tests/tcg: target/mips: Include isa/ase and group name in test output
For better appearance and usefullnes, include ISA/ASE name and
instruction group name in the output of tests. For example, all
this data will be displayed for FMAX_A.W test:
| MSA | Float Max Min | FMAX_A.W |
| PASS: 80 | FAIL: 0 | elapsed time: 0.16 ms |
(the data will be displayed in one row; they are presented here in two
rows not to exceed the width of the commit message)
# gpg: Signature made Tue 18 Jun 2019 15:44:12 BST
# gpg: using RSA key 7F09B272C88F2FD6
# gpg: Good signature from "Kevin Wolf <[email protected]>" [full]
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74 56FE 7F09 B272 C88F 2FD6
* remotes/kevin/tags/for-upstream:
block/null: Expose read-zeroes option in QAPI schema
iotests: Test failure to loosen restrictions
block: Ignore loosening perm restrictions failures
block: Add *tighten_restrictions to *check*_perm()
block: Fix order in bdrv_replace_child()
block/commit: Drop bdrv_child_try_set_perm()
block/mirror: Fix child permissions
block: Add bdrv_child_refresh_perms()
file-posix: Update open_flags in raw_set_perm()
block: drop bs->job
blockdev: blockdev_mark_auto_del: drop usage of bs->job
block/block-backend: blk_iostatus_reset: drop usage of bs->job
block/replication: drop usage of bs->job
iotests: Hide timestamps for skipped tests
Peter Maydell [Tue, 18 Jun 2019 14:47:16 +0000 (15:47 +0100)]
Merge remote-tracking branch 'remotes/ehabkost/tags/python-next-pull-request' into staging
Python queue, 2019-06-18
Use a different method to dump avocado job log, to work around
timing-dependent issues in the arm test cases.
# gpg: Signature made Tue 18 Jun 2019 15:39:31 BST
# gpg: using RSA key 2807936F984DC5A6
# gpg: Good signature from "Eduardo Habkost <[email protected]>" [full]
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF D1AA 2807 936F 984D C5A6
* remotes/ehabkost/tags/python-next-pull-request:
Travis: print acceptance tests logs in case of job failure
Revert "travis: Make check-acceptance job more verbose"
Kevin Wolf [Mon, 17 Jun 2019 11:54:48 +0000 (13:54 +0200)]
block/null: Expose read-zeroes option in QAPI schema
Commit cd219eb1e55 added the read-zeroes option for the null-co and
null-aio block driver, but forgot to add them to the QAPI schema.
Therefore, this option wasn't available in -blockdev and blockdev-add
until now.
Add the missing option in the schema to make it available there, too.
We generally assume that loosening permission restrictions can never
fail. We have seen in the past that this assumption is wrong. This has
led to crashes because we generally pass &error_abort when loosening
permissions.
However, a failure in such a case should actually be handled in quite
the opposite way: It is very much not fatal, so qemu may report it, but
still consider the operation successful. The only realistic problem is
that qemu may then retain permissions and thus locks on images it
actually does not require. But again, that is not fatal.
To implement this behavior, we make all functions that change
permissions and that pass &error_abort to the initiating function
(bdrv_check_perm() or bdrv_child_check_perm()) evaluate the
@loosen_restrictions value introduced in the previous patch. If it is
true and an error did occur, we abort the permission update, discard the
error, and instead report success to the caller.
bdrv_child_try_set_perm() itself does not pass &error_abort, but it is
the only public function to change permissions. As such, callers may
pass &error_abort to it, expecting dropping permission restrictions to
never fail.
Max Reitz [Wed, 22 May 2019 17:03:50 +0000 (19:03 +0200)]
block: Add *tighten_restrictions to *check*_perm()
This patch makes three functions report whether the necessary permission
change tightens restrictions or not. These functions are:
- bdrv_check_perm()
- bdrv_check_update_perm()
- bdrv_child_check_perm()
Callers can use this result to decide whether a failure is fatal or not
(see the next patch).
Max Reitz [Wed, 22 May 2019 17:03:47 +0000 (19:03 +0200)]
block/mirror: Fix child permissions
We cannot use bdrv_child_try_set_perm() to give up all restrictions on
the child edge, and still have bdrv_mirror_top_child_perm() request
BLK_PERM_WRITE. Fix this by making bdrv_mirror_top_child_perm() return
0/BLK_PERM_ALL when we want to give up all permissions, and replacing
bdrv_child_try_set_perm() by bdrv_child_refresh_perms().
The bdrv_child_try_set_perm() before removing the node with
bdrv_replace_node() is then unnecessary. No permissions have changed
since the previous invocation of bdrv_child_try_set_perm().
Max Reitz [Wed, 22 May 2019 17:03:46 +0000 (19:03 +0200)]
block: Add bdrv_child_refresh_perms()
If a block node uses bdrv_child_try_set_perm() to change the permission
it takes on its child, the result may be very short-lived. If anything
makes the block layer recalculate the permissions internally, it will
invoke the node driver's .bdrv_child_perm() implementation. The
permission/shared permissions masks that returns will then override the
values previously passed to bdrv_child_try_set_perm().
