block of
zeroes, such that the request returns before the zeroes
have been guaranteed to land on underlying media.

SCSI does not support FUA during WRITESAME(10/16); FUA is only
supported if it falls back to WRITE(10/16).  But where the
underlying device is new enough to not need a fallback, it
means that any upper layer request with FUA semantics was
silently ignoring BDRV_REQ_FUA.

Conversely, NBD has situations where it can support FUA but not
ZERO_WRITE; when that happens, the generic block layer fallback
to bdrv_driver_pwritev() (or the older bdrv_co_writev() in qemu
2.6) was losing the FUA flag.

The problem of losing flags unrelated to ZERO_WRITE has been
latent in bdrv_co_do_write_zeroes() since commit aa7bfbff, but
back then, it did not matter because there was no FUA flag.  It
became observable when commit 93f5e6d8 paved the way for flags
that can impact correctness, when we should have been using
bdrv_co_writev_flags() with modified flags.  Compare to commit
9eeb6dd1, which got flag manipulation right in
bdrv_co_do_zero_pwritev().

Symptoms: I tested with qemu-io with default writethrough cache
(which is supposed to use FUA semantics on every write), and
targetted an NBD client connected to a server that intentionally
did not advertise NBD_FLAG_SEND_FUA.  When doing 'write 0 512',
the NBD client sent two operations (NBD_CMD_WRITE then
NBD_CMD_FLUSH) to get the fallback FUA semantics; but when doing
'write -z 0 512', the NBD client sent only NBD_CMD_WRITE.

The fix is do to a cleanup bdrv_co_flush() at the end of the
operation if any step in the middle relied on a BDS that does
not natively support FUA for that step (note that we don't
need to flush after every operation, if the operation is broken
into chunks based on bounce-buffer sizing).  Each BDS gains a
new flag .supported_zero_flags, which parallels the use of
.supported_write_flags but only when accessing a zero write
operation (the flags MUST be different, because of SCSI having
different semantics based on WRITE vs. WRITESAME; and also
because BDRV_REQ_MAY_UNMAP only makes sense on zero writes).

Also fix some documentation to describe -ENOTSUP semantics,
particularly since iscsi depends on those semantics.

Down the road, we may want to add a driver where its
.bdrv_co_pwritev() honors all three of BDRV_REQ_FUA,
BDRV_REQ_ZERO_WRITE, and BDRV_REQ_MAY_UNMAP, and advertise
this via bs->supported_write_flags for blocks opened by that
driver; such a driver should NOT supply .bdrv_co_write_zeroes
nor .supported_zero_flags.  But none of the drivers touched
in this patch want to do that (the act of writing zeroes is
different enough from normal writes to deserve a second
callback).
Signed-off-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NFam Zheng <famz@redhat.com>
Acked-by: NStefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

Pre-patch, .supported_write_flags lives at the driver level, which
means we are blindly declaring that all block devices using a
given driver will either equally support FUA, or that we need a
fallback at the block layer.  But there are drivers where FUA
support is a per-block decision: the NBD block driver is dependent
on the remote server advertising NBD_FLAG_SEND_FUA (and has
fallback code to duplicate the flush that the block layer would do
if NBD had not set .supported_write_flags); and the iscsi block
driver is dependent on the mode sense bits advertised by the
underlying device (and is currently silently ignoring FUA requests
if the underlying device does not support FUA).

The fix is to make supported flags as a per-BDS option, set during
.bdrv_open().  This patch moves the variable and fixes NBD and iscsi
to set it only conditionally; later patches will then further
simplify the NBD driver to quit duplicating work done at the block
layer, as well as tackle the fact that SCSI does not support FUA
semantics on WRITESAME(10/16) but only on WRITE(10/16).
Signed-off-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NFam Zheng <famz@redhat.com>
Acked-by: NStefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

There is a possibility that qcow2_co_write_zeroes() will be called
with the partial block. This could be synthetically triggered with
    qemu-io -c "write -z 32k 4k"
and can happen in the real life in qemu-nbd. The latter happens under
the following conditions:
    (1) qemu-nbd is started with --detect-zeroes=on and is connected to the
        kernel NBD client
    (2) third party program opens kernel NBD device with O_DIRECT
    (3) third party program performs write operation with memory buffer
        not aligned to the page
In this case qcow2_co_write_zeroes() is unable to perform the operation
and mark entire cluster as zeroed and returns ENOTSUP. Thus the caller
switches to non-optimized version and writes real zeroes to the disk.

The patch creates a shortcut. If the block is read as zeroes, f.e. if
it is unallocated, the request is extended to cover full block.
User-visible situation with this block is not changed. Before the patch
the block is filled in the image with real zeroes. After that patch the
block is marked as zeroed in metadata. Thus any subsequent changes in
backing store chain are not affected.

Kevin, thank you for a cool suggestion.
Signed-off-by: NDenis V. Lunev <den@openvz.org>
Reviewed-by: NRoman Kagan <rkagan@virtuozzo.com>
CC: Kevin Wolf <kwolf@redhat.com>
CC: Max Reitz <mreitz@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

Now that there are no remaining clients, we can drop the
sector-based blk_read(), blk_write(), blk_aio_readv(), and
blk_aio_writev().  Sadly, there are still remaining
sector-based interfaces, such as blk_*discard(), or
blk_write_compressed(); those will have to wait for another
day.
Signed-off-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

qemu-io is the last user of several sector-based interfaces.
This patch upgrades to the new interfaces under the hood,
then deletes the resulting dead code.  Note that for maximum
back-compat, while the -p option is no longer required to get
blk_pread(), it is still needed to allow for unaligned access;
this is because qemu-iotest 23 relies on qemu-io rejecting
unaligned accesses without -p.  A later patch may clean up the
interface to be more user-friendly, but it's better to separate
what's done under the hood from what the user sees.
Signed-off-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

Sector-based blk_write() should die; switch to byte-based
blk_pwrite() instead.  Likewise for blk_read().
Signed-off-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

Sector-based blk_read() should die; switch to byte-based
blk_pread() instead.

Add a constant for our magic number 512, to make it obvious
that this size will NOT change even if BDRV_SECTOR_SIZE does,
even though the two happen to be the same for now.  Split
assignments from conditionals to keep checkpatch.pl happy.
Signed-off-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

Sector-based blk_read() should die; switch to byte-based
blk_pread() instead.

Add new defines ATAPI_SECTOR_BITS and ATAPI_SECTOR_SIZE to
use anywhere we were previously scaling BDRV_SECTOR_* by 4,
for better legibility.
Signed-off-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

Sector-based blk_read() should die; switch to byte-based
blk_pread() instead.

Likewise for blk_aio_readv() and blk_aio_writev().
Signed-off-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>