Patch Round-up for stable 2.2.1, freeze on 2015-03-05
Subsystems (2)
Hi everyone,
The following new patches are queued for QEMU stable v2.2.1:
https://github.com/mdroth/qemu/commits/stable-2.2-staging
The release is planned for 2015-03-10:
http://wiki.qemu.org/Planning/2.2
Please respond here or CC [email protected] on any patches you think should be included in the release.
Testing/feedback is greatly appreciated.
Thanks!
NOTE:
I also plan on including:
6d4e4cb pc: acpi: fix WindowsXP BSOD when memory hotplug is enabled
However, the follow-up ACPI blob updates:
e909c16 acpi: update generated hex files
215d8d1 acpi-test: update expected DSDT
Don't match what's generated by stable-2.2-staging, so I've backed it out for now while I figure out the best way to resolve this.
----------------------------------------------------------------
Alex Williamson (1):
vfio-pci: Fix missing unparent of dynamically allocated MemoryRegion
Cornelia Huck (1):
virtio: fix feature bit checks
Daniel P. Berrange (1):
libcacard: stop linking against every single 3rd party library
David Gibson (1):
PPC: Fix crash on spapr_tce_table_finalize()
Dinar Valeev (1):
hw/input/hid.c Fix capslock hid code
Fam Zheng (1):
qtest: Fix deadloop by running main loop AIO context's timers
Gary R Hook (1):
block migration: fix return value
Kevin Wolf (3):
qcow2: Fix header extension size check qcow2.py: Add required padding for header extensions
block: Don't probe for unknown backing file format
Marcel Apfelbaum (1):
vl.c: fix regression when reading machine type from config file
Max Filippov (2):
target-xtensa: fix translation for opcodes crossing page boundary
target-xtensa: test cross-page opcode
Max Reitz (13):
block: Make essential BlockDriver objects public
block: Omit bdrv_find_format for essential drivers block/vvfat: qcow driver may not be found block/nfs: Add create_opts
block: Check create_opts before image creation
qemu-img: Check create_opts before image creation
qemu-img: Check create_opts before image amendment
iotests: Only kill NBD server if it runs
iotests: Add test for unsupported image creation
qcow2: Prevent numerical overflow
qcow2: Flushing the caches in qcow2_close may fail
qcow2: Respect bdrv_truncate() error block/raw-posix: Fix ret in raw_open_common()
Paolo Bonzini (11):
linuxboot: fix loading old kernels
atomic: fix position of volatile qualifier
serial: reset thri_pending on IER writes with THRI=0
serial: refine serial_thr_ipending_needed
pckbd: set bits 2-3-6-7 of the output port by default
scsi: fix cancellation when I/O was completed but DMA was not.
target-i386: fix movntsd on big-endian hosts
vt82c686: avoid out-of-bounds read
sb16: fix interrupt acknowledgement
qemu-thread: fix qemu_event without futexes
exec: change default exception_index value for migration to -1
Peter Maydell (3):
audio: Don't free hw resources until after hw backend is stopped
linux-user: Fix broken m68k signal handling on 64 bit hosts target-arm/translate-a64: Fix wrong mmu_idx usage for LDT/STT
Peter Wu (1):
block/iscsi: fix uninitialized variable
Vladimir Sementsov-Ogievskiy (1):
migration/block: fix pending() return value
Zhang Haoyu (1):
fix mc146818rtc wrong subsection name to avoid vmstate_subsection_load() fail
audio/audio_template.h | 2 +- block-migration.c | 14 +++++++++++--- block.c | 36 ++++++++++++++++++++---------------- block/iscsi.c | 2 +- block/nfs.c | 15 +++++++++++++++ block/qcow2-cluster.c | 2 +- block/qcow2.c | 33 ++++++++++++++++++++++----------- block/raw-posix.c | 3 ++- block/raw-win32.c | 2 +- block/raw_bsd.c | 2 +- block/vvfat.c | 6 ++++++ cpus.c | 4 ++++ exec.c | 4 ++-- hw/audio/sb16.c | 4 ++-- hw/char/serial.c | 31 +++++++++++++++++++++++++++---- hw/input/hid.c | 2 +- hw/input/pckbd.c | 10 ++++++++-- hw/isa/vt82c686.c | 2 +- hw/misc/vfio.c | 1 + hw/ppc/spapr_iommu.c | 6 +++--- hw/scsi/scsi-bus.c | 2 ++ hw/scsi/virtio-scsi.c | 2 +- hw/timer/mc146818rtc.c | 2 +- hw/virtio/dataplane/vring.c | 4 ++-- include/block/block_int.h | 8 ++++++++ include/qemu/atomic.h | 4 ++-- libcacard/Makefile | 2 ++ linux-user/signal.c | 6 +++--- pc-bios/linuxboot.bin | Bin 1024 -> 1024 bytes pc-bios/optionrom/linuxboot.S | 37 +++++++++++++++++++++++++++---------- qemu-img.c | 21 +++++++++++++++++++++ target-arm/translate-a64.c | 2 +- target-i386/translate.c | 3 ++- target-xtensa/translate.c | 27 +++++++++++++++++++++++---- tests/qemu-iotests/026.out | 120 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/071.out | 8 ++++++++ tests/qemu-iotests/080 | 2 ++ tests/qemu-iotests/080.out | 2 ++ tests/qemu-iotests/089.out | 2 ++ tests/qemu-iotests/113 | 76 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/113.out | 15 +++++++++++++++ tests/qemu-iotests/114 | 61 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/114.out | 13 +++++++++++++ tests/qemu-iotests/common.rc | 4 +++- tests/qemu-iotests/group | 2 ++ tests/qemu-iotests/qcow2.py | 4 ++++ tests/tcg/xtensa/test_mmu.S | 26 +++++++++++++------------- util/qemu-thread-posix.c | 2 ++ vl.c | 15 ++++++++------- 49 files changed, 556 insertions(+), 97 deletions(-) create mode 100755 tests/qemu-iotests/113 create mode 100644 tests/qemu-iotests/113.out create mode 100755 tests/qemu-iotests/114 create mode 100644 tests/qemu-iotests/114.out
PATCH 1/43 block: Make essential BlockDriver objects public
From: Max Reitz <[email protected]>
There are some block drivers which are essential to QEMU and may not be
removed: These are raw, file and qcow2 (as the default non-raw format). Make their BlockDriver objects public so they can be directly referenced throughout the block layer without needing to call bdrv_find_format() and having to deal with an error at runtime, while the real problem occurred during linking (where raw, file or qcow2 were not linked into qemu).
--- block/qcow2.c | 2 +- block/raw-posix.c | 2 +- block/raw-win32.c | 2 +- block/raw_bsd.c | 2 +- include/block/block_int.h | 8 ++++++++ 5 files changed, 12 insertions(+), 4 deletions(-)
block/qcow2.c +1 -1
@@ -2847,7 +2847,7 @@ static QemuOptsList qcow2_create_opts = { } }; -static BlockDriver bdrv_qcow2 = { +BlockDriver bdrv_qcow2 = { .format_name ="qcow2" , .instance_size =sizeof (BDRVQcowState), .bdrv_probe = qcow2_probe,
block/raw-posix.c +1 -1
@@ -1684,7 +1684,7 @@ static QemuOptsList raw_create_opts = { } }; -static BlockDriver bdrv_file = { +BlockDriver bdrv_file = { .format_name ="file" , .protocol_name ="file" , .instance_size =sizeof (BDRVRawState),
block/raw-win32.c +1 -1
@@ -540,7 +540,7 @@ static QemuOptsList raw_create_opts = { } }; -static BlockDriver bdrv_file = { +BlockDriver bdrv_file = { .format_name ="file" , .protocol_name ="file" , .instance_size =sizeof (BDRVRawState),
block/raw_bsd.c +1 -1
@@ -173,7 +173,7 @@ static int raw_probe(const uint8_t *buf, int buf_size, const char *filename)return 1 ; } -static BlockDriver bdrv_raw = { +BlockDriver bdrv_raw = { .format_name ="raw" , .bdrv_probe = &raw_probe, .bdrv_reopen_prepare = &raw_reopen_prepare,
include/block/block_int.h +8 -0
@@ -411,6 +411,14 @@ struct BlockDriverState {Error *backing_blocker; }; + +/* Essential block drivers which must always be statically linked into qemu, and +* which therefore can be accessed without using bdrv_find_format() */ +extern BlockDriver bdrv_file; +extern BlockDriver bdrv_raw; +extern BlockDriver bdrv_qcow2; + +int get_tmp_filename (char *filename,int size); void bdrv_set_io_limits(BlockDriverState *bs,
Signature · 1 line
1.9.1
PATCH 2/43 block: Omit bdrv_find_format for essential drivers
In reply to Michael RothFrom: Max Reitz <[email protected]>
We can always assume raw, file and qcow2 being available; so do not use bdrv_find_format() to locate their BlockDriver objects but statically reference the respective objects.
--- block.c | 17 +++++------------ block/qcow2.c | 7 +++---- 2 files changed, 8 insertions(+), 16 deletions(-)
block.c +5 -12
@@ -629,7 +629,7 @@ BlockDriver *bdrv_find_protocol(const char *filename, }if (!path_has_protocol (filename) || !allow_protocol_prefix) { -return bdrv_find_format ("file" ); +return &bdrv_file; } p =strchr (filename,':' );
@@ -658,12 +658,7 @@ static int find_image_format(BlockDriverState *bs, const char *filename,/* Return the raw BlockDriver * to scsi-generic devices or empty drives */ if (bs->sg || !bdrv_is_inserted (bs) ||bdrv_getlength (bs) ==0 ) { - drv =bdrv_find_format ("raw" ); -if (!drv) { -error_setg (errp,"Could not find raw image format" ); - ret = -ENOENT ; - } - *pdrv = drv; + *pdrv = &bdrv_raw;return ret; }
@@ -1294,7 +1289,6 @@ int bdrv_append_temp_snapshot(BlockDriverState *bs, int flags, Error **errp)/* TODO: extra byte is a hack to ensure MAX_PATH space on Windows. */ char *tmp_filename =g_malloc0 (PATH_MAX +1 );int64_t total_size; -BlockDriver *bdrv_qcow2;QemuOpts *opts =NULL ;QDict *snapshot_options;BlockDriverState *bs_snapshot;
@@ -1319,11 +1313,10 @@ int bdrv_append_temp_snapshot(BlockDriverState *bs, int flags, Error **errp) goto out; } - bdrv_qcow2 =bdrv_find_format ("qcow2" ); - opts =qemu_opts_create (bdrv_qcow2->create_opts,NULL ,0 , + opts =qemu_opts_create (bdrv_qcow2.create_opts,NULL ,0 , &error_abort);qemu_opt_set_number (opts,BLOCK_OPT_SIZE , total_size); - ret =bdrv_create (bdrv_qcow2, tmp_filename, opts, &local_err); + ret =bdrv_create (&bdrv_qcow2, tmp_filename, opts, &local_err);qemu_opts_del (opts);if (ret <0 ) { error_setg_errno(errp, -ret, "Could not create temporary overlay "
@@ -1343,7 +1336,7 @@ int bdrv_append_temp_snapshot(BlockDriverState *bs, int flags, Error **errp) bs_snapshot =bdrv_new (); ret =bdrv_open (&bs_snapshot,NULL ,NULL , snapshot_options, - flags, bdrv_qcow2, &local_err); + flags, &bdrv_qcow2, &local_err);if (ret <0 ) {error_propagate (errp, local_err); goto out;
block/qcow2.c +3 -4
@@ -1915,10 +1915,9 @@ static int qcow2_create2(const char *filename, int64_t total_size,* refcount of the cluster that is occupied by the header and the refcount * table) */ -BlockDriver * drv =bdrv_find_format ("qcow2" ); -assert (drv !=NULL ); ret =bdrv_open (&bs, filename,NULL ,NULL , -BDRV_O_RDWR |BDRV_O_CACHE_WB |BDRV_O_NO_FLUSH , drv, &local_err); +BDRV_O_RDWR |BDRV_O_CACHE_WB |BDRV_O_NO_FLUSH , + &bdrv_qcow2, &local_err);if (ret <0 ) {error_propagate (errp, local_err); goto out;
@@ -1970,7 +1969,7 @@ static int qcow2_create2(const char *filename, int64_t total_size,/* Reopen the image without BDRV_O_NO_FLUSH to flush it before returning */ ret =bdrv_open (&bs, filename,NULL ,NULL ,BDRV_O_RDWR |BDRV_O_CACHE_WB |BDRV_O_NO_BACKING , - drv, &local_err); + &bdrv_qcow2, &local_err);if (local_err) {error_propagate (errp, local_err); goto out;
Signature · 1 line
1.9.1
PATCH 3/43 block/vvfat: qcow driver may not be found
In reply to Michael RothFrom: Max Reitz <[email protected]>
Although virtually impossible right now, bdrv_find_format("qcow") may fail. The vvfat block driver should heed that case.
