Amos Kong [Fri, 7 Sep 2012 03:11:03 +0000 (11:11 +0800)]
add a boot parameter to set reboot timeout
Added an option to let qemu transfer a configuration file to bios,
"etc/boot-fail-wait", which could be specified by command
-boot reboot-timeout=T
T have a max value of 0xffff, unit is ms.
With this option, guest will wait for a given time if not find
bootabled device, then reboot. If reboot-timeout is '-1', guest
will not reboot, qemu passes '-1' to bios by default.
This feature need the new seabios's support.
Seabios pulls the value from the fwcfg "file" interface, this
interface is used because SeaBIOS needs a reliable way of
obtaining a name, value size, and value. It in no way requires
that there be a real file on the user's host machine.
getaddrinfo can give us a list of addresses, but we only try to
connect to the first one. If that fails we never proceed to
the next one. This is common on desktop setups that often have ipv6
configured but not actually working.
To fix this make inet_connect_nonblocking retry connection with a different
address.
callers on inet_nonblocking_connect register a callback function that will
be called when connect opertion completes, in case of failure the fd will have
a negative value
Stefan Weil [Fri, 14 Sep 2012 17:02:30 +0000 (19:02 +0200)]
configure: Allow builds without any system or user emulation
The old code aborted configure when no emulation target was selected.
Even after removing the 'exit 1', it tried to read from STDIN
when QEMU was configured with
This patch adds a "use64" property which will make the ivshmem driver
register a 64bit memory bar when set, so you have something to play with
when testing 64bit pci bits. It also allows to have quite big shared
memory regions, like this:
[root@fedora ~]# lspci -vs1:1
01:01.0 RAM memory: Red Hat, Inc Device 1110
Subsystem: Red Hat, Inc Device 1100
Physical Slot: 1-1
Flags: fast devsel
Memory at fd400000 (32-bit, non-prefetchable) [disabled] [size=256]
Memory at 8040000000 (64-bit, prefetchable) [size=1G]
[ v5: rebase, update compat property for post-1.2 merge ]
[ v4: rebase & adapt to latest master again ]
[ v3: rebase & adapt to latest master ]
[ v2: default to on as suggested by avi,
turn off for pc-$old using compat property ]
Igor Mammedov [Wed, 5 Sep 2012 21:06:21 +0000 (23:06 +0200)]
Introduce powerdown_notifiers
Notifier will be used for signaling powerdown request to guest in
a more general way and intended to replace very specific
qemu_irq_rise(qemu_system_powerdown) and will allow to remove global
variable qemu_system_powerdown.
tcg: Streamline movcond_i64 using 32-bit arithmetic
Avoiding 64-bit arithmetic (outside of the compare) reduces the
generated op count from 15 to 12, and the generated code size on
i686 from 105 to 88 bytes.
Checking that we don't try for idx != [01] is trivial. Checking
that we don't issue more than one of any index requires a tad
more data and some ifdefs protecting that new variable.
For tcg_gen_concat_i32_i64 we only use deposit if the host supports it.
For tcg_gen_concat32_i64 even if the host does not, as we get identical
code before and after.
Note that this relies on the ANDI -> EXTU patch for the identity claim.
