Stefan Hajnoczi [Mon, 24 Jan 2011 15:35:00 +0000 (15:35 +0000)]
usb-msd: Propagate removable bit to SCSI device
USB Mass Storage Devices sometimes have the RMB (removable) bit set in
the SCSI INQUIRY response. Thumbdrives tend to have the bit set whereas
hard disks do not.
Operating systems differentiate between removable devices and fixed
devices. Under Linux, the anaconda installer looks for removable
devices. Under Windows, only fixed devices may have more than one
partition and AutoRun is also affected by the removable bit.
For these reasons, allow USB Mass Storage Devices to override the
removable bit:
Stefan Hajnoczi [Mon, 24 Jan 2011 15:34:58 +0000 (15:34 +0000)]
scsi-disk: Allow overriding SCSI INQUIRY removable bit
Provide the "removable" qdev property bit to override the SCSI INQUIRY
removable (RMB) bit for non-CDROM devices. This will be used by USB
Mass Storage Devices, which sometimes have this guest-visible bit set
and sometimes do not. They therefore requires a means for user
configuration.
Stefan Hajnoczi [Mon, 24 Jan 2011 09:32:20 +0000 (09:32 +0000)]
block: Use backing format driver during image creation
The backing format should be honored during image creation. For some
reason we currently use the image format to open the backing file. This
fails when the backing file has a different format than the image being
created. Keep the image and backing format drivers completely separate.
Also print the backing filename if there is an error opening the backing
file instead of the image filename.
Pierre Riteau [Fri, 21 Jan 2011 11:42:30 +0000 (12:42 +0100)]
Fix block migration when the device size is not a multiple of 1 MB
b02bea3a85cc939f09aa674a3f1e4f36d418c007 added a check on the return
value of bdrv_write and aborts migration when it fails. However, if the
size of the block device to migrate is not a multiple of BLOCK_SIZE
(currently 1 MB), the last bdrv_write will fail with -EIO.
Fixed by calling bdrv_write with the correct size of the last block.
Kevin Wolf [Fri, 14 Jan 2011 14:55:38 +0000 (15:55 +0100)]
qcow2: Batch flushes for COW
qcow2 calls bdrv_flush() after performing COW in order to ensure that the
L2 table change is never written before the copy is safe on disk. Now that the
L2 table is cached, we can wait with flushing until we write out the next L2
table.
Kevin Wolf [Mon, 10 Jan 2011 16:15:10 +0000 (17:15 +0100)]
qcow2: Add QcowCache
This adds some new cache functions to qcow2 which can be used for caching
refcount blocks and L2 tables. When used with cache=writethrough they work
like the old caching code which is spread all over qcow2, so for this case we
have merely a cleanup.
The interesting case is with writeback caching (this includes cache=none) where
data isn't written to disk immediately but only kept in cache initially. This
leads to some form of metadata write batching which avoids the current "write
to refcount block, flush, write to L2 table" pattern for each single request
when a lot of cluster allocations happen. Instead, cache entries are only
written out if its required to maintain the right order. In the pure cluster
allocation case this means that all metadata updates for requests are done in
memory initially and on sync, first the refcount blocks are written to disk,
then fsync, then L2 tables.
This improves performance of scenarios with lots of cluster allocations
noticably (e.g. installation or after taking a snapshot).
Merge ide_dma_submit_check into it's only caller. Also use tail recursion
using a goto instead of a real recursion - this avoid overflowing the
stack in the pathological situation of an recurring error that is ignored.
We'll still be busy looping in ide_dma_cb, but at least won't eat up
all stack space after this.
Currenly the code only resets the io_buffer_index field for reads,
but the code seems to expect this for all types of I/O. I guess
we simply don't hit large enough transfers that would require this
often enough.
Factor the DMA I/O path that is duplicated between read and write
commands, into common helpers using the s->is_read flag added for
the macio ATA controller.
Pierre Riteau [Wed, 12 Jan 2011 13:41:00 +0000 (14:41 +0100)]
Avoid divide by zero when there is no block device to migrate
When block migration is requested and no read-write block device is
present, a divide by zero exception is triggered because
total_sector_sum equals zero.
Blue Swirl [Sun, 23 Jan 2011 11:43:25 +0000 (11:43 +0000)]
gdbstub-xml: avoid a warning from sparse
Include a header to get the declaration for xml_builtin. This
avoids a warning from sparse:
CC m68k-softmmu/gdbstub-xml.o
gdbstub-xml.c:244:12: warning: symbol 'xml_builtin' was not declared. Should it be static?
* Optimize handling when carry is not updated.
* Optimize handling for adds with nop semantics.
* Move code from helper_addkc to the translator making
helper_addkc PURE and CONST.