If drivers want a child edge to have specific values for the
permissions/shared permissions mask, it must return them in
.bdrv_child_perm(). Consequentially, there is no need for them to pass
the same values to bdrv_child_try_set_perm() then: It is better to have
a function that invokes .bdrv_child_perm() and calls
bdrv_child_try_set_perm() with the result. This patch adds such a
function under the name of bdrv_child_refresh_perms().
Max Reitz [Wed, 22 May 2019 17:03:45 +0000 (19:03 +0200)]
file-posix: Update open_flags in raw_set_perm()
raw_check_perm() + raw_set_perm() can change the flags associated with
the current FD. If so, we have to update BDRVRawState.open_flags
accordingly. Otherwise, we may keep reopening the FD even though the
current one already has the correct flags.
Drop remaining users of bs->job:
1. assertions actually duplicated by assert(!bs->refcnt)
2. trace-point seems not enough reason to change stream_start to return
BlockJob pointer
3. Restricting creation of two jobs based on same bs is bad idea, as
3.1 Some jobs creates filters to be their main node, so, this check
don't actually prevent creating second job on same real node (which
will create another filter node) (but I hope it is restricted by
other mechanisms)
3.2 Even without bs->job we have two systems of permissions:
op-blockers and BLK_PERM
3.3 We may want to run several jobs on one node one day
blockdev: blockdev_mark_auto_del: drop usage of bs->job
We are going to remove bs->job pointer. Drop it's usage in
blockdev_mark_auto_del: instead of looking at bs->job let's check all
jobs for references to bs.
block/block-backend: blk_iostatus_reset: drop usage of bs->job
We are going to remove bs->job pointer. Drop it's usage in
blk_iostatus_reset.
blk_iostatus_reset() has only two callers:
1. blk_attach_dev(). This doesn't have anything to do with jobs and
attaching a new guest device won't solve any problem the job
encountered, so no reason to reset the iostatus for the job.
2. qmp_cont(). This resets the iostatus for everything. We can just
call block_job_iostatus_reset() for all block jobs instead of going
through BlockBackend.
Max Reitz [Thu, 13 Jun 2019 18:37:27 +0000 (20:37 +0200)]
iotests: Hide timestamps for skipped tests
Currently, the "thistime" variable is not reinitialized on every loop
iteration. This leads to tests that do not yield a run time (because
they failed or were skipped) printing the run time of the previous test
that did. Fix that by reinitializing "thistime" for every test.
Cleber Rosa [Fri, 7 Jun 2019 15:22:16 +0000 (11:22 -0400)]
Travis: print acceptance tests logs in case of job failure
Because Travis doesn't allow us to keep files produced during tests
(such as log files), let's print the complete job log to the "console"
in case of job failures.
This is a debugging aid, and given that there's been some timeouts
happening on some tests, we absolutely needs the logs to have a proper
action.
Peter Maydell [Mon, 17 Jun 2019 16:01:30 +0000 (17:01 +0100)]
target/arm: Check for dp support for dp VFM, not sp
In commit 1120827fa182f0e7622 we accidentally put the
"UNDEF unless FPU has double-precision support" check in
the single-precision VFM function. Put it in the dp
function where it belongs.
* remotes/armbru/tags/pull-monitor-2019-06-17-v2:
vl: Deprecate -mon pretty=... for HMP monitors
monitor: Replace monitor_init() with monitor_init_{hmp, qmp}()
monitor: Split Monitor.flags into separate bools
monitor: Split out monitor/monitor.c
monitor: Split out monitor/hmp.c
monitor: Split out monitor/qmp.c
monitor: Create monitor-internal.h with common definitions
monitor: Move {hmp, qmp}.c to monitor/{hmp, qmp}-cmds.c
Move monitor.c to monitor/misc.c
monitor: Rename HMP command type and tables
monitor: Remove Monitor.cmd_table indirection
monitor: Create MonitorHMP with readline state
monitor: Make MonitorQMP a child class of Monitor
monitor: Split monitor_init in HMP and QMP function
monitor: Remove unused password prompting fields
monitor: Fix return type of monitor_fdset_dup_fd_find
Kevin Wolf [Thu, 13 Jun 2019 15:34:05 +0000 (17:34 +0200)]
vl: Deprecate -mon pretty=... for HMP monitors
The -mon pretty=on|off switch of the -mon option applies only to QMP
monitors. It's silently ignored for HMP. Deprecate this combination so
that we can make it an error in future versions.
Kevin Wolf [Thu, 13 Jun 2019 15:34:04 +0000 (17:34 +0200)]
monitor: Replace monitor_init() with monitor_init_{hmp, qmp}()
Most callers know which monitor type they want to have. Instead of
calling monitor_init() with flags that can describe both types of
monitors, make monitor_init_{hmp,qmp}() public interfaces that take
specific bools instead of flags and call these functions directly.