Anthony Liguori [Tue, 25 Sep 2012 21:06:16 +0000 (16:06 -0500)]
Merge remote-tracking branch 'kwolf/for-anthony' into staging
* kwolf/for-anthony:
block: remove keep_read_only flag from BlockDriverState struct
block: convert bdrv_commit() to use bdrv_reopen()
block: vpc image file reopen
block: vdi image file reopen
block: vmdk image file reopen
block: qcow image file reopen
block: qcow2 image file reopen
block: qed image file reopen
block: raw image file reopen
block: raw-posix image file reopen
block: purge s->aligned_buf and s->aligned_buf_size from raw-posix.c
block: use BDRV_O_NOCACHE instead of s->aligned_buf in raw-posix.c
block: do not parse BDRV_O_CACHE_WB in block drivers
block: move open flag parsing in raw block drivers to helper functions
block: move aio initialization into a helper function
block: Framework for reopening files safely
block: make bdrv_set_enable_write_cache() modify open_flags
block: correctly set the keep_read_only flag
blockdev: preserve readonly and snapshot states across media changes
Anthony Liguori [Tue, 25 Sep 2012 21:06:16 +0000 (16:06 -0500)]
Merge remote-tracking branch 'stefanha/trivial-patches' into staging
* stefanha/trivial-patches:
w32: Always use standard instead of native format strings
net/socket: Fix compiler warning (regression for MinGW)
linux-user: Remove redundant null check and replace free by g_free
qemu-timer: simplify qemu_run_timers
TextConsole: saturate escape parameter in TTY_STATE_CSI
curses: don't initialize curses when qemu is daemonized
dtrace backend: add function to reserved words
pflash_cfi01: Fix warning caused by unreachable code
ioh3420: Remove unreachable code
lm4549: Fix buffer overflow
cadence_uart: Fix buffer overflow
qemu-sockets: Fix potential memory leak
qemu-ga: Remove unreachable code after g_error
target-i386: Allow tsc-frequency to be larger then 2.147G
Anthony Liguori [Tue, 25 Sep 2012 21:06:16 +0000 (16:06 -0500)]
Merge remote-tracking branch 'bonzini/scsi-next' into staging
* bonzini/scsi-next:
SCSI: Standard INQUIRY data should report HiSup flag as set.
scsi-disk: use scsi_data_cdb_length
scsi: introduce scsi_cdb_length and scsi_data_cdb_length
scsi-disk: fix check for out-of-range LBA
scsi-disk: introduce check_lba_range
iSCSI: We dont need to explicitely call qemu_notify_event() any more
iSCSI: We need to support SG_IO also from iscsi_ioctl()
Anthony Liguori [Tue, 25 Sep 2012 21:06:15 +0000 (16:06 -0500)]
Merge remote-tracking branch 'bonzini/nbd-next' into staging
* bonzini/nbd-next:
nbd: add nbd_export_get_blockdev
nbd: negotiate with named exports
nbd: register named exports
qemu-nbd: rewrite termination conditions to use a state machine
nbd: add notification for closing an NBDExport
nbd: track clients into NBDExport
nbd: add reference counting to NBDExport
nbd: do not leak nbd_trip coroutines when a connection is torn down
nbd: make refcount interface public
nbd: do not close BlockDriverState in nbd_export_close
nbd: pass NBDClient to nbd_send_negotiate
nbd: add more constants
Jeff Cody [Thu, 20 Sep 2012 19:13:30 +0000 (15:13 -0400)]
block: vmdk image file reopen
This patch supports reopen for VMDK image files. VMDK extents are added
to the existing reopen queue, so that the transactional model of reopen
is maintained with multiple image files.
Jeff Cody [Thu, 20 Sep 2012 19:13:25 +0000 (15:13 -0400)]
block: raw-posix image file reopen
This is derived from the Supriya Kannery's reopen patches.
This contains the raw-posix driver changes for the bdrv_reopen_*
functions. All changes are staged into a temporary scratch buffer
during the prepare() stage, and copied over to the live structure
during commit(). Upon abort(), all changes are abandoned, and the
live structures are unmodified.
The _prepare() will create an extra fd - either by means of a dup,
if possible, or opening a new fd if not (for instance, access
control changes). Upon _commit(), the original fd is closed and
the new fd is used. Upon _abort(), the duplicate/new fd is closed.
Jeff Cody [Thu, 20 Sep 2012 19:13:23 +0000 (15:13 -0400)]
block: use BDRV_O_NOCACHE instead of s->aligned_buf in raw-posix.c
Rather than check for a non-NULL aligned_buf to determine if
raw_aio_submit needs to check for alignment, check for the presence
of BDRV_O_NOCACHE in the bs->open_flags.