--- block/vvfat.c | 6 ++++++ 1 file changed, 6 insertions(+)
block/vvfat.c +6 -0
@@ -2917,6 +2917,12 @@ static int enable_write_target(BDRVVVFATState *s, Error **errp) } bdrv_qcow =bdrv_find_format ("qcow" ); +if (!bdrv_qcow) { +error_setg (errp,"Failed to locate qcow driver" ); + ret = -ENOENT ; + goto err; + } + opts =qemu_opts_create (bdrv_qcow->create_opts,NULL ,0 , &error_abort);qemu_opt_set_number (opts,BLOCK_OPT_SIZE , s->sector_count *512 );qemu_opt_set (opts,BLOCK_OPT_BACKING_FILE ,"fat:" );
Signature · 1 line
1.9.1
PATCH 4/43 block/nfs: Add create_opts
In reply to Michael RothFrom: Max Reitz <[email protected]>
The nfs protocol driver is capable of creating images, but did not specify any creation options. Fix it.
A way to test this issue is the following:
$ qemu-img create -f nfs nfs://127.0.0.1/foo.qcow2 64M
Without this patch, it segfaults. With this patch, it does not. However, this is not something that should really work; qemu-img should check whether the parameter for the -f option (and -O for convert) is indeed a format, and error out if it is not. Therefore, I am not making it an iotest.
--- block/nfs.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+)
block/nfs.c +15 -0
@@ -409,6 +409,19 @@ out:return ret; } +static QemuOptsList nfs_create_opts = { + .name ="nfs-create-opts" , + .head =QTAILQ_HEAD_INITIALIZER (nfs_create_opts.head), + .desc = { + { + .name =BLOCK_OPT_SIZE , + .type =QEMU_OPT_SIZE , + .help ="Virtual disk size" + }, + {/* end of list */ } + } +}; +static int nfs_file_create (const char *url,QemuOpts *opts,Error **errp) {int ret =0 ;
@@ -470,6 +483,8 @@ static BlockDriver bdrv_nfs = { .instance_size = sizeof(NFSClient), .bdrv_needs_filename = true, + .create_opts = &nfs_create_opts, + .bdrv_has_zero_init = nfs_has_zero_init, .bdrv_get_allocated_file_size = nfs_get_allocated_file_size, .bdrv_truncate = nfs_file_truncate,
Signature · 1 line
1.9.1
PATCH 5/43 block: Check create_opts before image creation
In reply to Michael RothFrom: Max Reitz <[email protected]>
If a driver supports image creation, it needs to set the .create_opts field. We can use that to make sure .create_opts for both drivers involved is not NULL in bdrv_img_create(), which is important so that the create_opts pointer in that function is not NULL after the qemu_opts_append() calls and when going into qemu_opts_create().
--- block.c | 12 ++++++++++++ 1 file changed, 12 insertions(+)
block.c +12 -0
@@ -5534,6 +5534,18 @@ void bdrv_img_create(const char *filename, const char *fmt,return ; } +if (!drv->create_opts) { +error_setg (errp,"Format driver '%s' does not support image creation" , + drv->format_name); +return ; + } + +if (!proto_drv->create_opts) { +error_setg (errp,"Protocol driver '%s' does not support image creation" , + proto_drv->format_name); +return ; + } + create_opts =qemu_opts_append (create_opts, drv->create_opts); create_opts =qemu_opts_append (create_opts, proto_drv->create_opts);
Signature · 1 line
1.9.1
PATCH 6/43 qemu-img: Check create_opts before image creation
In reply to Michael RothFrom: Max Reitz <[email protected]>
If a driver supports image creation, it needs to set the .create_opts field. We can use that to make sure .create_opts for both drivers involved is not NULL for the target image in qemu-img convert, which is important so that the create_opts pointer in img_convert() is not NULL after the qemu_opts_append() calls and when going into qemu_opts_create().
--- qemu-img.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+)
qemu-img.c +14 -0
@@ -1531,6 +1531,20 @@ static int img_convert(int argc, char **argv) goto out; } +if (!drv->create_opts) { +error_report ("Format driver '%s' does not support image creation" , + drv->format_name); + ret =-1 ; + goto out; + } + +if (!proto_drv->create_opts) { +error_report ("Protocol driver '%s' does not support image creation" , + proto_drv->format_name); + ret =-1 ; + goto out; + } + create_opts =qemu_opts_append (create_opts, drv->create_opts); create_opts =qemu_opts_append (create_opts, proto_drv->create_opts);
Signature · 1 line
1.9.1
PATCH 7/43 qemu-img: Check create_opts before image amendment
In reply to Michael RothFrom: Max Reitz <[email protected]>
The image options which can be amended are described by the .create_opts field for every driver. This field must therefore be non-NULL so that anything can be amended in the first place. Check that this holds true before going into qemu_opts_create() (because if .create_opts is NULL, the create_opts pointer in img_amend() will be NULL after qemu_opts_append()).
--- qemu-img.c | 7 +++++++ 1 file changed, 7 insertions(+)
qemu-img.c +7 -0
@@ -2986,6 +2986,13 @@ static int img_amend(int argc, char **argv) goto out; } +if (!bs->drv->create_opts) { +error_report ("Format driver '%s' does not support any options to amend" , + fmt); + ret =-1 ; + goto out; + } + create_opts =qemu_opts_append (create_opts, bs->drv->create_opts); opts =qemu_opts_create (create_opts,NULL ,0 , &error_abort);if (options &&qemu_opts_do_parse (opts, options,NULL )) {
Signature · 1 line
1.9.1
PATCH 8/43 iotests: Only kill NBD server if it runs
In reply to Michael RothFrom: Max Reitz <[email protected]>
There may be NBD tests which do not create a sample image and simply test whether wrong usage of the protocol is rejected as expected. In this case, there will be no NBD server and trying to kill it during clean-up will fail.
--- tests/qemu-iotests/common.rc | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tests/qemu-iotests/common.rc +3 -1
@@ -175,7 +175,9 @@ _cleanup_test_img() case "$IMGPROTO" in nbd) - kill $QEMU_NBD_PID + if [ -n "$QEMU_NBD_PID" ]; then + kill $QEMU_NBD_PID + fi rm -f "$TEST_IMG_FILE" ;; file)
Signature · 1 line
1.9.1
PATCH 9/43 iotests: Add test for unsupported image creation
In reply to Michael RothFrom: Max Reitz <[email protected]>
Add a test for creating and amending images (amendment uses the creation options) with formats not supporting creation over protocols not supporting creation.
Conflicts:
tests/qemu-iotests/group
*removed context dependency on iotest group 114
--- tests/qemu-iotests/113 | 76 ++++++++++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/113.out | 15 +++++++++ tests/qemu-iotests/group | 1 + 3 files changed, 92 insertions(+) create mode 100755 tests/qemu-iotests/113 create mode 100644 tests/qemu-iotests/113.out
tests/qemu-iotests/113 new file +76 -0
@@ -0,0 +1,76 @@ +#!/bin/bash +# +# Test case for accessing creation options on image formats and +# protocols not supporting image creation +# +# Copyright (C) 2014 Red Hat, Inc. +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. +# + +# creator [email protected] + +seq="$(basename $0)" +echo "QA output created by $seq" + +here="$PWD" +tmp=/tmp/$$ +status=1 # failure is the default! + +_cleanup() +{ + _cleanup_test_img +} +trap "_cleanup; exit \$status" 0 1 2 3 15 + +# get standard environment, filters and checks +. ./common.rc +. ./common.filter + +# We can only test one format here because we need its sample file +_supported_fmt bochs +_supported_proto nbd +_supported_os Linux + +echo +echo '=== Unsupported image creation in qemu-img create ===' +echo + +$QEMU_IMG create -f $IMGFMT nbd://example.com 2>&1 64M | _filter_imgfmt + +echo +echo '=== Unsupported image creation in qemu-img convert ===' +echo + +# We could use any input image format here, but this is a bochs test, so just +# use the bochs image +_use_sample_img empty.bochs.bz2 +$QEMU_IMG convert -f $IMGFMT -O $IMGFMT "$TEST_IMG" nbd://example.com 2>&1 \ + | _filter_imgfmt + +echo +echo '=== Unsupported format in qemu-img amend ===' +echo + +# The protocol does not matter here +_use_sample_img empty.bochs.bz2 +$QEMU_IMG amend -f $IMGFMT -o foo=bar "$TEST_IMG" 2>&1 | _filter_imgfmt + + +# success, all done +echo +echo '*** done' +rm -f $seq.full +status=0
tests/qemu-iotests/113.out new file +15 -0
@@ -0,0 +1,15 @@ +QA output created by 113 + +=== Unsupported image creation in qemu-img create === + +qemu-img: nbd://example.com: Format driver 'IMGFMT' does not support image creation + +=== Unsupported image creation in qemu-img convert === + +qemu-img: Format driver 'IMGFMT' does not support image creation + +=== Unsupported format in qemu-img amend === + +qemu-img: Format driver 'IMGFMT' does not support any options to amend + +*** done
tests/qemu-iotests/group +1 -0
@@ -112,3 +112,4 @@ 107 rw auto quick 108 rw auto quick 111 rw auto quick +113 rw auto quick
Signature · 1 line
1.9.1
PATCH 10/43 qcow2: Prevent numerical overflow
In reply to Michael RothFrom: Max Reitz <[email protected]>
In qcow2_alloc_cluster_offset(), *num is limited toINT_MAX >> BDRV_SECTOR_BITS by all callers. However, since remaining is of type uint64_t, we might as well cast *num to that type before performing the shift.
--- block/qcow2-cluster.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
block/qcow2-cluster.c +1 -1
@@ -1263,7 +1263,7 @@ int qcow2_alloc_cluster_offset(BlockDriverState *bs, uint64_t offset, again: start = offset; - remaining = *num <<BDRV_SECTOR_BITS ; + remaining = (uint64_t )*num <<BDRV_SECTOR_BITS ; cluster_offset =0 ; *host_offset =0 ; cur_bytes =0 ;
Signature · 1 line
1.9.1
PATCH 11/43 qcow2: Flushing the caches in qcow2_close may fail
In reply to Michael RothFrom: Max Reitz <[email protected]>
qcow2_cache_flush() may fail; if one of the caches failed to be flushed successfully to disk in qcow2_close() the image should not be marked clean, and we should emit a warning.
This breaks the (qcow2-specific) iotests 026, 071 and 089; change their output accordingly.
--- block/qcow2.c | 19 +++++-- tests/qemu-iotests/026.out | 120 +++++++++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/071.out | 8 +++ tests/qemu-iotests/089.out | 2 + 4 files changed, 146 insertions(+), 3 deletions(-)
block/qcow2.c +16 -3
@@ -1428,10 +1428,23 @@ static void qcow2_close(BlockDriverState *bs) s->l1_table =NULL ;if (!(bs->open_flags &BDRV_O_INCOMING )) { -qcow2_cache_flush (bs, s->l2_table_cache); -qcow2_cache_flush (bs, s->refcount_block_cache); +int ret1, ret2; -qcow2_mark_clean (bs); + ret1 =qcow2_cache_flush (bs, s->l2_table_cache); + ret2 =qcow2_cache_flush (bs, s->refcount_block_cache); + +if (ret1) { +error_report ("Failed to flush the L2 table cache: %s" , +strerror (-ret1)); + } +if (ret2) { +error_report ("Failed to flush the refcount block cache: %s" , +strerror (-ret2)); + } + +if (!ret1 && !ret2) { +qcow2_mark_clean (bs); + } }qcow2_cache_destroy (bs, s->l2_table_cache);
tests/qemu-iotests/026.out +120 -0
464 lines not on this page · raw message
tests/qemu-iotests/071.out +8 -0
@@ -30,10 +30,14 @@ blkverify: read sector_num=0 nb_sectors=4 contents mismatch in sector 0 === Testing blkdebug through filename === +Failed to flush the L2 table cache: Input/output error +Failed to flush the refcount block cache: Input/output error read failed: Input/output error === Testing blkdebug through file blockref === +Failed to flush the L2 table cache: Input/output error +Failed to flush the refcount block cache: Input/output error read failed: Input/output error === Testing blkdebug on existing block device ===
@@ -48,6 +52,8 @@ read failed: Input/output error {"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN"} {"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "DEVICE_TRAY_MOVED", "data": {"device": "ide1-cd0", "tray-open": true}} {"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "DEVICE_TRAY_MOVED", "data": {"device": "floppy0", "tray-open": true}} +qemu-system-x86_64: Failed to flush the L2 table cache: Input/output error +qemu-system-x86_64: Failed to flush the refcount block cache: Input/output error === Testing blkverify on existing block device ===
@@ -86,5 +92,7 @@ read failed: Input/output error {"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN"} {"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "DEVICE_TRAY_MOVED", "data": {"device": "ide1-cd0", "tray-open": true}} {"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "DEVICE_TRAY_MOVED", "data": {"device": "floppy0", "tray-open": true}} +qemu-system-x86_64: Failed to flush the L2 table cache: Input/output error +qemu-system-x86_64: Failed to flush the refcount block cache: Input/output error *** done
tests/qemu-iotests/089.out +2 -0
@@ -24,6 +24,8 @@ read 512/512 bytes at offset 0 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 wrote 512/512 bytes at offset 229376 512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +Failed to flush the L2 table cache: Input/output error +Failed to flush the refcount block cache: Input/output error read failed: Input/output error === Testing qemu-img info output ===
Signature · 1 line
1.9.1
PATCH 12/43 qcow2: Respect bdrv_truncate() error
In reply to Michael RothFrom: Max Reitz <[email protected]>
bdrv_truncate() may fail and qcow2_write_compressed() should return the error code in that case.