Kevin Wolf [Thu, 13 Jun 2019 15:34:03 +0000 (17:34 +0200)]
monitor: Split Monitor.flags into separate bools
Monitor.flags contains three different flags: One to distinguish HMP
from QMP; one specific to HMP (MONITOR_USE_READLINE) that is ignored
with QMP; and another one specific to QMP (MONITOR_USE_PRETTY) that is
ignored with HMP.
Split the flags field into three bools and move them to the right
subclass. Flags are still in use for the monitor_init() interface.
Kevin Wolf [Thu, 13 Jun 2019 15:34:02 +0000 (17:34 +0200)]
monitor: Split out monitor/monitor.c
Move the monitor core infrastructure from monitor/misc.c to
monitor/monitor.c. This is code that can be shared for all targets, so
compile it only once.
What remains in monitor/misc.c after this patch is mostly monitor
command implementations (which could move to hmp-cmds.c or qmp-cmds.c
later) and code that requires a system emulator or is even
target-dependent (including HMP command completion code).
The amount of function and particularly extern variables in
monitor_int.h is probably a bit larger than it needs to be, but this way
no non-trivial code modifications are needed. The interfaces between all
monitor parts can be cleaned up later.
Kevin Wolf [Thu, 13 Jun 2019 15:34:01 +0000 (17:34 +0200)]
monitor: Split out monitor/hmp.c
Move HMP infrastructure from monitor/misc.c to monitor/hmp.c. This is
code that can be shared for all targets, so compile it only once.
The amount of function and particularly extern variables in
monitor_int.h is probably a bit larger than it needs to be, but this way
no non-trivial code modifications are needed. The interfaces between HMP
and the monitor core can be cleaned up later.
Kevin Wolf [Thu, 13 Jun 2019 15:34:00 +0000 (17:34 +0200)]
monitor: Split out monitor/qmp.c
Move QMP infrastructure from monitor/misc.c to monitor/qmp.c. This is
code that can be shared for all targets, so compile it only once.
The amount of function and particularly extern variables in
monitor_int.h is probably a bit larger than it needs to be, but this way
no non-trivial code modifications are needed. The interfaces between QMP
and the monitor core can be cleaned up later.
Kevin Wolf [Thu, 13 Jun 2019 15:33:59 +0000 (17:33 +0200)]
monitor: Create monitor-internal.h with common definitions
Before we can split monitor/misc.c, we need to create a header file that
contains the common definitions that will be used by multiple source
files.
For a start, add the type definitions for Monitor, MonitorHMP and
MonitorQMP and their dependencies. We'll add functions as needed when
splitting monitor/misc.c.
Kevin Wolf [Thu, 13 Jun 2019 15:33:58 +0000 (17:33 +0200)]
monitor: Move {hmp, qmp}.c to monitor/{hmp, qmp}-cmds.c
Now that we have a monitor/ subdirectory, let's move hmp.c and qmp.c
from the root directory there. As they contain implementations of
monitor commands, rename them to {hmp,qmp}-cmds.c, so that {hmp,qmp}.c
are free for the HMP and QMP infrastructure.
Kevin Wolf [Thu, 13 Jun 2019 15:33:57 +0000 (17:33 +0200)]
Move monitor.c to monitor/misc.c
Create a new monitor/ subdirectory and move monitor.c there. As the plan
is to move the monitor core into separate files, use the chance to
rename it to misc.c.
Kevin Wolf [Thu, 13 Jun 2019 15:33:56 +0000 (17:33 +0200)]
monitor: Rename HMP command type and tables
This renames the type for HMP monitor commands and the tables holding
the commands to make clear that they are related to HMP and to allow
making them public later:
* mon_cmd_t -> HMPCommand (fixing use of a reserved name, too)
* mon_cmds -> hmp_cmds
* info_cmds -> hmp_info_cmds
Kevin Wolf [Thu, 13 Jun 2019 15:33:55 +0000 (17:33 +0200)]
monitor: Remove Monitor.cmd_table indirection
Monitor.cmd_table is initialised to point to mon_cmds and never changed
afterwards. We can remove the indirection and just reference mon_cmds
directly instead.
Kevin Wolf [Thu, 13 Jun 2019 15:33:53 +0000 (17:33 +0200)]
monitor: Make MonitorQMP a child class of Monitor
Currently, struct Monitor mixes state that is only relevant for HMP,
state that is only relevant for QMP, and some actually shared state.
In particular, a MonitorQMP field is present in the state of any
monitor, even if it's not a QMP monitor and therefore doesn't use the
state.
As a first step towards a clean separation between QMP and HMP, let
MonitorQMP extend Monitor and create a MonitorQMP object only when the
monitor is actually a QMP monitor.
Some places accessed Monitor.qmp unconditionally, even for HMP monitors.
They can't keep doing this now, so during the conversion, they are
either changed to become conditional on monitor_is_qmp() or to assert()
that they always get a QMP monitor.
Kevin Wolf [Thu, 13 Jun 2019 15:33:52 +0000 (17:33 +0200)]
monitor: Split monitor_init in HMP and QMP function
Instead of mixing HMP and QMP monitors in the same function, separate
the monitor creation function for both.