Jeff Cody [Thu, 20 Sep 2012 19:13:19 +0000 (15:13 -0400)]
block: Framework for reopening files safely
This is based on Supriya Kannery's bdrv_reopen() patch series.
This provides a transactional method to reopen multiple
images files safely.
Image files are queue for reopen via bdrv_reopen_queue(), and the
reopen occurs when bdrv_reopen_multiple() is called. Changes are
staged in bdrv_reopen_prepare() and in the equivalent driver level
functions. If any of the staged images fails a prepare, then all
of the images left untouched, and the staged changes for each image
abandoned.
Block drivers are passed a reopen state structure, that contains:
* BDS to reopen
* flags for the reopen
* opaque pointer for any driver-specific data that needs to be
persistent from _prepare to _commit/_abort
* reopen queue pointer, if the driver needs to queue additional
BDS for a reopen
Jeff Cody [Thu, 20 Sep 2012 19:13:18 +0000 (15:13 -0400)]
block: make bdrv_set_enable_write_cache() modify open_flags
bdrv_set_enable_write_cache() sets the bs->enable_write_cache flag,
but without the flag recorded in bs->open_flags, then next time
a reopen() is performed the enable_write_cache setting may be
inadvertently lost.
This will set the flag in open_flags, so it is preserved across
reopens.
Jeff Cody [Thu, 20 Sep 2012 19:13:17 +0000 (15:13 -0400)]
block: correctly set the keep_read_only flag
I believe the bs->keep_read_only flag is supposed to reflect
the initial open state of the device. If the device is initially
opened R/O, then commit operations, or reopen operations changing
to R/W, are prohibited.
Currently, the keep_read_only flag is only accurate for the active
layer, and its backing file. Subsequent images end up always having
the keep_read_only flag set.
For instance, what happens now:
[ base ] kro = 1, ro = 1
|
v
[ snap-1 ] kro = 1, ro = 1
|
v
[ snap-2 ] kro = 0, ro = 1
|
v
[ active ] kro = 0, ro = 0
What we want:
[ base ] kro = 0, ro = 1
|
v
[ snap-1 ] kro = 0, ro = 1
|
v
[ snap-2 ] kro = 0, ro = 1
|
v
[ active ] kro = 0, ro = 0
Kevin Shanahan [Thu, 20 Sep 2012 23:20:22 +0000 (08:50 +0930)]
blockdev: preserve readonly and snapshot states across media changes
If readonly=on is given at device creation time, the ->readonly flag
needs to be set in the block driver state for this device so that
readonly-ness is preserved across media changes (qmp change command).
Similarly, to preserve the snapshot property requires ->open_flags to
be correct.
Stefan Weil [Sat, 22 Sep 2012 20:26:19 +0000 (22:26 +0200)]
w32: Add implementation of gmtime_r, localtime_r
Those functions are missing in MinGW.
Some versions of MinGW-w64 include defines for gmtime_r and localtime_r.
Older versions of these macros are buggy (they return a pointer to a
static variable), therefore we don't want them. Newer versions are
similar to the code used here, but without the memset.
The implementation which is used here is not strictly reentrant,
but sufficiently good for QEMU on w32 or w64.
Stefan Weil [Wed, 22 Aug 2012 19:42:32 +0000 (21:42 +0200)]
w32: Always use standard instead of native format strings
GLib 2.0 include files use __printf__ for the format attribute
which resolves to native format strings on w32 hosts.
QEMU wants standard format strings instead of native format
strings, so we simply change any declaration with __printf__
to use __gnu_printf__.
This works because all basic printf functions support both
kinds of format strings.