Aurelien Jarno [Fri, 21 Jan 2011 16:56:32 +0000 (17:56 +0100)]
sm501: fix screen redraw
Due to signed/unsigned comparison, the dirty bits are never reset, and
the screen redrawn each time. Fix that by only using ram_addr_t types,
and looking for page_min != addr_max instead.
Aurelien Jarno [Thu, 20 Jan 2011 20:40:53 +0000 (21:40 +0100)]
gt64xxx: set isa_mem_base during registration
isa_mem_base is computed from registers during reset, but due to QEMU
limitations some devices (e.g. VGA card) need to know it earlier when
they are registered.
Workaround this by setting the value during registration instead of
reset.
Peter Maydell [Thu, 20 Jan 2011 16:04:52 +0000 (16:04 +0000)]
hw/pl190.c: Fix writing of default vector address
The PL190 implementation keeps the default vector address
in vect_addr[16], but we weren't using this for writes to
the DEFVECTADDR register. As a result of this fix the
default_addr structure member is unused and we can delete it.
Aurelien Jarno [Mon, 17 Jan 2011 18:29:33 +0000 (19:29 +0100)]
target-ppc: fix wrong NaN tests
Some tests in FPU emulation code were wrongly using float64_is_nan()
before commit 185698715dfb18c82ad2a5dbc169908602d43e81, and wrongly
using float64_is_quiet_nan() after. Fix them by using float64_is_any_nan()
instead.
Aurelien Jarno [Mon, 17 Jan 2011 18:29:33 +0000 (19:29 +0100)]
target-ppc: fix sNaN propagation
The current FPU code returns 0.0 if one of the operand is a
signaling NaN and the VXSNAN exception is disabled.
fload_invalid_op_excp() doesn't return a qNaN in case of a VXSNAN
exception as the operand should be propagated instead of a new
qNaN to be generated. Fix that by calling fload_invalid_op_excp()
only for the exception generation (if enabled), and use the softfloat
code to correctly compute the result.
Isaku Yamahata [Thu, 20 Jan 2011 07:21:38 +0000 (16:21 +0900)]
pci: deassert intx on reset.
deassert intx on device reset.
So far pci_device_reset() is used for system reset.
In that case, interrupt controller is reset at the same time so that
all irq is are deasserted.
But now pci bus reset/flr is supported, and in that case irq needs to be
disabled explicitly.
Recently PXA2xx lcd have stopped to be updated incrementally (picture
frozen). This patch fixes that by passing non min/max x/y, but rather
(correctly) x/y and w/h.
vmstate_pxa2xx_i2c incorrectly recursed to itself instead of going
to store slave device. Fix that stop stop qemu from segfaulting
during savevm for pxa2xx-based devices.
scoop: fix access to registers from second instance
Second instance of scoop contains registers shifted to 0x40 from the start
of the page. Instead of messing with register mapping, just limit register
address to 0x00..0x3f.
mainstone: fix name of the allocated memory for roms
Mainstone board has two flash chips (emulated by two ram regions), however
currently code tries to allocate them with the same name, which fails.
Fix that to make mainstone emulation work again.
Fred Boiteux [Sun, 9 Jan 2011 13:24:59 +0000 (14:24 +0100)]
add bepo (french dvorak) keyboard layout
I'm using the Qemu program with VNC I/O, and I had some problems with
my keyboard layout, so I've prepared a definition to be included in
Qemu, built from Xorg description.
floatx80_is_{quiet,signaling}_nan() functions are incorrectly detecting
the type of NaN, depending on SNAN_BIT_IS_ONE, one of the two is
returning the correct value, and the other true for any kind of NaN.
This patch fixes that by applying the same kind of comparison as for
other float formats, but taking into account the explicit bit.
Patch a6a7005d14b3c32d4864a718fb1cb19c789f58a5 generated
broken device paths. We snprintf with a length shorter
than the output, so the last character is discarded and replaced
by the null byte. Fix it up by snprintf to a buffer
which is larger by 1 byte and then memcpy the data (without
the null byte) to where we need it.
Blue Swirl [Tue, 18 Jan 2011 21:34:51 +0000 (21:34 +0000)]
sparc: fix NaN handling
Fix several bugs in NaN handling:
* e in fcmpe* only changes qNaN handling
* FCC is unchanged if an exception is raised
* clear previous FTT before setting it
When reading cp0_count from a timer with a late trigger that should
already have expired, expire it and raise the timer irq.
This makes it possible for guest code (e.g, Linux) that first read
cp0_count, then compare it with cp0_compare and check for raised
timer interrupt lines to run reliably.