While in theory, one could pass both MONITOR_USE_CONTROL and
MONITOR_USE_READLINE before this patch and both flags would do
something, readline support is tightly coupled with HMP: QMP never feeds
its input to readline, and the tab completion function treats the input
as an HMP command. Therefore, this configuration is useless.
After this patch, the QMP path asserts that MONITOR_USE_READLINE is not
set. The HMP path can be used with or without MONITOR_USE_READLINE, like
before.
Kevin Wolf [Thu, 13 Jun 2019 15:33:51 +0000 (17:33 +0200)]
monitor: Remove unused password prompting fields
Commit 788cf9f8c removed the code for password prompting from the
monitor. Since then, the Monitor fields password_completion_cb and
password_opaque have been unused. Remove them.
Yury Kotov [Thu, 23 May 2019 09:44:33 +0000 (12:44 +0300)]
monitor: Fix return type of monitor_fdset_dup_fd_find
monitor_fdset_dup_fd_find_remove() and monitor_fdset_dup_fd_find()
return mon_fdset->id which is int64_t. Downcasting from int64_t to int
leads to a bug with removing fd from fdset with id >= 2^32.
So, fix return types for these function.
Peter Maydell [Mon, 17 Jun 2019 15:41:25 +0000 (16:41 +0100)]
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-docs-20190617' into staging
docs infrastructure queue:
* fix some minor syntax issues in docs/specs/index.rst
* build and install the 'specs' manual, since it now has some content
* delete the "QEMU compared to other emulators" section of the docs
* Convert "translator internals" docs to RST, move to devel manual
Peter Maydell [Mon, 17 Jun 2019 14:35:31 +0000 (15:35 +0100)]
docs/specs/index.rst: Fix minor syntax issues
The docs/specs/index.rst has a couple of minor issues which
we didn't notice because we weren't building the manual:
* the ToC entry for the new PPC XIVE docs points to
a nonexistent file
* the initial comment needs to be marked by '..', not '.',
or it will appear in the output
* the title doesn't match the capitialization used by
the existing interop or devel manuals, and uses
'full-system emulation' rather than the 'system emulation'
that the interop manual title uses
Fix these minor issues before we start trying to build the manual.
Peter Maydell [Mon, 17 Jun 2019 14:35:31 +0000 (15:35 +0100)]
qemu-tech.texi: Remove "QEMU compared to other emulators" section
The "QEMU compared to other emulators" section of our documentation
hasn't been updated since 2015 (and parts of the text are even older).
We're clearly not very well placed to track the evolution of a
dozen other emulation projects, and an inaccurate or out of date
comparison doesn't serve anybody, so we're best off just removing
the whole documentation section.
If anybody cares strongly about maintaining a comparison page,
it's probably better to do that on the project's wiki where
we can update it more dynamically.
Peter Maydell [Mon, 17 Jun 2019 14:35:30 +0000 (15:35 +0100)]
Convert "translator internals" docs to RST, move to devel manual
Our user-facing manual currently has a section "translator internals"
which has some high-level information about the design of the
TCG translator. This should really be in our new devel/ manual.
Convert it to RST format and move it there.
Peter Maydell [Mon, 17 Jun 2019 14:35:21 +0000 (15:35 +0100)]
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20190617' into staging
target-arm queue:
* support large kernel images in bootloader (by avoiding
putting the initrd over the top of them)
* correctly disable FPU/DSP in the CPU for the mps2-an521, musca-a boards
* arm_gicv3: Fix decoding of ID register range
* arm_gicv3: GICD_TYPER.SecurityExtn is RAZ if GICD_CTLR.DS == 1
* some code cleanups following on from the VFP decodetree conversion
* Only implement doubles if the FPU supports them
(so we now correctly model Cortex-M4, -M33 as single precision only)
* remotes/pmaydell/tags/pull-target-arm-20190617: (24 commits)
target/arm: Only implement doubles if the FPU supports them
target/arm: Fix typos in trans function prototypes
target/arm: Remove unused cpu_F0s, cpu_F0d, cpu_F1s, cpu_F1d
target/arm: Stop using deprecated functions in NEON_2RM_VCVT_F32_F16
target/arm: stop using deprecated functions in NEON_2RM_VCVT_F16_F32
target/arm: Stop using cpu_F0s in Neon VCVT fixed-point ops
target/arm: Stop using cpu_F0s for Neon f32/s32 VCVT
target/arm: Stop using cpu_F0s for NEON_2RM_VRECPE_F and NEON_2RM_VRSQRTE_F
target/arm: Stop using cpu_F0s for NEON_2RM_VCVT[ANPM][US]
target/arm: Stop using cpu_F0s for NEON_2RM_VRINT*
target/arm: Stop using cpu_F0s for NEON_2RM_VNEG_F
target/arm: Stop using cpu_F0s for NEON_2RM_VABS_F
target/arm: Use vfp_expand_imm() for AArch32 VFP VMOV_imm
target/arm: Move vfp_expand_imm() to translate.[ch]
hw/intc/arm_gicv3: GICD_TYPER.SecurityExtn is RAZ if GICD_CTLR.DS == 1
hw/intc/arm_gicv3: Fix decoding of ID register range
hw/arm: Correctly disable FPU/DSP for some ARMSSE-based boards
hw/arm/armv7m: Forward "vfp" and "dsp" properties to CPU
target/arm: Allow M-profile CPUs to disable the DSP extension via CPU property
target/arm: Allow VFP and Neon to be disabled via a CPU property
...