--- block/qcow2.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
block/qcow2.c +1 -2
@@ -2162,8 +2162,7 @@ static int qcow2_write_compressed(BlockDriverState *bs, int64_t sector_num,/* align end of file to a sector boundary to ease reading with sector based I/Os */ cluster_offset =bdrv_getlength (bs->file); -bdrv_truncate (bs->file, cluster_offset); -return 0 ; +return bdrv_truncate (bs->file, cluster_offset); }if (nb_sectors != s->cluster_sectors) {
Signature · 1 line
1.9.1
PATCH 13/43 block/raw-posix: Fix ret in raw_open_common()
In reply to Michael RothFrom: Max Reitz <[email protected]>
The return value must be negative on error; there is one place in raw_open_common() where errp is set, but ret remains 0. Fix it.
--- block/raw-posix.c | 1 + 1 file changed, 1 insertion(+)
block/raw-posix.c +1 -0
@@ -446,6 +446,7 @@ static int raw_open_common(BlockDriverState *bs, QDict *options, }if (fstat (s->fd, &st) <0 ) { + ret = -errno;error_setg_errno (errp, errno,"Could not stat file" ); goto fail; }
Signature · 1 line
1.9.1
PATCH 14/43 block migration: fix return value
In reply to Michael RothFrom: Gary R Hook <[email protected]>
Modify block_save_iterate() to return positive/zero/negative (success/not done/failure) return status. The computation of the blocks transferred (an int64_t) exceeds the size of an int return value.
Message-id: [email protected]
--- block-migration.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-)
block-migration.c +9 -1
@@ -653,6 +653,7 @@ static int block_save_iterate(QEMUFile *f, void *opaque) {int ret;int64_t last_ftell =qemu_ftell (f); +int64_t delta_ftell;DPRINTF ("Enter save live iterate submitted %d transferred %d\n" , block_mig_state.submitted, block_mig_state.transferred);
@@ -702,7 +703,14 @@ static int block_save_iterate(QEMUFile *f, void *opaque) }qemu_put_be64 (f,BLK_MIG_FLAG_EOS ); -return qemu_ftell (f) - last_ftell; + delta_ftell =qemu_ftell (f) - last_ftell; +if (delta_ftell >0 ) { +return 1 ; + }else if (delta_ftell <0 ) { +return -1 ; + }else { +return 0 ; + } }/* Called with iothread lock taken. */
Signature · 1 line
1.9.1
PATCH 15/43 qcow2: Fix header extension size check
In reply to Michael RothFrom: Kevin Wolf <[email protected]>
After reading the extension header, offset is incremented, but not checked against end_offset any more. This way an integer overflow could happen when checking whether the extension end is within the allowed range, effectively disabling the check.
This patch adds the missing check and a test case for it.
Message-id: [email protected]
--- block/qcow2.c | 2 +- tests/qemu-iotests/080 | 2 ++ tests/qemu-iotests/080.out | 2 ++ 3 files changed, 5 insertions(+), 1 deletion(-)
block/qcow2.c +1 -1
@@ -117,7 +117,7 @@ static int qcow2_read_extensions(BlockDriverState *bs, uint64_t start_offset,#ifdef DEBUG_EXT printf ("ext.magic = 0x%x\n" , ext.magic);#endif -if (ext.len > end_offset - offset) { +if (offset > end_offset || ext.len > end_offset - offset) {error_setg (errp,"Header extension too large" );return -EINVAL ; }
tests/qemu-iotests/080 +2 -0
@@ -78,6 +78,8 @@ poke_file "$TEST_IMG" "$offset_backing_file_offset" "\xff\xff\xff\xff\xff\xff\xf poke_file "$TEST_IMG" "$offset_ext_magic" "\x12\x34\x56\x78" poke_file "$TEST_IMG" "$offset_ext_size" "\x7f\xff\xff\xff" { $QEMU_IO -c "read 0 512" $TEST_IMG; } 2>&1 | _filter_qemu_io | _filter_testdir +poke_file "$TEST_IMG" "$offset_backing_file_offset" "\x00\x00\x00\x00\x00\x00\x00\x$(printf %x $offset_ext_size)" +{ $QEMU_IO -c "read 0 512" $TEST_IMG; } 2>&1 | _filter_qemu_io | _filter_testdir poke_file "$TEST_IMG" "$offset_backing_file_offset" "\x00\x00\x00\x00\x00\x00\x00\x00" { $QEMU_IO -c "read 0 512" $TEST_IMG; } 2>&1 | _filter_qemu_io | _filter_testdir
tests/qemu-iotests/080.out +2 -0
@@ -13,6 +13,8 @@ qemu-io: can't open device TEST_DIR/t.qcow2: Invalid backing file offset no file open, try 'help open' qemu-io: can't open device TEST_DIR/t.qcow2: Header extension too large no file open, try 'help open' +qemu-io: can't open device TEST_DIR/t.qcow2: Header extension too large +no file open, try 'help open' == Huge refcount table size == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
Signature · 1 line
1.9.1
PATCH 16/43 qcow2.py: Add required padding for header extensions
In reply to Michael RothFrom: Kevin Wolf <[email protected]>
The qcow2 specification requires that the header extension data be padded to round up the extension size to the next multiple of 8 bytes.
Message-id: [email protected]
--- tests/qemu-iotests/qcow2.py | 4 ++++ 1 file changed, 4 insertions(+)
tests/qemu-iotests/qcow2.py +4 -0
@@ -7,6 +7,10 @@ import string class QcowHeaderExtension: def __init__(self, magic, length, data): + if length % 8 != 0: + padding = 8 - (length % 8) + data += "\0" * padding + self.magic = magic self.length = length self.data = data
Signature · 1 line
1.9.1
PATCH 17/43 block: Don't probe for unknown backing file format
In reply to Michael RothFrom: Kevin Wolf <[email protected]>
If a qcow2 image specifies a backing file format that doesn't correspond to any format driver that qemu knows, we shouldn't fall back to probing, but simply error out.
Not looking up the backing file driver in bdrv_open_backing_file(), but just filling in the "driver" option if it isn't there moves us closer to the goal of having everything in QDict options and gets us the error handling of bdrv_open(), which correctly refuses unknown drivers.
Message-id: [email protected]
Conflicts:
tests/qemu-iotests/group
*resolved context conflict due to group 113 being present locally
--- block.c | 7 +++--- tests/qemu-iotests/114 | 61 ++++++++++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/114.out | 13 ++++++++++ tests/qemu-iotests/group | 1 + 4 files changed, 78 insertions(+), 4 deletions(-) create mode 100755 tests/qemu-iotests/114 create mode 100644 tests/qemu-iotests/114.out
block.c +3 -4
@@ -1175,7 +1175,6 @@ int bdrv_open_backing_file(BlockDriverState *bs, QDict *options, Error **errp) {char *backing_filename =g_malloc0 (PATH_MAX );int ret =0 ; -BlockDriver *back_drv =NULL ;BlockDriverState *backing_hd;Error *local_err =NULL ;
@@ -1208,14 +1207,14 @@ int bdrv_open_backing_file(BlockDriverState *bs, QDict *options, Error **errp) backing_hd =bdrv_new (); -if (bs->backing_format[0 ] !='\0' ) { - back_drv =bdrv_find_format (bs->backing_format); +if (bs->backing_format[0 ] !='\0' && !qdict_haskey (options,"driver" )) { +qdict_put (options,"driver" ,qstring_from_str (bs->backing_format)); }assert (bs->backing_hd ==NULL ); ret =bdrv_open (&backing_hd, *backing_filename ? backing_filename :NULL ,NULL , options, -bdrv_backing_flags (bs->open_flags), back_drv, &local_err); +bdrv_backing_flags (bs->open_flags),NULL , &local_err);if (ret <0 ) {bdrv_unref (backing_hd); backing_hd =NULL ;
tests/qemu-iotests/114 new file +61 -0
61 lines not on this page · raw message
tests/qemu-iotests/114.out new file +13 -0
@@ -0,0 +1,13 @@ +QA output created by 114 +Formatting 'TEST_DIR/t.IMGFMT.base', fmt=IMGFMT size=67108864 +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 backing_file='TEST_DIR/t.IMGFMT.base' +image: TEST_DIR/t.IMGFMT +file format: IMGFMT +virtual size: 64M (67108864 bytes) +cluster_size: 65536 +backing file: TEST_DIR/t.IMGFMT.base +backing file format: foo +qemu-io: can't open device TEST_DIR/t.qcow2: Could not open backing file: Unknown driver 'foo' +read 4096/4096 bytes at offset 0 +4 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +*** done
tests/qemu-iotests/group +1 -0
@@ -113,3 +113,4 @@ 108 rw auto quick 111 rw auto quick 113 rw auto quick +114 rw auto quick
Signature · 1 line
1.9.1
PATCH 18/43 linuxboot: fix loading old kernels
In reply to Michael RothFrom: Paolo Bonzini <[email protected]>
Old kernels that used high memory only allowed the initrd to be in the first 896MB of memory. If you load the initrd above, they complain that "initrd extends beyond end of memory".
In order to fix this, while not breaking machines with small amounts of memory fixed by cdebec5 (linuxboot: compute initrd loading address, 2014-10-06), we need to distinguish two cases. If pc.c placed the initrd at end of memory, use the new algorithm based on the e801 memory map. If instead pc.c placed the initrd at the maximum address specified by the bzImage, leave it there.
The only interesting part is that the low-memory info block is now loaded very early, in real mode, and thus the 32-bit address has to be converted into a real mode segment. The initrd address is also patched in the info block before entering real mode, it is simpler that way.
This fixes booting the RHEL4.8 32-bit installation image with 1GB of RAM.
--- pc-bios/linuxboot.bin | Bin 1024 -> 1024 bytes pc-bios/optionrom/linuxboot.S | 37 +++++++++++++++++++++++++++---------- 2 files changed, 27 insertions(+), 10 deletions(-)
pc-bios/linuxboot.bin binary +0 -0
pc-bios/optionrom/linuxboot.S +27 -10
@@ -76,7 +76,31 @@ boot_kernel: copy_kernel: - /* Compute initrd address */ + /* Read info block in low memory (0x10000 or 0x90000) */ + read_fw FW_CFG_SETUP_ADDR + shr $4, %eax + mov %eax, %es + xor %edi, %edi + read_fw_blob_addr32_edi(FW_CFG_SETUP) + + cmpw $0x203, %es:0x206 // if protocol >= 0x203 + jae 1f // have initrd_max + movl $0x37ffffff, %es:0x22c // else assume 0x37ffffff +1: + + /* Check if using kernel-specified initrd address */ + read_fw FW_CFG_INITRD_ADDR + mov %eax, %edi // (load_kernel wants it in %edi) + read_fw FW_CFG_INITRD_SIZE // find end of initrd + add %edi, %eax + xor %es:0x22c, %eax // if it matches es:0x22c + and $-4096, %eax // (apart from padding for page) + jz load_kernel // then initrd is not at top + // of memory + + /* pc.c placed the initrd at end of memory. Compute a better + * initrd address based on e801 data. + */ mov $0xe801, %ax xor %cx, %cx xor %dx, %dx
@@ -107,7 +131,9 @@ copy_kernel: read_fw FW_CFG_INITRD_SIZE subl %eax, %edi andl $-4096, %edi /* EDI = start of initrd */ + movl %edi, %es:0x218 /* put it in the header */ +load_kernel: /* We need to load the kernel into memory we can't access in 16 bit mode, so let's get into 32 bit mode, write the kernel and jump back again. */
@@ -139,19 +165,10 @@ copy_kernel: /* We're now running in 16-bit CS, but 32-bit ES! */ /* Load kernel and initrd */ - pushl %edi read_fw_blob_addr32_edi(FW_CFG_INITRD) read_fw_blob_addr32(FW_CFG_KERNEL) read_fw_blob_addr32(FW_CFG_CMDLINE) - read_fw FW_CFG_SETUP_ADDR - mov %eax, %edi - mov %eax, %ebx - read_fw_blob_addr32_edi(FW_CFG_SETUP) - - /* Update the header with the initrd address we chose above */ - popl %es:0x218(%ebx) - /* And now jump into Linux! */ mov $0, %eax mov %eax, %cr0
Signature · 1 line
1.9.1
PATCH 19/43 audio: Don't free hw resources until after hw backend is stopped
In reply to Michael RothFrom: Peter Maydell <[email protected]>
When stopping an audio voice, call the audio backend's fini method before calling audio_pcm_hw_free_resources_ rather than afterwards. This allows backends which use helper threads (like pulseaudio) to terminate those threads before the conv_buf or mix_buf are freed and avoids race conditions where the helper may access a NULL pointer or freed memory.