Stefan Weil [Sun, 16 Jan 2011 15:28:20 +0000 (16:28 +0100)]
bsd-user: Fix possible memory leaks and wrong realloc call
These errors were reported by cppcheck:
[bsd-user/elfload.c:1108]: (error) Common realloc mistake: "syms" nulled but not freed upon failure
[bsd-user/elfload.c:1076]: (error) Memory leak: s
[bsd-user/elfload.c:1079]: (error) Memory leak: syms
v2:
* The previous fix for memory leaks was incomplete (thanks to Peter Maydell for te hint).
* Fix wrong realloc usage, too.
Alex Williamson [Tue, 11 Jan 2011 21:39:43 +0000 (14:39 -0700)]
savevm: Fix no_migrate
The no_migrate save state flag is currently only checked in the
last phase of migration. This means that we potentially waste
a lot of time and bandwidth with the live state handlers before
we ever check the no_migrate flags. The error message printed
when we catch a non-migratable device doesn't get printed for
a detached migration. And, no_migrate does nothing to prevent
an incoming migration to a target that includes a non-migratable
device. This attempts to fix all of these.
One notable difference in behavior is that an outgoing migration
now checks for non-migratable devices before ever connecting to
the target system. This means the target will remain listening
rather than exit from failure.
Michael Tokarev [Mon, 18 Oct 2010 12:55:25 +0000 (16:55 +0400)]
USB keyboard emulation key mapping error
The USB keyboard emulation's translation table in hw/usb-hid.c doesn't
match the codes actually sent for the Logo (a.k.a. "Windows") or Menu
keys. This results in the guest OS not being able to receive these keys
at all when the USB keyboard emulation is being used.
In particular, both the keymap in /usr/share/kvm/keymaps/modifiers and
the evdev table in x_keymap.c map these keys to 0xdb, 0xdc, and 0xdd,
while usb_hid_usage_keys[] seems to be expecting them to be mapped to
0x7d, 0x7e, and 0x7f.
The attached patch seems to fix the problem, at least in my (limited)
testing.
http://bugs.debian.org/578846
http://bugs.debian.org/600593 (cloned from the above against different pkg)
https://bugs.launchpad.net/qemu/+bug/584139
Blue Swirl [Sun, 16 Jan 2011 08:32:27 +0000 (08:32 +0000)]
tcg/sparc64: fix segfault
With current OpenBSD, code_gen_buffer was mapped 8GB away from
text segment. Then any helpers were beyond the 2GB range of call
instruction genereated by TCG and so the calls would go nowhere,
leading to a segfault.
Fix by specifying an address for the code_gen_buffer,
hopefully free and nearby the helpers.
Blue Swirl [Sat, 15 Jan 2011 08:31:00 +0000 (08:31 +0000)]
monitor: fix a typo
Fix usage of wrong variable, spotted by clang:
/src/qemu/monitor.c:2278:36: warning: The left operand of '&' is a garbage value
prot = pde & (PG_USER_MASK | PG_RW_MASK |
Stefan Weil [Thu, 30 Dec 2010 12:04:57 +0000 (12:04 +0000)]
make_device_config: Fix non-fatal error message with dash and other shells
ORS=" " adds a blank to the name of the include file.
Some shells (e.g. dash) don't accept input redirection
(tr -d '\r' < $f) when $f ends with a blank, so they
print an error message instead of reading pci.mak.
This is a non-fatal error because pci.mak does not
contain an include line. It was introduced by commit 5d6b423c5cd6f9dfac30959ff1d5c088996719c3.
Using printf avoids adding a blank and is also supported
by older awk versions (this solution was suggested by
Paolo Bonzini, thank you).
Aurelien Jarno [Fri, 14 Jan 2011 19:39:19 +0000 (20:39 +0100)]
MAINTAINERS: add entries for TCG
The MAINTAINERS file was lacking entries concerning the TCG code, add
them based on the git history.
For the common TCG code, is probably better to keep [email protected]
as this code can break easily, so it's better to get it reviewed by a few
persons.
Peter Maydell [Fri, 14 Jan 2011 19:39:19 +0000 (20:39 +0100)]
target-arm: Restore IT bits when resuming after an exception
We were not correctly restoring the IT bits when resuming execution
after taking an unexpected exception in the middle of an IT block.
Fix this by tracking them along with PC changes and restoring in
gen_pc_load().
This fixes bug https://bugs.launchpad.net/qemu/+bug/581335
Peter Maydell [Fri, 14 Jan 2011 19:39:19 +0000 (20:39 +0100)]
linux-user: ARM: clear the IT bits when invoking a signal handler
When invoking a signal handler for an ARM target, make sure the IT
bits in the CPSR are cleared. (This would otherwise cause incorrect
execution if the IT state was non-zero when an exception occured.
This bug has been masked previously because we weren't getting the
IT state bits at exception entry right anyway.)
Also use the proper cpsr_read()/cpsr_write() interface to update
the CPSR rather than manipulating CPUState fields directly.
projects / qemu.git / log