Peter Maydell [Fri, 14 Jun 2019 10:44:57 +0000 (11:44 +0100)]
target/arm: Only implement doubles if the FPU supports them
The architecture permits FPUs which have only single-precision
support, not double-precision; Cortex-M4 and Cortex-M33 are
both like that. Add the necessary checks on the MVFR0 FPDP
field so that we UNDEF any double-precision instructions on
CPUs like this.
Note that even if FPDP==0 the insns like VMOV-to/from-gpreg,
VLDM/VSTM, VLDR/VSTR which take double precision registers
still exist.
Peter Maydell [Fri, 14 Jun 2019 10:44:56 +0000 (11:44 +0100)]
target/arm: Fix typos in trans function prototypes
In several places cut and paste errors meant we were using the wrong
type for the 'arg' struct in trans_ functions called by the
decodetree decoder, because we were using the _sp version of the
struct in the _dp function. These were harmless, because the two
structs were identical and so decodetree made them typedefs of the
same underlying structure (and we'd have had a compile error if they
were not harmless), but we should clean them up anyway.
Peter Maydell [Thu, 13 Jun 2019 16:39:13 +0000 (17:39 +0100)]
target/arm: Stop using cpu_F0s for Neon f32/s32 VCVT
Stop using cpu_F0s for the Neon f32/s32 VCVT operations.
Since this is the last user of cpu_F0s in the Neon 2rm-op
loop, we can remove the handling code for it too.
Peter Maydell [Thu, 13 Jun 2019 16:39:08 +0000 (17:39 +0100)]
target/arm: Stop using cpu_F0s for NEON_2RM_VABS_F
Where Neon instructions are floating point operations, we
mostly use the old VFP utility functions like gen_vfp_abs()
which work on the TCG globals cpu_F0s and cpu_F1s. The
Neon for-each-element loop conditionally loads the inputs
into either a plain old TCG temporary for most operations
or into cpu_F0s for float operations, and similarly stores
back either cpu_F0s or the temporary.
Switch NEON_2RM_VABS_F away from using cpu_F0s, and
update neon_2rm_is_float_op() accordingly.
Peter Maydell [Thu, 13 Jun 2019 16:39:07 +0000 (17:39 +0100)]
target/arm: Use vfp_expand_imm() for AArch32 VFP VMOV_imm
The AArch32 VMOV (immediate) instruction uses the same VFP encoded
immediate format we already handle in vfp_expand_imm(). Use that
function rather than hand-decoding it.
Peter Maydell [Thu, 13 Jun 2019 16:39:06 +0000 (17:39 +0100)]
target/arm: Move vfp_expand_imm() to translate.[ch]
We want to use vfp_expand_imm() in the AArch32 VFP decode;
move it from the a64-only header/source file to the
AArch32 one (which is always compiled even for AArch64).
Peter Maydell [Fri, 24 May 2019 12:42:48 +0000 (13:42 +0100)]
hw/intc/arm_gicv3: GICD_TYPER.SecurityExtn is RAZ if GICD_CTLR.DS == 1
The GICv3 specification says that the GICD_TYPER.SecurityExtn bit
is RAZ if GICD_CTLR.DS is 1. We were incorrectly making it RAZ
if the security extension is unsupported. "Security extension
unsupported" always implies GICD_CTLR.DS == 1, but the guest can
also set DS on a GIC which does support the security extension.
Fix the condition to correctly check the GICD_CTLR.DS bit.
Peter Maydell [Fri, 17 May 2019 17:40:46 +0000 (18:40 +0100)]
hw/arm: Correctly disable FPU/DSP for some ARMSSE-based boards
The SSE-200 hardware has configurable integration settings which
determine whether its two CPUs have the FPU and DSP:
* CPU0_FPU (default 0)
* CPU0_DSP (default 0)
* CPU1_FPU (default 1)
* CPU1_DSP (default 1)
Similarly, the IoTKit has settings for its single CPU:
* CPU0_FPU (default 1)
* CPU0_DSP (default 1)
Of our four boards that use either the IoTKit or the SSE-200:
* mps2-an505, mps2-an521 and musca-a use the default settings
* musca-b1 enables FPU and DSP on both CPUs
Currently QEMU models all these boards using CPUs with
both FPU and DSP enabled. This means that we are incorrect
for mps2-an521 and musca-a, which should not have FPU or DSP
on CPU0.
Create QOM properties on the ARMSSE devices corresponding to the
default h/w integration settings, and make the Musca-B1 board
enable FPU and DSP on both CPUs. This fixes the mps2-an521
and musca-a behaviour, and leaves the musca-b1 and mps2-an505
behaviour unchanged.