Message-id: [email protected]
--- audio/audio_template.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
audio/audio_template.h +1 -1
@@ -191,9 +191,9 @@ static void glue (audio_pcm_hw_gc_, TYPE) (HW **hwp)audio_detach_capture (hw);#endif QLIST_REMOVE (hw, entries); +glue (hw->pcm_ops->fini_,TYPE ) (hw);glue (s->nb_hw_voices_,TYPE ) +=1 ;glue (audio_pcm_hw_free_resources_ ,TYPE ) (hw); -glue (hw->pcm_ops->fini_,TYPE ) (hw);g_free (hw); *hwp =NULL ; }
Signature · 1 line
1.9.1
PATCH 20/43 target-xtensa: fix translation for opcodes crossing page boundary
In reply to Michael RothFrom: Max Filippov <[email protected]>
If TB ends with an opcode that crosses page boundary and the following page is not executable then EPC1 for the code fetch exception wrongly points at the beginning of the TB. Always treat instruction that crosses page boundary as a separate TB.
--- target-xtensa/translate.c | 27 +++++++++++++++++++++++---- 1 file changed, 23 insertions(+), 4 deletions(-)
target-xtensa/translate.c +23 -4
@@ -887,6 +887,11 @@ static TCGv_i32 gen_mac16_m(TCGv_i32 v, bool hi, bool is_unsigned)return m; } +static inline unsigned xtensa_op0_insn_len (unsigned op0) +{ +return op0 >=8 ?2 :3 ; +} +static void disas_xtensa_insn (CPUXtensaState *env,DisasContext *dc) {#define HAS_OPTION_BITS (opt) do { \
@@ -989,6 +994,7 @@ static void disas_xtensa_insn(CPUXtensaState *env, DisasContext *dc)uint8_t b0 =cpu_ldub_code (env, dc->pc);uint8_t b1 =cpu_ldub_code (env, dc->pc +1 );uint8_t b2 =0 ; +unsigned len =xtensa_op0_insn_len (OP0 );static const uint32_t B4CONST [] = {0xffffffff ,1 ,2 ,3 ,4 ,5 ,6 ,7 ,8 ,10 ,12 ,16 ,32 ,64 ,128 ,256
@@ -998,13 +1004,19 @@ static void disas_xtensa_insn(CPUXtensaState *env, DisasContext *dc)32768 ,65536 ,2 ,3 ,4 ,5 ,6 ,7 ,8 ,10 ,12 ,16 ,32 ,64 ,128 ,256 }; -if (OP0 >=8 ) { - dc->next_pc = dc->pc +2 ; +switch (len) { +case 2 :HAS_OPTION (XTENSA_OPTION_CODE_DENSITY ); - }else { - dc->next_pc = dc->pc +3 ; +break ; + +case 3 : b2 =cpu_ldub_code (env, dc->pc +2 ); +break ; + +default : +RESERVED (); } + dc->next_pc = dc->pc + len;switch (OP0 ) {case 0 :/*QRST*/
@@ -2949,6 +2961,12 @@ invalid_opcode:#undef HAS_OPTION } +static inline unsigned xtensa_insn_len (CPUXtensaState *env,DisasContext *dc) +{ +uint8_t b0 =cpu_ldub_code (env, dc->pc); +return xtensa_op0_insn_len (OP0 ); +} +static void check_breakpoint (CPUXtensaState *env,DisasContext *dc) {CPUState *cs =CPU (xtensa_env_get_cpu (env));
@@ -3081,6 +3099,7 @@ void gen_intermediate_code_internal(XtensaCPU *cpu, }while (dc.is_jmp ==DISAS_NEXT && insn_count < max_insns && dc.pc < next_page_start && + dc.pc +xtensa_insn_len (env, &dc) <= next_page_start && tcg_ctx.gen_opc_ptr < gen_opc_end);reset_litbase (&dc);
Signature · 1 line
1.9.1
PATCH 21/43 target-xtensa: test cross-page opcode
In reply to Michael RothFrom: Max Filippov <[email protected]>
Alter cross-page TB test to also test cross-page opcode.
--- tests/tcg/xtensa/test_mmu.S | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-)
tests/tcg/xtensa/test_mmu.S +13 -13
@@ -641,7 +641,7 @@ test cross_page_tb witlb a2, a3 wdtlb a2, a3 - movi a2, 0x00007ffd + movi a2, 0x00007ffc movi a3, 20f movi a4, 21f sub a4, a4, a3
@@ -651,7 +651,7 @@ test cross_page_tb addi a2, a2, 1 addi a3, a3, 1 1: - movi a2, 0x00007ffd + movi a2, 0x00007ffc movi a3, 0x00008000 /* DTLB: OK, ITLB: OK */ jx a2
@@ -668,10 +668,10 @@ test cross_page_tb movi a3, 1 assert eq, a2, a3 rsr a2, epc1 - movi a3, 0x8000 + movi a3, 0x7fff assert eq, a2, a3 rsr a2, excsave1 - movi a3, 0x00007ffd + movi a3, 0x00007ffc assert ne, a2, a3 reset_ps
@@ -680,7 +680,7 @@ test cross_page_tb movi a2, 0x0400000c /* PPN */ movi a3, 0x00008000 /* VPN */ wdtlb a2, a3 - movi a2, 0x00007ffd + movi a2, 0x00007ffc movi a3, 0x00008000 /* DTLB: FAIL, ITLB: OK */ jx a2
@@ -689,10 +689,10 @@ test cross_page_tb movi a3, 28 assert eq, a2, a3 rsr a2, epc1 - movi a3, 0x7ffd + movi a3, 0x7ffc assert eq, a2, a3 rsr a2, excsave1 - movi a3, 0x00007ffd + movi a3, 0x00007ffc assert eq, a2, a3 reset_ps
@@ -703,7 +703,7 @@ test cross_page_tb witlb a2, a3 movi a2, 0x04000003 /* PPN */ wdtlb a2, a3 - movi a2, 0x00007ffd + movi a2, 0x00007ffc movi a3, 0x00008000 /* DTLB: OK, ITLB: FAIL */ jx a2
@@ -712,10 +712,10 @@ test cross_page_tb movi a3, 20 assert eq, a2, a3 rsr a2, epc1 - movi a3, 0x8000 + movi a3, 0x7fff assert eq, a2, a3 rsr a2, excsave1 - movi a3, 0x00007ffd + movi a3, 0x00007ffc assert ne, a2, a3 reset_ps
@@ -724,7 +724,7 @@ test cross_page_tb movi a2, 0x0400000c /* PPN */ movi a3, 0x00008000 /* VPN */ wdtlb a2, a3 - movi a2, 0x00007ffd + movi a2, 0x00007ffc movi a3, 0x00008000 /* DTLB: FAIL, ITLB: FAIL */ jx a2
@@ -733,10 +733,10 @@ test cross_page_tb movi a3, 28 assert eq, a2, a3 rsr a2, epc1 - movi a3, 0x7ffd + movi a3, 0x7ffc assert eq, a2, a3 rsr a2, excsave1 - movi a3, 0x00007ffd + movi a3, 0x00007ffc assert eq, a2, a3 test_end
Signature · 1 line
1.9.1
PATCH 22/43 migration/block: fix pending() return value
In reply to Michael RothFrom: Vladimir Sementsov-Ogievskiy <[email protected]>
Because of wrong return value of .save_live_pending() in migration/block.c, migration finishes before the whole disk is transferred. Such situation occurs when the migration process is fast enough, for example when source and dest are on the same host.
If in the bulk phase we return something < max_size, we will skip transferring the tail of the device. Currently we have "set pending to BLOCK_SIZE if it is zero" for bulk phase, but there no guarantee, that it will be < max_size.
True approach is to return, for example, max_size+1 when we are in the bulk phase.
Message-id: [email protected]
--- block-migration.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
block-migration.c +2 -2
@@ -765,8 +765,8 @@ static uint64_t block_save_pending(QEMUFile *f, void *opaque, uint64_t max_size) block_mig_state.read_done *BLOCK_SIZE ;/* Report at least one block pending during bulk phase */ -if (pending ==0 && !block_mig_state.bulk_completed) { - pending =BLOCK_SIZE ; +if (pending <= max_size && !block_mig_state.bulk_completed) { + pending = max_size +BLOCK_SIZE ; }blk_mig_unlock ();qemu_mutex_unlock_iothread ();
Signature · 1 line
1.9.1
PATCH 23/43 atomic: fix position of volatile qualifier
In reply to Michael RothFrom: Paolo Bonzini <[email protected]>
What needs to be volatile is not the pointer, but the pointed-to value!
--- include/qemu/atomic.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
include/qemu/atomic.h +2 -2
@@ -122,11 +122,11 @@#endif #ifndef atomic_read -#define atomic_read (ptr) (*(__typeof__(*ptr) *volatile) (ptr)) +#define atomic_read (ptr) (*(__typeof__(*ptr) volatile*) (ptr))#endif #ifndef atomic_set -#define atomic_set (ptr, i) ((*(__typeof__(*ptr) *volatile) (ptr)) = (i)) +#define atomic_set (ptr, i) ((*(__typeof__(*ptr) volatile*) (ptr)) = (i))#endif /* These have the same semantics as Java volatile variables.
Signature · 1 line
1.9.1
PATCH 24/43 PPC: Fix crash on spapr_tce_table_finalize()
In reply to Michael RothFrom: David Gibson <[email protected]>
spapr_tce_table_finalize() can SEGV if the object was not previously realized. In particular this can be triggered by running
qemu-system-ppc -device spapr-tce-table,?
The basic problem is that we have mismatched initialization versus
finalization: spapr_tce_table_finalize() is attempting to undo things that are done in spapr_tce_table_realize(), not an instance_init function.
Therefore, replace spapr_tce_table_finalize() with spapr_tce_table_unrealize().
--- hw/ppc/spapr_iommu.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
hw/ppc/spapr_iommu.c +3 -3
@@ -173,9 +173,9 @@ sPAPRTCETable *spapr_tce_new_table(DeviceState *owner, uint32_t liobn,return tcet; } -static void spapr_tce_table_finalize (Object *obj) +static void spapr_tce_table_unrealize (DeviceState *dev,Error **errp) { -sPAPRTCETable *tcet =SPAPR_TCE_TABLE (obj); +sPAPRTCETable *tcet =SPAPR_TCE_TABLE (dev);QLIST_REMOVE (tcet, list);
@@ -420,6 +420,7 @@ static void spapr_tce_table_class_init(ObjectClass *klass, void *data)DeviceClass *dc =DEVICE_CLASS (klass); dc->init = spapr_tce_table_realize; dc->reset = spapr_tce_reset; + dc->unrealize = spapr_tce_table_unrealize;QLIST_INIT (&spapr_tce_tables);
@@ -435,7 +436,6 @@ static TypeInfo spapr_tce_table_info = { .parent = TYPE_DEVICE, .instance_size = sizeof(sPAPRTCETable), .class_init = spapr_tce_table_class_init, - .instance_finalize = spapr_tce_table_finalize, };static void register_types (void )
Signature · 1 line
1.9.1
PATCH 25/43 vl.c: fix regression when reading machine type from config file
In reply to Michael RothFrom: Marcel Apfelbaum <[email protected]>
After 'Machine as QOM' series the machine type input triggers the creation of the machine class.
If the machine type is set in the configuration file, the machine class is not updated accordingly and remains the default.
Fixed that by querying the machine options after the configuration file is loaded.