This fixes a compiler warning:
qapi/string-output-visitor.c: In function ‘print_type_int’:
qapi/string-output-visitor.c:34:5: warning: unknown conversion type character ‘l’ in format [-Wformat]
qapi/string-output-visitor.c:34:5: warning: too many arguments for format [-Wformat-extra-args]
net/socket.c:136: warning:
pointer targets in passing argument 2 of ‘sendto’ differ in signedness
/usr/lib/gcc/amd64-mingw32msvc/4.4.4/../../../../amd64-mingw32msvc/include/winsock2.h:1313: note:
expected ‘const char *’ but argument is of type ‘const uint8_t *’
Add a 'qemu_sendto' macro which provides that type cast where needed
and use the new macro instead of 'sendto'.
curses: don't initialize curses when qemu is daemonized
Current qemu initializes curses even if -daemonize option is
passed. This cause problem because shell prompt appears without
calling endwin().
This patch adds new function, is_daemonized(), to OS dependent
code. With this function, curses_display_init() can check that qemu is
daemonized or not. If daemonized, curses_display_init() isn't called
and the problem is avoided.
Of course, -daemonize && -curses doesn't make sense. Users shouldn't
pass the arguments at the same time. But the problem is very painful
because Ctrl-C cannot be delivered to the terminal.
Stefan Weil [Sat, 1 Sep 2012 11:00:48 +0000 (13:00 +0200)]
pflash_cfi01: Fix warning caused by unreachable code
Report from smatch:
hw/pflash_cfi01.c:431 pflash_write(180) info: ignoring unreachable code.
Instead of removing the return statement after the switch statement,
the patch replaces the return statements in the switch statement by
break statements. Other switch statements in the same code do it also
like that.
There must be enough space to add two entries starting with index
s->buffer_level, therefore the old check was wrong.
[Peter Maydell <[email protected]> clarifies the nature of the
analyser warning:
I don't object to making the change to placate the analyser,
but I don't think this is actually a buffer overrun. We always
add and remove samples from the buffer two at a time, so it's
not possible to get here with s->buffer_level == BUFFER_SIZE-1
(which is the only case where the old and new conditions
give different answers).]
Blue Swirl [Sat, 22 Sep 2012 17:59:15 +0000 (17:59 +0000)]
Merge branch 'tcg-sparc' of git://repo.or.cz/qemu/rth
* 'tcg-sparc' of git://repo.or.cz/qemu/rth:
tcg-sparc: Preserve branch destinations during retranslation
tcg-sparc: Fix and enable direct TB chaining.
tcg-sparc: Add %g/%o registers to alloc_order
tcg-sparc: Use defines for temporaries.
tcg-sparc: Mask shift immediates to avoid illegal insns.
tcg-sparc: Clean up cruft stemming from attempts to use global registers.
tcg-sparc: Change AREG0 in generated code to %i0.
tcg-sparc: Support GUEST_BASE.
tcg-sparc: Fix qemu_ld/st to handle 32-bit host.
tcg-sparc: Assume v9 cpu always, i.e. force v8plus in 32-bit mode.
tcg-sparc: Don't MAP_FIXED on top of the program
tcg-sparc: Fix ADDX opcode.
tcg-sparc: Hack in qemu_ld/st64 for 32-bit.
linux-user: Use memcpy in get_user/put_user.
In case Coprocessor Context option is enabled CPENABLE SR bits control
whether access to coprocessors is allowed or would rise one of
CoprocessorXDisabled exceptions.
Max Filippov [Wed, 19 Sep 2012 00:23:51 +0000 (04:23 +0400)]
softfloat: add NO_SIGNALING_NANS
Architectures that don't have signaling NaNs can define
NO_SIGNALING_NANS, it will make float*_is_quiet_nan return 1 for any NaN
and float*_is_signaling_nan always return 0.
Max Filippov [Wed, 19 Sep 2012 00:23:50 +0000 (04:23 +0400)]
softfloat: make float_muladd_negate_* flags independent
Flags passed into float{32,64}_muladd are treated as bits; assign
independent bits to float_muladd_negate_* to allow precise control over
what gets negated in float{32,64}_muladd.