Peter Maydell [Fri, 17 May 2019 17:40:45 +0000 (18:40 +0100)]
hw/arm/armv7m: Forward "vfp" and "dsp" properties to CPU
Create "vfp" and "dsp" properties on the armv7m container object
which will be forwarded to its CPU object, so that SoCs can
configure whether the CPU has these features.
Peter Maydell [Fri, 17 May 2019 17:40:44 +0000 (18:40 +0100)]
target/arm: Allow M-profile CPUs to disable the DSP extension via CPU property
Allow the DSP extension to be disabled via a CPU property for
M-profile CPUs. (A and R-profile CPUs don't have this extension
as a defined separate optional architecture extension, so
they don't need the property.)
Peter Maydell [Fri, 17 May 2019 17:40:43 +0000 (18:40 +0100)]
target/arm: Allow VFP and Neon to be disabled via a CPU property
Allow VFP and neon to be disabled via a CPU property. As with
the "pmu" property, we only allow these features to be removed
from CPUs which have it by default, not added to CPUs which
don't have it.
The primary motivation here is to be able to optionally
create Cortex-M33 CPUs with no FPU, but we provide switches
for both VFP and Neon because the two interact:
* AArch64 can't have one without the other
* Some ID register fields only change if both are disabled
Peter Maydell [Thu, 16 May 2019 14:47:33 +0000 (15:47 +0100)]
hw/arm/boot: Honour image size field in AArch64 Image format kernels
Since Linux v3.17, the kernel's Image header includes a field image_size,
which gives the total size of the kernel including unpopulated data
sections such as the BSS). If this is present, then return it from
load_aarch64_image() as the true size of the kernel rather than
just using the size of the Image file itself. This allows the code
which calculates where to put the initrd to avoid putting it in
the kernel's BSS area.
This means that we should be able to reliably load kernel images
which are larger than 128MB without accidentally putting the
initrd or dtb in locations that clash with the kernel itself.
Peter Maydell [Thu, 16 May 2019 14:47:32 +0000 (15:47 +0100)]
hw/arm/boot: Avoid placing the initrd on top of the kernel
We currently put the initrd at the smaller of:
* 128MB into RAM
* halfway into the RAM
(with the dtb following it).
However for large kernels this might mean that the kernel
overlaps the initrd. For some kinds of kernel (self-decompressing
32-bit kernels, and ELF images with a BSS section at the end)
we don't know the exact size, but even there we have a
minimum size. Put the initrd at least further into RAM than
that. For image formats that can give us an exact kernel size, this
will mean that we definitely avoid overlaying kernel and initrd.
Peter Maydell [Thu, 16 May 2019 14:47:31 +0000 (15:47 +0100)]
hw/arm/boot: Diagnose layouts that put initrd or DTB off the end of RAM
We calculate the locations in memory where we want to put the
initrd and the DTB based on the size of the kernel, since they
come after it. Add some explicit checks that these aren't off the
end of RAM entirely.
(At the moment the way we calculate the initrd_start means that
it can't ever be off the end of RAM, but that will change with
the next commit.)
Peter Maydell [Thu, 16 May 2019 14:47:30 +0000 (15:47 +0100)]
hw/arm/boot: Don't assume RAM starts at address zero
In the Arm kernel/initrd loading code, in some places we make the
incorrect assumption that info->ram_size can be treated as the
address of the end of RAM, as for instance when we calculate the
available space for the initrd using "info->ram_size - info->initrd_start".
This is wrong, because many Arm boards (including "virt") specify
a non-zero info->loader_start to indicate that their RAM area
starts at a non-zero physical address.
Correct the places which make this incorrect assumption.
# gpg: Signature made Fri 14 Jun 2019 21:09:29 BST
# gpg: using RSA key D39DA71E0D496CFA
# gpg: Good signature from "Laszlo Ersek <[email protected]>" [marginal]
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg: It is not certain that the signature belongs to the owner.
# Primary key fingerprint: F5D9 660F 1BA5 F310 A95A C5E0 466A EAE0 6125 3988
# Subkey fingerprint: B3A5 5D3F 88A8 90ED 2E63 3E8D D39D A71E 0D49 6CFA
* remotes/lersek/tags/edk2-pull-2019-06-14:
pc-bios: update the README file with edk2-stable201905 information
pc-bios: refresh edk2 build artifacts for edk2-stable201905
roms/Makefile.edk2: update input file list for "pc-bios/edk2-licenses.txt"
roms/Makefile.edk2: remove edk2-stable201903 network feature test macros
roms/edk2: update submodule from edk2-stable201903 to edk2-stable201905
roms/Makefile.edk2: define edk2-stable201905 network feature test macros
Gerd Hoffmann [Fri, 7 Jun 2019 07:34:29 +0000 (09:34 +0200)]
q35: fix mmconfig and PCI0._CRS
This patch changes the handling of the mmconfig area. Thanks to the
pci(e) expander devices we already have the logic to exclude address
ranges from PCI0._CRS. We can simply add the mmconfig address range
to the list get it excluded as well.