--- vl.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-)
vl.c +8 -7
@@ -2886,9 +2886,6 @@ int main(int argc, char **argv, char **envp)exit (1 ); }switch (popt->index) { -case QEMU_OPTION_M : - machine_class =machine_parse (optarg); -break ;case QEMU_OPTION_no_kvm_irqchip: { olist =qemu_find_opts ("machine" );qemu_opts_parse (olist,"kernel_irqchip=off" ,0 );
@@ -3506,16 +3503,13 @@ int main(int argc, char **argv, char **envp) olist =qemu_find_opts ("machine" );qemu_opts_parse (olist,"accel=kvm" ,0 );break ; +case QEMU_OPTION_M :case QEMU_OPTION_machine: olist =qemu_find_opts ("machine" ); opts =qemu_opts_parse (olist, optarg,1 );if (!opts) {exit (1 ); } - optarg =qemu_opt_get (opts,"type" ); -if (optarg) { - machine_class =machine_parse (optarg); - }break ;case QEMU_OPTION_no_kvm: olist =qemu_find_opts ("machine" );
@@ -3807,6 +3801,13 @@ int main(int argc, char **argv, char **envp) } } } + + opts =qemu_get_machine_opts (); + optarg =qemu_opt_get (opts,"type" ); +if (optarg) { + machine_class =machine_parse (optarg); + } +loc_set_none ();os_daemonize ();
Signature · 1 line
1.9.1
PATCH 26/43 serial: reset thri_pending on IER writes with THRI=0
In reply to Michael RothFrom: Paolo Bonzini <[email protected]>
This is responsible for failure of migration from 2.2 to 2.1, because thr_ipending is always one in practice.
serial.c is setting thr_ipending unconditionally. However, thr_ipending is not used at all if THRI=0, and it will be overwritten again the next time THRE or THRI changes. For that reason, we can set thr_ipending to zero every time THRI is reset.
There is disagreement on whether LSR.THRE should be resampled when IER.THRI goes from 1 to 1. This patch does not touch the code, leaving that for QEMU 2.3+.
This has no semantic change and is enough to fix migration in the common case where the interrupt is not pending or is reported in IIR. It does not change the migration format, so 2.2.0 -> 2.1 will remain broken but we can fix 2.2.1 -> 2.1 without breaking 2.2.1 <-> 2.2.0.
The case that remains broken (the one in which the subsection is strictly necessary) is when THRE=1, the THRI interrupt has *not* been acknowledged yet, and a higher-priority interrupt comes. In this case, you need the subsection to tell the source that the lower-priority THRI interrupt is pending. The subsection's breakage of migration, in this case, prevents continuing the VM on the destination with an invalid state.
--- hw/char/serial.c | 18 ++++++++++++++++-- 1 file changed, 16 insertions(+), 2 deletions(-)
hw/char/serial.c +16 -2
@@ -350,10 +350,24 @@ static void serial_ioport_write(void *opaque, hwaddr addr, uint64_t val, s->poll_msl =0 ; } } -if (s->lsr &UART_LSR_THRE ) { + +/* Turning on the THRE interrupt on IER can trigger the interrupt +* if LSR.THRE=1, even if it had been masked before by reading IIR. +* This is not in the datasheet, but Windows relies on it. It is +* unclear if THRE has to be resampled every time THRI becomes +* 1, or only on the rising edge. Bochs does the latter, and Windows +* always toggles IER to all zeroes and back to all ones. But for +* now leave it as it has always been in QEMU. +* +* If IER.THRI is zero, thr_ipending is not used. Set it to zero +* so that the thr_ipending subsection is not migrated. +*/ +if ((s->ier &UART_IER_THRI ) && (s->lsr &UART_LSR_THRE )) { s->thr_ipending =1 ; -serial_update_irq (s); + }else { + s->thr_ipending =0 ; } +serial_update_irq (s); }break ;case 2 :
Signature · 1 line
1.9.1
PATCH 27/43 serial: refine serial_thr_ipending_needed
In reply to Michael RothFrom: Paolo Bonzini <[email protected]>
If the THR interrupt is disabled, there is no need to migrate thr_ipending because LSR.THRE will be sampled again when the interrupt is enabled. (This is the behavior that is not documented in the datasheet, but relied on by Windows!)
Note that in this case IIR will never be 0x2 so, if thr_ipending were to be one, QEMU would produce the subsection.
--- hw/char/serial.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-)
hw/char/serial.c +11 -2
@@ -637,8 +637,17 @@ static int serial_post_load(void *opaque, int version_id)static bool serial_thr_ipending_needed (void *opaque) {SerialState *s = opaque; -bool expected_value = ((s->iir &UART_IIR_ID ) ==UART_IIR_THRI ); -return s->thr_ipending != expected_value; + +if (s->ier &UART_IER_THRI ) { +bool expected_value = ((s->iir &UART_IIR_ID ) ==UART_IIR_THRI ); +return s->thr_ipending != expected_value; + }else { +/* LSR.THRE will be sampled again when the interrupt is +* enabled. thr_ipending is not used in this case, do +* not migrate it. +*/ +return false; + } }const VMStateDescription vmstate_serial_thr_ipending = {
Signature · 1 line
1.9.1
PATCH 28/43 pckbd: set bits 2-3-6-7 of the output port by default
In reply to Michael RothFrom: Paolo Bonzini <[email protected]>
OSes typically write 0xdd/0xdf to turn the A20 line off and on. This has bits 2-3-6-7 on, so that the output port subsection is migrated. Change the reset value and migration default to include those four bits, thus avoiding that the subsection is migrated.
This strictly speaking changes guest ABI, but the long time during which we have not migrated the value means that the guests really do not care much; so the change is for all machine types.
--- hw/input/pckbd.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-)
hw/input/pckbd.c +8 -2
@@ -101,6 +101,12 @@#define KBD_OUT_OBF 0x10/* Keyboard output buffer full */ #define KBD_OUT_MOUSE_OBF 0x20/* Mouse output buffer full */ +/* OSes typically write 0xdd/0xdf to turn the A20 line off and on. +* We make the default value of the outport include these four bits, +* so that the subsection is rarely necessary. +*/ +#define KBD_OUT_ONES 0xcc +/* Mouse Commands */ #define AUX_SET_SCALE11 0xE6/* Set 1:1 scaling */ #define AUX_SET_SCALE21 0xE7/* Set 2:1 scaling */
@@ -367,13 +373,13 @@ static void kbd_reset(void *opaque) s->mode =KBD_MODE_KBD_INT |KBD_MODE_MOUSE_INT ; s->status =KBD_STAT_CMD |KBD_STAT_UNLOCKED ; - s->outport =KBD_OUT_RESET |KBD_OUT_A20 ; + s->outport =KBD_OUT_RESET |KBD_OUT_A20 |KBD_OUT_ONES ; s->outport_present = false; }static uint8_t kbd_outport_default (KBDState *s) { -return KBD_OUT_RESET |KBD_OUT_A20 +return KBD_OUT_RESET |KBD_OUT_A20 |KBD_OUT_ONES | (s->status &KBD_STAT_OBF ?KBD_OUT_OBF :0 ) | (s->status &KBD_STAT_MOUSE_OBF ?KBD_OUT_MOUSE_OBF :0 ); }
Signature · 1 line
1.9.1
PATCH 29/43 linux-user: Fix broken m68k signal handling on 64 bit hosts
In reply to Michael RothFrom: Peter Maydell <[email protected]>
The m68k signal frame setup code which writes the signal return trampoline code to the stack was assuming that a 'long' was 32 bits; on 64 bit systems this meant we would end up writing the 32 bit (2 insn) trampoline sequence to retaddr+4,retaddr+6 instead of the intended retaddr+0,retaddr+2, resulting in a guest crash when it tried to execute the invalid zero-bytes at retaddr+0. Fix by using uint32_t instead; also use uint16_t rather than short for consistency. This fixes bug LP:1404690.
--- linux-user/signal.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
linux-user/signal.c +3 -3
@@ -5091,7 +5091,7 @@ static void setup_frame(int sig, struct target_sigaction *ka,/* moveq #,d0; trap #0 */ __put_user (0x70004e40 + (TARGET_NR_sigreturn <<16 ), - (long *)(frame->retcode)); + (uint32_t *)(frame->retcode));/* Set up to return from userspace */
@@ -5225,8 +5225,8 @@ static void setup_rt_frame(int sig, struct target_sigaction *ka,/* moveq #,d0; notb d0; trap #0 */ __put_user (0x70004600 + ((TARGET_NR_rt_sigreturn ^0xff ) <<16 ), - (long *)(frame->retcode +0 )); -__put_user (0x4e40 , (short *)(frame->retcode +4 )); + (uint32_t *)(frame->retcode +0 )); +__put_user (0x4e40 , (uint16_t *)(frame->retcode +4 ));if (err) goto give_sigsegv;
Signature · 1 line
1.9.1
Re: [Qemu-devel] [PATCH 29/43] linux-user: Fix broken m68k signal handling on 64 bit hosts
Hi,
I think you should use abi_long instead of uint32_t.
abi_long has an "aligned" attribute, and on m68k long are aligned on a short boundary.
#ifdef TARGET_M68K
#define ABI_INT_ALIGNMENT 2
#define ABI_LONG_ALIGNMENT 2
#define ABI_LLONG_ALIGNMENT 2
#endif
typedef uint32_t abi_ulong __attribute__((aligned(ABI_LONG_ALIGNMENT)));
Regards,
Laurent
Quoted 53 lines · Michael Roth
From: Peter Maydell <[email protected]>
The m68k signal frame setup code which writes the signal return trampoline code to the stack was assuming that a 'long' was 32 bits; on 64 bit systems this meant we would end up writing the 32 bit (2 insn) trampoline sequence to retaddr+4,retaddr+6 instead of the intended retaddr+0,retaddr+2, resulting in a guest crash when it tried to execute the invalid zero-bytes at retaddr+0. Fix by using
uint32_tinstead; also useuint16_trather than short for consistency. This fixes bug LP:1404690.Reported-by Michel BoaventuraSigned-off-by Peter Maydell <[email protected]>Signed-off-by Riku Voipio <[email protected]>(cherry picked from commit 1669add752d9f29283f8ebf6a863d7b1e2d0f146)Signed-off-by Michael Roth <[email protected]>---
linux-user/signal.c| 6 +++---1 file changed, 3 insertions(+), 3 deletions(-)linux-user/signal.c +3
-3@@ -5091,7 +5091,7 @@ static void setup_frame(int sig, struct target_sigaction *ka,/* moveq #,d0; trap #0 */ __put_user (0x70004e40 + (TARGET_NR_sigreturn <<16 ), - (long *)(frame->retcode)); + (uint32_t *)(frame->retcode));/* Set up to return from userspace */ @@ -5225,8 +5225,8 @@ static void setup_rt_frame(int sig, struct target_sigaction *ka,/* moveq #,d0; notb d0; trap #0 */ __put_user (0x70004600 + ((TARGET_NR_rt_sigreturn ^0xff ) <<16 ), - (long *)(frame->retcode +0 )); -__put_user (0x4e40 , (short *)(frame->retcode +4 )); + (uint32_t *)(frame->retcode +0 )); +__put_user (0x4e40 , (uint16_t *)(frame->retcode +4 ));if (err)goto
give_sigsegv;
--
1.9.1
Re: [Qemu-devel] [PATCH 29/43] linux-user: Fix broken m68k signal handling on 64 bit hosts
Quoted 8 lines · Laurent Vivier
Hi,
I think you should use
abi_longinstead ofuint32_t.
abi_longhas an "aligned" attribute, and on m68k long are aligned on a short boundary.
Quoted 7 lines
#ifdef
TARGET_M68K
#defineABI_INT_ALIGNMENT2
#defineABI_LONG_ALIGNMENT2
#defineABI_LLONG_ALIGNMENT2
#endiftypedef
uint32_tabi_ulong__attribute__((aligned(ABI_LONG_ALIGNMENT)));
In this instance it doesn't matter because get_sigframe() aligns the pointer we're going to write the sigframe to, and the elements in the struct before retcode[] won't make it worse than 4-aligned, I think?
-- PMM
Re: [Qemu-devel] [PATCH 29/43] linux-user: Fix broken m68k signal handling on 64 bit hosts
Quoted 19 lines · Peter Maydell
Quoted 14 lines · Laurent Vivier
Hi,
I think you should use
abi_longinstead ofuint32_t.
abi_longhas an "aligned" attribute, and on m68k long are aligned on a short boundary.#ifdef
TARGET_M68K
#defineABI_INT_ALIGNMENT2
#defineABI_LONG_ALIGNMENT2
#defineABI_LLONG_ALIGNMENT2
#endif
Quoted 6 lines
Quoted 1 line
typedef
uint32_tabi_ulong__attribute__((aligned(ABI_LONG_ALIGNMENT)));In this instance it doesn't matter because
get_sigframe()aligns the pointer we're going to write the sigframe to, and the elements in the struct before retcode[] won't make it worse than 4-aligned, I think?
Yes, I agree. But the aim of the abi_* types is to define the target ABI. Thus, for consistency it should better to use the abi_long (or abi_ulong) instead of uint32_t.