Stefan Weil [Thu, 13 Sep 2012 17:37:44 +0000 (19:37 +0200)]
tcg/i386: Add shortcuts for registers used in L constraint
While 64 bit hosts use the first three registers which are also used
as function input parameters, 32 bit hosts use TCG_REG_EAX and
TCG_REG_EDX which are not used in parameter passing.
After defining new register macros for the registers used in L
constraint, the patch replaces most occurrences of
tcg_target_call_iarg_regs[0], tcg_target_call_iarg_regs[1] and
tcg_target_call_iarg_regs[2] by those new macros.
tcg_target_call_iarg_regs remains unchanged when it is used for input
arguments (only with 64 bit hosts) before tcg_out_calli.
A comment related to those registers was fixed, too.
Signed-off-by: Stefan Weil <[email protected]>
[aurel32: build fix on i386, small optimization for i386 in the prologue] Signed-off-by: Aurelien Jarno <[email protected]>
Stefan Weil [Thu, 13 Sep 2012 17:37:43 +0000 (19:37 +0200)]
w64: Fix TCG helper functions with 5 arguments
TCG uses 6 registers for function arguments on 64 bit Linux hosts,
but only 4 registers on W64 hosts.
Commit 2999a0b20074a7e4a58f56572bb1436749368f59 increased the number
of arguments for some important helper functions from 4 to 5
which triggered a bug for W64 hosts: QEMU aborts when executing
helper_lcall_real in the guest's BIOS because function
tcg_target_get_call_iarg_regs_count always returned 6.
As W64 has only 4 registers for arguments, the 5th argument must be
passed on the stack using a correct stack offset.
Commit 25c4d9cc changed all TCGOpcode enums to be available, so we don't
need to #ifdef #endif the one that are available only on some targets.
This makes the code easier to read.
tcg/optimize: prefer the "op a, a, b" form for commutative ops
The "op a, a, b" form is better handled on non-RISC host than the "op
a, b, a" form, so swap the arguments to this form when possible, and
when b is not a constant.
This reduces the number of generated instructions by a tiny bit.
tcg/optimize: further optimize brcond/movcond/setcond
When both argument of brcond/movcond/setcond are the same or when one
of the two values is a constant equal to zero, it's possible to do
further optimizations.
Now that it's possible to detect copies, we can optimize the case
the "op r, a, a => movi r, 0". This helps in the computation of
overflow flags when one of the two args is 0.
tcg/optimize: do copy propagation for all operations
It is possible to due copy propagation for all operations, even the one
that have side effects or clobber arguments (it only concerns input
arguments). That said, the call operation should be handled differently
due to the variable number of arguments.
The copy propagation pass tries to keep track what is a copy of what
and what has copy of what, and in addition it keep a circular list of
of all the copies. Unfortunately this doesn't fully work: a mov from
a temp which has a state "COPY" changed it into a state "HAS_COPY".
Later when this temp is used again, it is considered has not having
copy and thus no propagation is done.
This patch fixes that by removing the hiearchy between copies, and thus
only keeping a "COPY" state both meaning "is a copy" and "has a copy".
The decision of which copy to use is deferred to the actual temp
replacement. At this stage there is not one best choice to do, but only
better choices than others. For doing the best choice the operation
would have to be parsed in reversed to know if a temp is going to be
used later or not. That what is done by the liveness analysis. At this
stage it is known that globals will be always live, that local temps
will be dead at the end of the translation block, and that the temps
will be dead at the end of the basic block. This means that this stage
should try to replace temps by local temps or globals and local temps
by globals.
The copy propagation doesn't check the types of the temps during copy
propagation. However TCG is using the mov_i32 for the i64 to i32
conversion and thus the two are not equivalent.
With this patch tcg_opt_gen_mov() doesn't consider two temps of
different type as copies anymore.
So far it seems the optimization was not aggressive enough to trigger
this bug, but it will be triggered later in this series once the copy
propagation is improved.
projects / qemu.git / log