With that in place we can go with a fixed pci hole which covers the
whole area from the end of (low) ram to the ioapic.
This will make the whole logic alot less fragile. No matter where the
firmware places the mmconfig xbar, things should work correctly. The
guest also gets a bit more PCI address space (seabios boot):
Wei Yang [Mon, 10 Jun 2019 01:18:30 +0000 (09:18 +0800)]
hw/acpi: extract acpi_add_rom_blob()
arm and i386 has almost the same function acpi_add_rom_blob(), except
giving different FWCfgCallback function.
This patch moves acpi_add_rom_blob() to utils.c by passing
FWCfgCallback to it.
Signed-off-by: Wei Yang <[email protected]> Reviewed-by: Igor Mammedov <[email protected]>
v7:
* rebase on top of current master because of conflict
v6:
* change author from Igor to Michael
v5:
* remove unnecessary header glib/gprintf.h
* rearrange include header to make it more suitable
v4:
* extract -> moves
* adjust comment in source to make checkpatch happy
v3:
* put acpi_add_rom_blob() to hw/acpi/utils.c
v2:
* remove unused header in original source file
Message-Id: <20190610011830[email protected]> Reviewed-by: Michael S. Tsirkin <[email protected]> Signed-off-by: Michael S. Tsirkin <[email protected]>
Li Hangjing [Mon, 3 Jun 2019 06:15:24 +0000 (14:15 +0800)]
vhost: fix vhost_log size overflow during migration
When a guest which doesn't support multiqueue is migrated with a multi queues
vhost-user-blk deivce, a crash will occur like:
0 qemu_memfd_alloc (name=<value optimized out>, size=562949953421312, seals=<value optimized out>, fd=0x7f87171fe8b4, errp=0x7f87171fe8a8) at util/memfd.c:153
1 0x00007f883559d7cf in vhost_log_alloc (size=70368744177664, share=true) at hw/virtio/vhost.c:186
2 0x00007f88355a0758 in vhost_log_get (listener=0x7f8838bd7940, enable=1) at qemu-2-12/hw/virtio/vhost.c:211
3 vhost_dev_log_resize (listener=0x7f8838bd7940, enable=1) at hw/virtio/vhost.c:263
4 vhost_migration_log (listener=0x7f8838bd7940, enable=1) at hw/virtio/vhost.c:787
5 0x00007f88355463d6 in memory_global_dirty_log_start () at memory.c:2503
6 0x00007f8835550577 in ram_init_bitmaps (f=0x7f88384ce600, opaque=0x7f8836024098) at migration/ram.c:2173
7 ram_init_all (f=0x7f88384ce600, opaque=0x7f8836024098) at migration/ram.c:2192
8 ram_save_setup (f=0x7f88384ce600, opaque=0x7f8836024098) at migration/ram.c:2219
9 0x00007f88357a419d in qemu_savevm_state_setup (f=0x7f88384ce600) at migration/savevm.c:1002
10 0x00007f883579fc3e in migration_thread (opaque=0x7f8837530400) at migration/migration.c:2382
11 0x00007f8832447893 in start_thread () from /lib64/libpthread.so.0
12 0x00007f8832178bfd in clone () from /lib64/libc.so.6
This is because vhost_get_log_size() returns a overflowed vhost-log size.
In this function, it uses the uninitialized variable vqs->used_phys and
vqs->used_size to get the vhost-log size.
This check shouldn't be necessary, since &error_fatal is given as
argument and will exit() on failure. However, this change should
silence coverity CID 1401761 & 1401705.
Laszlo Ersek [Thu, 6 Jun 2019 10:49:02 +0000 (12:49 +0200)]
pc-bios: update the README file with edk2-stable201905 information
Refresh the "pc-bios/README" file with edk2, OpenSSL, and Berkeley
SoftFloat release info, matching the edk2-stable201905 firmware images
added in the previous patch.
Edk2 commit f53de8aecb12 ("edk2: Add License-History.txt", 2019-04-09)
added a new file called "License-History.txt", describing license and
contributor agreement changes over time.
(2) For advacing edk2's OpenSSL git submodule to the 1.1.1b release, a
more up-to-date ArmSoftFloatLib instance became necessary, on 32-bit
ARM. Therefore, the embedded (and by now, obsolete) Berkeley SoftFloat
library was replaced with a git submodule of the same, checked out at
the latest release ("3e", plus a typo fix on top -- commit b64af41c3276). This was implemented for TianoCore Bugzilla
Include the edk2 project's "License-History.txt" file, and the Berkeley
SoftFloat library's "COPYING.txt", in our generated
"pc-bios/edk2-licenses.txt" text file.
Laszlo Ersek [Thu, 6 Jun 2019 11:50:56 +0000 (13:50 +0200)]
roms/Makefile.edk2: remove edk2-stable201903 network feature test macros
At tag "edk2-stable201905", the build flags HTTP_BOOT_ENABLE and
TLS_ENABLE have become useless. Their roles have been taken by
NETWORK_HTTP_BOOT_ENABLE and NETWORK_TLS_ENABLE, respectively. Remove the
old build flags.