Regards,
Laurent
Re: [Qemu-devel] [PATCH 29/43] linux-user: Fix broken m68k signal handling on 64 bit hosts
Quoted 23 lines · Laurent Vivier
Quoted 23 lines
Le 25 février 2015 à 12:28, Peter Maydell <[email protected]> a écrit :
Quoted 16 lines · Laurent Vivier
Hi,
I think you should use
abi_longinstead ofuint32_t.
abi_longhas an "aligned" attribute, and on m68k long are aligned on a short
boundary.#ifdef
TARGET_M68K
#defineABI_INT_ALIGNMENT2
#defineABI_LONG_ALIGNMENT2
#defineABI_LLONG_ALIGNMENT2
#endiftypedef
uint32_tabi_ulong__attribute__((aligned(ABI_LONG_ALIGNMENT)));
Quoted 8 lines
Quoted 4 lines
In this instance it doesn't matter because
get_sigframe()aligns the pointer we're going to write the sigframe to, and the elements in the struct before retcode[] won't make it worse than 4-aligned, I think?Yes, I agree. But the aim of the abi_* types is to define the target ABI. Thus, for consistency it should better to use the
abi_long(orabi_ulong) instead ofuint32_t.
Well, strictly speaking this is writing instructions into a byte array rather than ABI longs, so it ought to do it byte-at-a-time to respect the definition of the struct. Anyway, if you want to submit a patch I don't object.
-- PMM
PATCH 30/43 scsi: fix cancellation when I/O was completed but DMA was not.
In reply to Michael RothFrom: Paolo Bonzini <[email protected]>
Commit d577646 (scsi: Introduce scsi_req_cancel_complete, 2014-09-25) was supposed to have no semantic change, but it missed a case. When r->aiocb has already been NULLed, but DMA was not complete and the SCSI layer was waiting for scsi_req_continue, after the patch the SCSI layer will not call the .cancel callback of SCSIBusInfo.
--- hw/scsi/scsi-bus.c | 2 ++ 1 file changed, 2 insertions(+)
hw/scsi/scsi-bus.c +2 -0
@@ -1770,6 +1770,8 @@ void scsi_req_cancel(SCSIRequest *req) req->io_canceled = true;if (req->aiocb) {blk_aio_cancel (req->aiocb); + }else { +scsi_req_cancel_complete (req); } }
Signature · 1 line
1.9.1
PATCH 31/43 target-i386: fix movntsd on big-endian hosts
In reply to Michael RothFrom: Paolo Bonzini <[email protected]>
This was accessing an XMM register's low half without going through XMM_Q.
--- target-i386/translate.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
target-i386/translate.c +2 -1
@@ -3069,7 +3069,8 @@ static void gen_sse(CPUX86State *env, DisasContext *s, int b, goto illegal_op;gen_lea_modrm (env, s, modrm);if (b1 &1 ) { - gen_stq_env_A0(s, offsetof(CPUX86State, xmm_regs[reg])); +gen_stq_env_A0 (s, offsetof(CPUX86State , + xmm_regs[reg].XMM_Q(0))); }else {tcg_gen_ld32u_tl (cpu_T[0 ], cpu_env, offsetof(CPUX86State , xmm_regs[reg].XMM_L(0)));
Signature · 1 line
1.9.1
PATCH 32/43 vt82c686: avoid out-of-bounds read
In reply to Michael RothFrom: Paolo Bonzini <[email protected]>
superio_ioport_readb can read the 256th element of the array. Coverity reports an out-of-bounds write in superio_ioport_writeb, but it does not show the corresponding out-of-bounds read because it cannot prove that it can happen. Fix the root cause of the problem (zhanghailang's patch instead fixes the logic in superio_ioport_writeb).
--- hw/isa/vt82c686.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
hw/isa/vt82c686.c +1 -1
@@ -36,7 +36,7 @@typedef struct SuperIOConfig { -uint8_t config[0xff ]; +uint8_t config[0x100 ];uint8_t index;uint8_t data; }SuperIOConfig ;
Signature · 1 line
1.9.1
PATCH 33/43 virtio: fix feature bit checks
In reply to Michael RothFrom: Cornelia Huck <[email protected]>
Several places check against the feature bit number instead of against the feature bit. Fix them.
--- hw/scsi/virtio-scsi.c | 2 +- hw/virtio/dataplane/vring.c | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-)
hw/scsi/virtio-scsi.c +1 -1
@@ -144,7 +144,7 @@ static int virtio_scsi_parse_req(VirtIOSCSIReq *req,* * TODO: always disable this workaround for virtio 1.0 devices. */ -if ((vdev->guest_features &VIRTIO_F_ANY_LAYOUT ) ==0 ) { +if ((vdev->guest_features & (1 <<VIRTIO_F_ANY_LAYOUT )) ==0 ) { req_size = req->elem.out_sg[0 ].iov_len; resp_size = req->elem.in_sg[0 ].iov_len; }
hw/virtio/dataplane/vring.c +2 -2
@@ -133,12 +133,12 @@ bool vring_should_notify(VirtIODevice *vdev, Vring *vring)* interrupts. */ smp_mb (); -if ((vdev->guest_features &VIRTIO_F_NOTIFY_ON_EMPTY ) && +if ((vdev->guest_features & (1 <<VIRTIO_F_NOTIFY_ON_EMPTY )) &&unlikely (vring->vr.avail->idx == vring->last_avail_idx)) {return true; } -if (!(vdev->guest_features &VIRTIO_RING_F_EVENT_IDX )) { +if (!(vdev->guest_features & (1 <<VIRTIO_RING_F_EVENT_IDX ))) {return !(vring->vr.avail->flags &VRING_AVAIL_F_NO_INTERRUPT ); } old = vring->signalled_used;
Signature · 1 line
1.9.1
PATCH 34/43 sb16: fix interrupt acknowledgement
In reply to Michael RothFrom: Paolo Bonzini <[email protected]>
SoundBlaster 16 emulation is very broken and consumes a lot of CPU, but a small fix was suggested offlist and it is enough to fix some games. I got Epic Pinball to work with the "SoundBlaster Clone" option.
The processing of the interrupt register is wrong due to two missing "not"s. This causes the interrupt flag to remain set even after the Acknowledge ports have been read (0x0e and 0x0f).
The line was introduced by commit 85571bc (audio merge (malc), 2004-11-07), but the code might have been broken before because I did not look closely at the huge patches from 10 years ago.
--- hw/audio/sb16.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
hw/audio/sb16.c +2 -2
@@ -999,7 +999,7 @@ static IO_READ_PROTO (dsp_read) retval = (!s->out_data_len || s->highspeed) ?0 :0x80 ;if (s->mixer_regs[0x82 ] &1 ) { ack =1 ; - s->mixer_regs[0x82 ] &=1 ; + s->mixer_regs[0x82 ] &= ~1 ;qemu_irq_lower (s->pic); }break ;
@@ -1008,7 +1008,7 @@ static IO_READ_PROTO (dsp_read) retval =0xff ;if (s->mixer_regs[0x82 ] &2 ) { ack =1 ; - s->mixer_regs[0x82 ] &=2 ; + s->mixer_regs[0x82 ] &= ~2 ;qemu_irq_lower (s->pic); }break ;
Signature · 1 line
1.9.1
PATCH 35/43 hw/input/hid.c Fix capslock hid code
In reply to Michael RothFrom: Dinar Valeev <[email protected]>
When ever USB keyboard is used, e.g. '-usbdevice keyboard' pressing caps lock key send 0x32 hid code, which is treated as backslash. Instead it should be 0x39 code. This affects sending uppercase keys, as they typed whith caps lock active.
While on x86 this can be workarounded by using ps/2 protocol. On Power it is crusial as we don't have anything else than USB.
This is fixes guest automation tasts over vnc.
--- hw/input/hid.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
hw/input/hid.c +1 -1
@@ -41,7 +41,7 @@ static const uint8_t hid_usage_keys[0x100] = {0x07 ,0x09 ,0x0a ,0x0b ,0x0d ,0x0e ,0x0f ,0x33 ,0x34 ,0x35 ,0xe1 ,0x31 ,0x1d ,0x1b ,0x06 ,0x19 ,0x05 ,0x11 ,0x10 ,0x36 ,0x37 ,0x38 ,0xe5 ,0x55 , -0xe2 ,0x2c ,0x32 ,0x3a ,0x3b ,0x3c ,0x3d ,0x3e , +0xe2 ,0x2c ,0x39 ,0x3a ,0x3b ,0x3c ,0x3d ,0x3e ,0x3f ,0x40 ,0x41 ,0x42 ,0x43 ,0x53 ,0x47 ,0x5f ,0x60 ,0x61 ,0x56 ,0x5c ,0x5d ,0x5e ,0x57 ,0x59 , 0x5a, 0x5b, 0x62, 0x63, 0x00, 0x00, 0x00, 0x44,
Signature · 1 line
1.9.1
PATCH 36/43 target-arm/translate-a64: Fix wrong mmu_idx usage for LDT/STT
In reply to Michael RothFrom: Peter Maydell <[email protected]>
The LDT/STT (load/store unprivileged) instruction decode was using the wrong MMU index value. This meant that instead of these insns being "always access as if user-mode regardless of current privilege" they were "always access as if kernel-mode regardless of current privilege". This went unnoticed because AArch64 Linux doesn't use these instructions.
---
I'm not counting this as a security issue because I'm assuming nobody treats TCG guests as a security boundary (certainly I would not recommend doing so...)
--- target-arm/translate-a64.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
target-arm/translate-a64.c +1 -1
@@ -2107,7 +2107,7 @@ static void disas_ldst_reg_imm9(DisasContext *s, uint32_t insn) } }else {TCGv_i64 tcg_rt =cpu_reg (s, rt); -int memidx = is_unpriv ?1 :get_mem_index (s); +int memidx = is_unpriv ?MMU_USER_IDX :get_mem_index (s);if (is_store) {do_gpr_st_memidx (s, tcg_rt, tcg_addr, size, memidx);
Signature · 1 line
1.9.1
PATCH 37/43 vfio-pci: Fix missing unparent of dynamically allocated MemoryRegion
In reply to Michael RothFrom: Alex Williamson <[email protected]>
Commit d8d95814609e added explicit object_unparent() calls for dynamically allocated MemoryRegions. The VFIOMSIXInfo structure also contains such a MemoryRegion, covering the mmap'd region of a PCI BAR above the MSI-X table. This structure is freed as part of the class exit function and therefore also needs an explicit object_unparent(). Failing to do this results in random segfaults due to fields within the structure, often the class pointer, being reclaimed and corrupted by the time object_finalize_child_property() is called for the object.
--- hw/misc/vfio.c | 1 + 1 file changed, 1 insertion(+)
hw/misc/vfio.c +1 -0
@@ -4049,6 +4049,7 @@ static void vfio_put_device(VFIODevice *vdev)DPRINTF ("vfio_put_device: close vdev->fd\n" );close (vdev->fd);if (vdev->msix) { +object_unparent (OBJECT (&vdev->msix->mmap_mem));g_free (vdev->msix); vdev->msix =NULL ; }
Signature · 1 line
1.9.1
PATCH 38/43 qemu-thread: fix qemu_event without futexes
In reply to Michael RothFrom: Paolo Bonzini <[email protected]>
This had a possible deadlock that was visible with rcutorture.
qemu_event_set qemu_event_wait
----------------------------------------------------------------
cmpxchg reads FREE, writes BUSY
futex_wait: pthread_mutex_lock
futex_wait: value == BUSY
xchg reads BUSY, writes SET
futex_wake: pthread_cond_broadcast
futex_wait: pthread_cond_wait <deadlock>
The fix is simply to avoid condvar tricks and do the obvious locking around pthread_cond_broadcast:
qemu_event_set qemu_event_wait
----------------------------------------------------------------
cmpxchg reads FREE, writes BUSY
futex_wait: pthread_mutex_lock
futex_wait: value == BUSY
xchg reads BUSY, writes SET
futex_wake: pthread_mutex_lock
(blocks)
futex_wait: pthread_cond_wait
(mutex unlocked)
futex_wake: pthread_cond_broadcast
futex_wake: pthread_mutex_unlock
futex_wait: pthread_mutex_unlock
--- util/qemu-thread-posix.c | 2 ++ 1 file changed, 2 insertions(+)
util/qemu-thread-posix.c +2 -0
@@ -306,11 +306,13 @@ static inline void futex_wait(QemuEvent *ev, unsigned val)#else static inline void futex_wake (QemuEvent *ev,int n) { +pthread_mutex_lock (&ev->lock);if (n ==1 ) {pthread_cond_signal (&ev->cond); }else {pthread_cond_broadcast (&ev->cond); } +pthread_mutex_unlock (&ev->lock); }static inline void futex_wait (QemuEvent *ev,unsigned val)
Signature · 1 line
1.9.1
PATCH 39/43 libcacard: stop linking against every single 3rd party library
In reply to Michael RothFrom: "Daniel P. Berrange" <[email protected]>
Building QEMU results in a libcacard.so that links against practically the entire world
linux-vdso.so.1 => (0x00007fff71e99000) libssl3.so => /usr/lib64/libssl3.so (0x00007f49f94b6000) libsmime3.so => /usr/lib64/libsmime3.so (0x00007f49f928e000) libnss3.so => /usr/lib64/libnss3.so (0x00007f49f8f67000) libnssutil3.so => /usr/lib64/libnssutil3.so (0x00007f49f8d3b000) libplds4.so => /usr/lib64/libplds4.so (0x00007f49f8b36000) libplc4.so => /usr/lib64/libplc4.so (0x00007f49f8931000) libnspr4.so => /usr/lib64/libnspr4.so (0x00007f49f86f2000) libdl.so.2 => /usr/lib64/libdl.so.2 (0x00007f49f84ed000) libm.so.6 => /usr/lib64/libm.so.6 (0x00007f49f81e5000) libgthread-2.0.so.0 => /usr/lib64/libgthread-2.0.so.0 (0x00007f49f7fe3000) librt.so.1 => /usr/lib64/librt.so.1 (0x00007f49f7dda000) libz.so.1 => /usr/lib64/libz.so.1 (0x00007f49f7bc4000) libcap-ng.so.0 => /usr/lib64/libcap-ng.so.0 (0x00007f49f79be000) libuuid.so.1 => /usr/lib64/libuuid.so.1 (0x00007f49f77b8000) libgnutls.so.28 => /usr/lib64/libgnutls.so.28 (0x00007f49f749a000) libSDL-1.2.so.0 => /usr/lib64/libSDL-1.2.so.0 (0x00007f49f71fd000) libpthread.so.0 => /usr/lib64/libpthread.so.0 (0x00007f49f6fe0000) libvte.so.9 => /usr/lib64/libvte.so.9 (0x00007f49f6d3f000) libXext.so.6 => /usr/lib64/libXext.so.6 (0x00007f49f6b2d000) libgtk-x11-2.0.so.0 => /usr/lib64/libgtk-x11-2.0.so.0 (0x00007f49f64a0000) libgdk-x11-2.0.so.0 => /usr/lib64/libgdk-x11-2.0.so.0 (0x00007f49f61de000) libpangocairo-1.0.so.0 => /usr/lib64/libpangocairo-1.0.so.0 (0x00007f49f5fd1000) libatk-1.0.so.0 => /usr/lib64/libatk-1.0.so.0 (0x00007f49f5daa000) libcairo.so.2 => /usr/lib64/libcairo.so.2 (0x00007f49f5a9d000) libgdk_pixbuf-2.0.so.0 => /usr/lib64/libgdk_pixbuf-2.0.so.0 (0x00007f49f5878000) libgio-2.0.so.0 => /usr/lib64/libgio-2.0.so.0 (0x00007f49f5500000) libpangoft2-1.0.so.0 => /usr/lib64/libpangoft2-1.0.so.0 (0x00007f49f52eb000) libpango-1.0.so.0 => /usr/lib64/libpango-1.0.so.0 (0x00007f49f50a0000) libgobject-2.0.so.0 => /usr/lib64/libgobject-2.0.so.0 (0x00007f49f4e4e000) libglib-2.0.so.0 => /usr/lib64/libglib-2.0.so.0 (0x00007f49f4b15000) libfontconfig.so.1 => /usr/lib64/libfontconfig.so.1 (0x00007f49f48d6000) libfreetype.so.6 => /usr/lib64/libfreetype.so.6 (0x00007f49f462b000) libX11.so.6 => /usr/lib64/libX11.so.6 (0x00007f49f42e8000) libxenstore.so.3.0 => /usr/lib64/libxenstore.so.3.0 (0x00007f49f40de000) libxenctrl.so.4.4 => /usr/lib64/libxenctrl.so.4.4 (0x00007f49f3eb6000) libxenguest.so.4.4 => /usr/lib64/libxenguest.so.4.4 (0x00007f49f3c8b000) libseccomp.so.2 => /usr/lib64/libseccomp.so.2 (0x00007f49f3a74000) librdmacm.so.1 => /usr/lib64/librdmacm.so.1 (0x00007f49f385d000) libibverbs.so.1 => /usr/lib64/libibverbs.so.1 (0x00007f49f364a000) libutil.so.1 => /usr/lib64/libutil.so.1 (0x00007f49f3447000) libc.so.6 => /usr/lib64/libc.so.6 (0x00007f49f3089000) /lib64/ld-linux-x86-64.so.2 (0x00007f49f9902000) libp11-kit.so.0 => /usr/lib64/libp11-kit.so.0 (0x00007f49f2e23000) libtspi.so.1 => /usr/lib64/libtspi.so.1 (0x00007f49f2bb2000) libtasn1.so.6 => /usr/lib64/libtasn1.so.6 (0x00007f49f299f000) libnettle.so.4 => /usr/lib64/libnettle.so.4 (0x00007f49f276d000) libhogweed.so.2 => /usr/lib64/libhogweed.so.2 (0x00007f49f2545000) libgmp.so.10 => /usr/lib64/libgmp.so.10 (0x00007f49f22cd000) libncurses.so.5 => /usr/lib64/libncurses.so.5 (0x00007f49f20a5000) libtinfo.so.5 => /usr/lib64/libtinfo.so.5 (0x00007f49f1e7a000) libgmodule-2.0.so.0 => /usr/lib64/libgmodule-2.0.so.0 (0x00007f49f1c76000) libXfixes.so.3 => /usr/lib64/libXfixes.so.3 (0x00007f49f1a6f000) libXrender.so.1 => /usr/lib64/libXrender.so.1 (0x00007f49f1865000) libXinerama.so.1 => /usr/lib64/libXinerama.so.1 (0x00007f49f1662000) libXi.so.6 => /usr/lib64/libXi.so.6 (0x00007f49f1452000) libXrandr.so.2 => /usr/lib64/libXrandr.so.2 (0x00007f49f1247000) libXcursor.so.1 => /usr/lib64/libXcursor.so.1 (0x00007f49f103c000) libXcomposite.so.1 => /usr/lib64/libXcomposite.so.1 (0x00007f49f0e39000) libXdamage.so.1 => /usr/lib64/libXdamage.so.1 (0x00007f49f0c35000) libharfbuzz.so.0 => /usr/lib64/libharfbuzz.so.0 (0x00007f49f09dd000) libpixman-1.so.0 => /usr/lib64/libpixman-1.so.0 (0x00007f49f072f000) libEGL.so.1 => /usr/lib64/libEGL.so.1 (0x00007f49f0505000) libpng16.so.16 => /usr/lib64/libpng16.so.16 (0x00007f49f02d2000) libxcb-shm.so.0 => /usr/lib64/libxcb-shm.so.0 (0x00007f49f00cd000) libxcb-render.so.0 => /usr/lib64/libxcb-render.so.0 (0x00007f49efec3000) libxcb.so.1 => /usr/lib64/libxcb.so.1 (0x00007f49efca1000) libGL.so.1 => /usr/lib64/libGL.so.1 (0x00007f49efa06000) libffi.so.6 => /usr/lib64/libffi.so.6 (0x00007f49ef7fe000) libselinux.so.1 => /usr/lib64/libselinux.so.1 (0x00007f49ef5d8000) libresolv.so.2 => /usr/lib64/libresolv.so.2 (0x00007f49ef3be000) libexpat.so.1 => /usr/lib64/libexpat.so.1 (0x00007f49ef193000) libbz2.so.1 => /usr/lib64/libbz2.so.1 (0x00007f49eef83000) libgcc_s.so.1 => /usr/lib64/libgcc_s.so.1 (0x00007f49eed6c000) liblzma.so.5 => /usr/lib64/liblzma.so.5 (0x00007f49eeb46000) libnl-route-3.so.200 => /usr/lib64/libnl-route-3.so.200 (0x00007f49ee8e2000) libnl-3.so.200 => /usr/lib64/libnl-3.so.200 (0x00007f49ee6c4000) libcrypto.so.10 => /usr/lib64/libcrypto.so.10 (0x00007f49ee2d6000) libssl.so.10 => /usr/lib64/libssl.so.10 (0x00007f49ee067000) libgraphite2.so.3 => /usr/lib64/libgraphite2.so.3 (0x00007f49ede48000) libX11-xcb.so.1 => /usr/lib64/libX11-xcb.so.1 (0x00007f49edc46000) libxcb-dri2.so.0 => /usr/lib64/libxcb-dri2.so.0 (0x00007f49eda41000) libxcb-xfixes.so.0 => /usr/lib64/libxcb-xfixes.so.0 (0x00007f49ed838000) libxcb-shape.so.0 => /usr/lib64/libxcb-shape.so.0 (0x00007f49ed634000) libgbm.so.1 => /usr/lib64/libgbm.so.1 (0x00007f49ed426000) libwayland-client.so.0 => /usr/lib64/libwayland-client.so.0 (0x00007f49ed217000) libwayland-server.so.0 => /usr/lib64/libwayland-server.so.0 (0x00007f49ed005000) libglapi.so.0 => /usr/lib64/libglapi.so.0 (0x00007f49ecddb000) libdrm.so.2 => /usr/lib64/libdrm.so.2 (0x00007f49ecbce000) libXau.so.6 => /usr/lib64/libXau.so.6 (0x00007f49ec9ca000) libxcb-glx.so.0 => /usr/lib64/libxcb-glx.so.0 (0x00007f49ec7b0000) libxcb-dri3.so.0 => /usr/lib64/libxcb-dri3.so.0 (0x00007f49ec5ad000) libxcb-present.so.0 => /usr/lib64/libxcb-present.so.0 (0x00007f49ec3aa000) libxcb-randr.so.0 => /usr/lib64/libxcb-randr.so.0 (0x00007f49ec19b000) libxcb-sync.so.1 => /usr/lib64/libxcb-sync.so.1 (0x00007f49ebf94000) libxshmfence.so.1 => /usr/lib64/libxshmfence.so.1 (0x00007f49ebd91000) libXxf86vm.so.1 => /usr/lib64/libXxf86vm.so.1 (0x00007f49ebb8a000) libpcre.so.1 => /usr/lib64/libpcre.so.1 (0x00007f49eb91d000) libgssapi_krb5.so.2 => /usr/lib64/libgssapi_krb5.so.2 (0x00007f49eb6cf000) libkrb5.so.3 => /usr/lib64/libkrb5.so.3 (0x00007f49eb3ec000) libcom_err.so.2 => /usr/lib64/libcom_err.so.2 (0x00007f49eb1e8000) libk5crypto.so.3 => /usr/lib64/libk5crypto.so.3 (0x00007f49eafb4000) libkrb5support.so.0 => /usr/lib64/libkrb5support.so.0 (0x00007f49eada5000) libkeyutils.so.1 => /usr/lib64/libkeyutils.so.1 (0x00007f49eaba0000)
All libcacard actually needs are the NSS libs. Linking against the entire world is a regression caused by
commit 9d171bd9375e4d08feff9adda15163e0811f5f42
Author: Michael Tokarev <[email protected]>
Date: Thu May 8 16:48:27 2014 +0400
libcacard: remove libcacard-specific CFLAGS and LIBS from global vars
Which removed the setting of the LIBS variable in libcacard/Makefile.
Adding it back as an empty assignment brings the linked libs back to a more reasonable set
linux-vdso.so.1 => (0x00007fff575c1000) libssl3.so => /usr/lib64/libssl3.so (0x00007f7f753b1000) libsmime3.so => /usr/lib64/libsmime3.so (0x00007f7f75189000) libnss3.so => /usr/lib64/libnss3.so (0x00007f7f74e62000) libnssutil3.so => /usr/lib64/libnssutil3.so (0x00007f7f74c36000) libplds4.so => /usr/lib64/libplds4.so (0x00007f7f74a31000) libplc4.so => /usr/lib64/libplc4.so (0x00007f7f7482c000) libnspr4.so => /usr/lib64/libnspr4.so (0x00007f7f745ed000) libpthread.so.0 => /usr/lib64/libpthread.so.0 (0x00007f7f743d0000) libdl.so.2 => /usr/lib64/libdl.so.2 (0x00007f7f741cc000) libgthread-2.0.so.0 => /usr/lib64/libgthread-2.0.so.0 (0x00007f7f73fca000) libglib-2.0.so.0 => /usr/lib64/libglib-2.0.so.0 (0x00007f7f73c90000) libc.so.6 => /usr/lib64/libc.so.6 (0x00007f7f738d3000) libz.so.1 => /usr/lib64/libz.so.1 (0x00007f7f736bd000) librt.so.1 => /usr/lib64/librt.so.1 (0x00007f7f734b4000) /lib64/ld-linux-x86-64.so.2 (0x00007f7f757fd000)
--- libcacard/Makefile | 2 ++ 1 file changed, 2 insertions(+)
libcacard/Makefile +2 -0
@@ -19,6 +19,8 @@ vscclient$(EXESUF): libcacard/vscclient.o libcacard.la libcacard.la: LDFLAGS += -rpath $(libdir) -no-undefined \ -export-symbols $(SRC_PATH)/libcacard/libcacard.syms +# Prevent libcacard.so linking against the entire world of 3rd party libs +libcacard.la: LIBS = libcacard.la: $(libcacard-lobj-y) $(call LINK,$^)
Signature · 1 line
1.9.1
PATCH 40/43 fix mc146818rtc wrong subsection name to avoid vmstate_subsection_load() fail
In reply to Michael RothFrom: Zhang Haoyu <[email protected]>
fix mc146818rtc wrong subsection name to avoid vmstate_subsection_load() fail during incoming migration or loadvm.