Laszlo Ersek [Thu, 6 Jun 2019 09:59:09 +0000 (11:59 +0200)]
roms/Makefile.edk2: define edk2-stable201905 network feature test macros
Due to edk2 commits 631195044ff0 ("OvmfPkg: Update DSC/FDF to use
NetworkPkg's include fragment file.", 2019-05-16) and 157a3b1aa50f
("ArmVirtPkg: Update DSC/FDF to use NetworkPkg's include fragment file.",
2019-05-16), we'll have to pefix the HTTP_BOOT_ENABLE and TLS_ENABLE build
flags with "NETWORK_", for building edk2 at tag "edk2-stable201905".
Right now the new flags are no-ops, so we can insert them separately,
before advancing our edk2 submodule reference. We will remove the old
flags afterwards.
Peter Maydell [Fri, 14 Jun 2019 13:46:13 +0000 (14:46 +0100)]
Merge remote-tracking branch 'remotes/maxreitz/tags/pull-block-2019-06-14' into staging
Block patches:
- Allow blockdev-backup from nodes that are not in qemu's main AIO
context to newly added nodes
- Add salvaging mode to qemu-img convert
- Minor fixes to tests, documentation, and for less Valgrind annoyance
* remotes/maxreitz/tags/pull-block-2019-06-14:
iotests: Test qemu-img convert -C --salvage
iotests: Test qemu-img convert --salvage
blkdebug: Inject errors on .bdrv_co_block_status()
blkdebug: Add "none" event
blkdebug: Add @iotype error option
qemu-img: Add salvaging mode to convert
qemu-img: Move quiet into ImgConvertState
blockdev: Overlays are not snapshots
qapi/block-core: Overlays are not snapshots
qemu-img: Fix options leakage in img_rebase()
iotests: restrict 254 to support only qcow2
hw/block/fdc: floppy command FIFO memory initialization
iotests: Fix intermittent failure in 219
iotests: Filter 175's allocation information
event_match: always match on None value
iotests: add iotest 256 for testing blockdev-backup across iothread contexts
iotests.py: rewrite run_job to be pickier
QEMUMachine: add events_wait method
iotests.py: do not use infinite waits
blockdev-backup: don't check aio_context too early
Max Reitz [Tue, 7 May 2019 20:35:07 +0000 (22:35 +0200)]
iotests: Test qemu-img convert --salvage
This test converts a simple image to another, but blkdebug injects
block_status and read faults at some offsets. The resulting image
should be the same as the input image, except that sectors that could
not be read have to be 0.
Max Reitz [Tue, 7 May 2019 20:35:05 +0000 (22:35 +0200)]
blkdebug: Add "none" event
Together with @iotypes and @sector, this can be used to trap e.g. the
first read or write access to a certain sector without having to know
what happens internally in the block layer, i.e. which "real" events
happen right before such an access.
Max Reitz [Tue, 7 May 2019 20:35:04 +0000 (22:35 +0200)]
blkdebug: Add @iotype error option
This new error option allows users of blkdebug to inject errors only on
certain kinds of I/O operations. Users usually want to make a very
specific operation fail, not just any; but right now they simply hope
that the event that triggers the error injection is followed up with
that very operation. That may not be true, however, because the block
layer is changing (including blkdebug, which may increase the number of
types of I/O operations on which to inject errors).
The new option's default has been chosen to keep backwards
compatibility.
Note that similar to the internal representation, we could choose to
expose this option as a list of I/O types. But there is no practical
use for this, because as described above, users usually know exactly
which kind of operation they want to make fail, so there is no need to
specify multiple I/O types at once. In addition, exposing this option
as a list would require non-trivial changes to qemu_opts_absorb_qdict().
Max Reitz [Tue, 7 May 2019 20:35:03 +0000 (22:35 +0200)]
qemu-img: Add salvaging mode to convert
This adds a salvaging mode (--salvage) to qemu-img convert which ignores
read errors and treats the respective areas as containing only zeroes.
This can be used for instance to at least partially recover the data
from terminally corrupted qcow2 images.
Max Reitz [Tue, 7 May 2019 20:35:02 +0000 (22:35 +0200)]
qemu-img: Move quiet into ImgConvertState
Move img_convert()'s quiet flag into the ImgConvertState so it is
accessible by nested functions. -q dictates that it suppresses anything
but errors, so if those functions want to emit warnings, they need to
query this flag first. (There currently are no such warnings, but there
will be as of the next patch.)
Max Reitz [Mon, 3 Jun 2019 20:22:35 +0000 (22:22 +0200)]
qapi/block-core: Overlays are not snapshots
A snapshot is something that reflects the state of something at a
certain point in time. It does not change.
The file our snapshot commands create (or the node they install) is not
a snapshot, as it does change over time. It is an overlay. We cannot
do anything about the parameter names, but we can at least adjust the
descriptions to reflect that fact.
projects / qemu.git / log