--- hw/timer/mc146818rtc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
hw/timer/mc146818rtc.c +1 -1
@@ -734,7 +734,7 @@ static int rtc_post_load(void *opaque, int version_id) }static const VMStateDescription vmstate_rtc_irq_reinject_on_ack_count = { - .name ="irq_reinject_on_ack_count" , + .name ="mc146818rtc/irq_reinject_on_ack_count" , .version_id =1 , .minimum_version_id =1 , .fields = (VMStateField []) {
Signature · 1 line
1.9.1
PATCH 41/43 block/iscsi: fix uninitialized variable
In reply to Michael RothFrom: Peter Wu <[email protected]>
'ret' was never initialized in the success path.
--- block/iscsi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
block/iscsi.c +1 -1
@@ -1286,7 +1286,7 @@ static int iscsi_open(BlockDriverState *bs, QDict *options, int flags,QemuOpts *opts;Error *local_err =NULL ;const char *filename; -int i, ret; +int i, ret =0 ;if ((BDRV_SECTOR_SIZE %512 ) !=0 ) { error_setg(errp, "iSCSI: Invalid BDRV_SECTOR_SIZE. "
Signature · 1 line
1.9.1
PATCH 42/43 qtest: Fix deadloop by running main loop AIO context's timers
In reply to Michael RothFrom: Fam Zheng <[email protected]>
qemu_clock_run_timers() only takes care of main_loop_tlg, we shouldn't forget aio timer list groups.
Currently, the qemu_clock_deadline_ns_all (a few lines above) counts all the timergroups of this clock type, including aio tlg, but we don't fire them, so they are never cleared, which makes a dead loop.
For example, this function hangs when trying to drive throttled block request queue with qtest clock_step.
Message-id: [email protected]
--- cpus.c | 4 ++++ 1 file changed, 4 insertions(+)
cpus.c +4 -0
@@ -372,15 +372,19 @@ static void icount_warp_rt(void *opaque)void qtest_clock_warp (int64_t dest) {int64_t clock =qemu_clock_get_ns (QEMU_CLOCK_VIRTUAL ); +AioContext *aio_context;assert (qtest_enabled ()); + aio_context =qemu_get_aio_context ();while (clock < dest) {int64_t deadline =qemu_clock_deadline_ns_all (QEMU_CLOCK_VIRTUAL );int64_t warp =qemu_soonest_timeout (dest - clock, deadline); +seqlock_write_lock (&timers_state.vm_clock_seqlock); timers_state.qemu_icount_bias += warp;seqlock_write_unlock (&timers_state.vm_clock_seqlock);qemu_clock_run_timers (QEMU_CLOCK_VIRTUAL ); +timerlist_run_timers (aio_context->tlg.tl[QEMU_CLOCK_VIRTUAL ]); clock =qemu_clock_get_ns (QEMU_CLOCK_VIRTUAL ); }qemu_clock_notify (QEMU_CLOCK_VIRTUAL );
Signature · 1 line
1.9.1
PATCH 43/43 exec: change default exception_index value for migration to -1
In reply to Michael RothFrom: Paolo Bonzini <[email protected]>
In QEMU 2.2 the exception_index value was added to the migration stream through a subsection. The default was set to 0, which is wrong and should have been -1.
However, 2.2 does not have commit e511b4d (cpu-exec: reset exception_index correctly, 2014-11-26), hence in 2.2 the exception_index is never used and is set to -1 on the next call to cpu_exec. So we can change the migration stream to make the default -1. The effects are:
- 2.2.1 -> 2.2.0: cpu->exception_index set incorrectly to 0 if it were -1 on the source; then reset to -1 in cpu_exec. This is TCG only; KVM does not use exception_index.
- 2.2.0 -> 2.2.1: cpu->exception_index set incorrectly to -1 if it were 0 on the source; but it would be reset to -1 in cpu_exec anyway. This is TCG only; KVM does not use exception_index.
- 2.2.1 -> 2.1: two bugs fixed: 1) can migrate backwards if cpu->exception_index is set to -1; 2) should not migrate backwards (but 2.2.0 allows it) if cpu->exception_index is set to 0
- 2.2.0 -> 2.3.0: 2.2.0 will send the subsection unnecessarily if exception_index is -1, but that is not a problem. 2.3.0 will set cpu->exception_index to -1 if it is 0 on the source, but this would be anyway a problem for 2.2.0 -> 2.2.x migration (due to lack of commit e511b4d in 2.2.x) so we can ignore it
- 2.2.1 -> 2.3.0: everything works.
In addition, play it safe and never send the subsection unless TCG is in use. KVM does not use exception_index (PPC KVM stores values in it for use in the subsequent call to ppc_cpu_do_interrupt, but does not need it as soon as kvm_handle_debug returns). Xen and qtest do not run any code for the CPU at all.
Message-id: [email protected]
--- exec.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
exec.c +2 -2
@@ -434,7 +434,7 @@ static int cpu_common_pre_load(void *opaque) {CPUState *cpu = opaque; - cpu->exception_index =0 ; + cpu->exception_index =-1 ;return 0 ; }
@@ -443,7 +443,7 @@ static bool cpu_common_exception_index_needed(void *opaque) {CPUState *cpu = opaque; -return cpu->exception_index !=0 ; +return tcg_enabled () && cpu->exception_index !=-1 ; }static const VMStateDescription vmstate_cpu_common_exception_index = {
Signature · 1 line
1.9.1
Quoted 6 lines
Subject: [Qemu-stable] Patch Round-up for stable 2.2.1, freeze on 2015-03-05
Hi everyone,
The following new patches are queued for QEMU stable v2.2.1:
Quoted 9 lines
https://github.com/mdroth/qemu/commits/stable-2.2-staging
The release is planned for 2015-03-10:
http://wiki.qemu.org/Planning/2.2
Please respond here or CC [email protected] on any patches you think should be included in the release.
Hi, Michael
I think the update for ipxe rom should be considered to be included in this release.
commit c246cee4eedb17ae3932d699e009a8b63240235f
Author: Gerd Hoffmann <[email protected]> Date: Wed Dec 17 13:56:50 2014 +0100
update ipxe from 69313ed to 35c5379
I had reported a bug about ipxe, and Gerd pulled the new ipxe version:
[Question] about ipxe version update
http://lists.nongnu.org/archive/html/qemu-devel/2014-12/msg00475.html
Ccing Gerd Hoffmann.
Regards,
-Gonglei
Quoted 155 lines
Testing/feedback is greatly appreciated.
Thanks!
NOTE:
I also plan on including:
6d4e4cb pc: acpi: fix WindowsXP BSOD when memory hotplug is enabled
However, the follow-up ACPI blob updates:
e909c16 acpi: update generated hex files
215d8d1 acpi-test: update expected DSDTDon't match what's generated by stable-2.2-staging, so I've backed it out for now while I figure out the best way to resolve this.
----------------------------------------------------------------
Alex Williamson (1):
vfio-pci: Fix missing unparent of dynamically allocated MemoryRegionCornelia Huck (1):
virtio: fix feature bit checksDaniel P. Berrange (1):
libcacard: stop linking against every single 3rd party libraryDavid Gibson (1):
PPC: Fix crash onspapr_tce_table_finalize()Dinar Valeev (1):
hw/input/hid.cFix capslock hid codeFam Zheng (1):
qtest: Fix deadloop by running main loop AIO context's timersGary R Hook (1):
block migration: fix return valueKevin Wolf (3):
qcow2: Fix header extension size checkqcow2.py: Add required padding for header extensions
block: Don't probe for unknown backing file formatMarcel Apfelbaum (1):
vl.c: fix regression when reading machine type from config fileMax Filippov (2):
target-xtensa: fix translation for opcodes crossing page boundary
target-xtensa: test cross-page opcodeMax Reitz (13):
block: Make essential BlockDriver objects public
block: Omitbdrv_find_formatfor essential drivers block/vvfat: qcow driver may not be found block/nfs: Addcreate_opts
block: Checkcreate_optsbefore image creation
qemu-img: Checkcreate_optsbefore image creation
qemu-img: Checkcreate_optsbefore image amendment
iotests: Only kill NBD server if it runs
iotests: Add test for unsupported image creation
qcow2: Prevent numerical overflow
qcow2: Flushing the caches inqcow2_closemay fail
qcow2: Respectbdrv_truncate()errorblock/raw-posix: Fix ret inraw_open_common()Paolo Bonzini (11):
linuxboot: fix loading old kernels
atomic: fix position of volatile qualifier
serial: resetthri_pendingon IER writes with THRI=0
serial: refineserial_thr_ipending_needed
pckbd: set bits 2-3-6-7 of the output port by default
scsi: fix cancellation when I/O was completed but DMA was not.
target-i386: fix movntsd on big-endian hosts
vt82c686: avoid out-of-bounds read
sb16: fix interrupt acknowledgement
qemu-thread: fixqemu_eventwithout futexes
exec: change defaultexception_indexvalue for migration to -1Peter Maydell (3):
audio: Don't free hw resources until after hw backend is stopped
linux-user: Fix broken m68k signal handling on 64 bit hosts target-arm/translate-a64: Fix wrongmmu_idxusage for LDT/STTPeter Wu (1):
block/iscsi: fix uninitialized variableVladimir Sementsov-Ogievskiy (1):
migration/block: fixpending()return valueZhang Haoyu (1):
fix mc146818rtc wrong subsection name to avoidvmstate_subsection_load()failaudio/audio_template.h | 2 +- block-migration.c | 14 +++++++++++--- block.c | 36 ++++++++++++++++++++---------------- block/iscsi.c | 2 +- block/nfs.c | 15 +++++++++++++++ block/qcow2-cluster.c | 2 +- block/qcow2.c | 33 ++++++++++++++++++++++----------- block/raw-posix.c | 3 ++- block/raw-win32.c | 2 +- block/raw_bsd.c | 2 +- block/vvfat.c | 6 ++++++ cpus.c | 4 ++++ exec.c | 4 ++-- hw/audio/sb16.c | 4 ++-- hw/char/serial.c | 31 +++++++++++++++++++++++++++---- hw/input/hid.c | 2 +- hw/input/pckbd.c | 10 ++++++++-- hw/isa/vt82c686.c | 2 +- hw/misc/vfio.c | 1 + hw/ppc/spapr_iommu.c | 6 +++--- hw/scsi/scsi-bus.c | 2 ++ hw/scsi/virtio-scsi.c | 2 +- hw/timer/mc146818rtc.c | 2 +- hw/virtio/dataplane/vring.c | 4 ++-- include/block/block_int.h | 8 ++++++++ include/qemu/atomic.h | 4 ++-- libcacard/Makefile | 2 ++ linux-user/signal.c | 6 +++--- pc-bios/linuxboot.bin | Bin 1024 -> 1024 bytes pc-bios/optionrom/linuxboot.S | 37 +++++++++++++++++++++++++++---------- qemu-img.c | 21 +++++++++++++++++++++ target-arm/translate-a64.c | 2 +- target-i386/translate.c | 3 ++- target-xtensa/translate.c | 27 +++++++++++++++++++++++---- tests/qemu-iotests/026.out | 120 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/071.out | 8 ++++++++ tests/qemu-iotests/080 | 2 ++ tests/qemu-iotests/080.out | 2 ++ tests/qemu-iotests/089.out | 2 ++ tests/qemu-iotests/113 | 76 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ++++++++++++ tests/qemu-iotests/113.out | 15 +++++++++++++++ tests/qemu-iotests/114 | 61 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/114.out | 13 +++++++++++++ tests/qemu-iotests/common.rc | 4 +++- tests/qemu-iotests/group | 2 ++ tests/qemu-iotests/qcow2.py | 4 ++++ tests/tcg/xtensa/test_mmu.S | 26 +++++++++++++------------- util/qemu-thread-posix.c | 2 ++ vl.c | 15 ++++++++------- 49 files changed, 556 insertions(+), 97 deletions(-) create mode 100755 tests/qemu-iotests/113 create mode 100644 tests/qemu-iotests/113.out create mode 100755 tests/qemu-iotests/114 create mode 100644 tests/qemu-iotests/114.out
Hi Michael,
Quoted 4 lines · Michael Roth
Hi everyone,
The following new patches are queued for QEMU stable v2.2.1:
Quoted 8 lines
https://github.com/mdroth/qemu/commits/stable-2.2-staging
The release is planned for 2015-03-10:
http://wiki.qemu.org/Planning/2.2
Please respond here or CC [email protected] on any patches you think should be included in the release.
Please include this commit which fixes a regression:
b40a153 target-mips: fix broken snapshotting
Thanks,
Leon