aurel32 [Sun, 29 Mar 2009 01:18:03 +0000 (01:18 +0000)]
target-ppc: avoid nop to override next instruction
While searching PC, always store the pc of a new instruction.
Instructions that didn't generate tcg code (such as nop) prevented the
next one to be referenced.
aurel32 [Sun, 29 Mar 2009 01:04:39 +0000 (01:04 +0000)]
target-alpha: bug fix: avoid nop to override next instruction
While searching PC, always store the pc of a new instruction.
Instructions that didn't generate tcg code (such as nop) prevented the next
one to be referenced.
aurel32 [Sat, 28 Mar 2009 23:14:52 +0000 (23:14 +0000)]
hw/eeprom93xx.c: support 93xx EEPROMs with more than 255 words
In the head of eeprom93xx.c we promise to support chips with 256 words,
but store the size in an unsigned byte. This patch replaces this with an
16 bit variable and changes the load/store code accordingly (introducing a
new version).
aurel32 [Sat, 28 Mar 2009 23:14:41 +0000 (23:14 +0000)]
hw/eeprom93xx.c: substitute structure dump with discrete dump in eeprom_save/load
The EEPROM 93xx device used to dump a C structure to the migration stream.
This structure includes mixed 8 and 16bit variables and is thus subject to
compiler dependent padding. Replace this with discrete dumps of each member
(and add a padding byte to ensure compatibility, a version update is
included in the following patch).
aurel32 [Sat, 28 Mar 2009 23:14:32 +0000 (23:14 +0000)]
target-sh4: r2d --append option support
Add linux kernel command line ("--append" option) support.
Fix kernel loading address to appropriate position when --append used.
Using --kernel but --append case is left untouched for backward compatibility.
This also change the host<->SH address mapping for r2d to
host addr == phys_ram_base + SH addr.
aliguori [Sat, 28 Mar 2009 18:05:53 +0000 (18:05 +0000)]
gdbstub: Allow re-instantiation (Jan Kiszka)
[ Note: depends on char closing fixes ]
Properly clean up the gdbstub when the user tries to re-open it
(possibly under a different address). Moreover, allow to shut it down
from the monitor via 'gdbserver none'.
aliguori [Sat, 28 Mar 2009 17:58:14 +0000 (17:58 +0000)]
char: Fix closing of various char devices (Jan Kiszka)
This patch fixes several issues around closing char devices. Affected
were pty (timer was left behind, even running), udp (no close handling
at all) and tcp (missing async IO handler cleanup). The bugs either
caused segfaults or stalled the qemu process. So far, hot-unplugging USB
serial adapters suffered from this.
aliguori [Sat, 28 Mar 2009 17:55:19 +0000 (17:55 +0000)]
qemu-img: adding a "-F base_fmt" option to "qemu-img create -b" (Uri Lublin)
If the user specifies the backing file format,
then when opening the backing file, there is no need
to probe the (backing file) image to figure out its format.
This follows my previous patches implementing bdrv_create2
which keeps (for qcow2 only) the backing file format
as a qcow2-extension
aliguori [Sat, 28 Mar 2009 17:55:14 +0000 (17:55 +0000)]
block-qcow2: keep backing file format in a qcow2 extension (Uri Lublin)
Use a qcow2 extension to keep the backing file format.
By keeping the backing file format, we can:
1. Provide a way to know the backing file format without probing
it (setting the format at creation time).
2. Enable using qcow2 format over host block devices.
(only if the user specifically asks for it, by providing the format
at creation time).
Also fixes a security flaw found by Daniel P. Berrange on [1]
which summarizes: "Autoprobing: just say no."
aliguori [Sat, 28 Mar 2009 17:55:10 +0000 (17:55 +0000)]
block: support known backing format for image create and open (Uri Lublin)
Added a backing_format field to BlockDriverState.
Added bdrv_create2 and drv->bdrv_create2 to create an image with
a known backing file format.
Upon bdrv_open2 if backing format is known use it, instead of
probing the (backing) image.
aliguori [Sat, 28 Mar 2009 17:55:06 +0000 (17:55 +0000)]
Introducing qcow2 extensions (Uri Lublin)
Qcow2 extensions are build of magic (id) len (in bytes) and data.
They reside right after the qcow2 header.
If a backing filename exists it follows the qcow2 extension (if exist)
Qcow2 extensions are read upon image open.
Qcow2 extensions are identified by their magic.
Unknown qcow2 extensions (unknown magic) are skipped.
A Special magic of 0 means end-of-qcow2-extensions.
In this patchset, to be used to keep backing file format.
aliguori [Sat, 28 Mar 2009 17:51:36 +0000 (17:51 +0000)]
ROM write access for debugging (Jan Kiszka)
Enhance cpu_memory_rw_debug so that it can write even to ROM regions.
This allows to modify ROM via gdb (I see no point in denying this to the
user), and it will enable us to drop kvm_patch_opcode_byte().
aliguori [Sat, 28 Mar 2009 17:46:14 +0000 (17:46 +0000)]
virtio-blk: use generic vectored I/O APIs (Christoph Hellwig)
Use the generic bdrv_aio_readv/bdrv_aio_writev APIs instead of linearizing
buffers directly. This enables using the future native preadv/pwritev
support.
aliguori [Sat, 28 Mar 2009 17:46:10 +0000 (17:46 +0000)]
add qemu_iovec_init_external (Christoph Hellwig)
Allow to initialize a QEMUIOVector from an externally allocated iovec.
qiov->nalloc is initialized to -1 to indicate external storage for qiov->iov
and all functions dealing with memory management assert on the iovec beeing
an internally managed first.
aliguori [Sat, 28 Mar 2009 17:29:07 +0000 (17:29 +0000)]
Clean some PCI defines (Stefan Weil)
this patch adds some more defines from linux/pci_regs.h to
hw/pci.h. There is now no longer a need to define them in
eepro100.c, so they were removed there.
Some defines from linux/pci_regs.h had similar, but not
the same defines in hw/pci.h (PCI_REVISION_ID / PCI_REVISION,
PCI_SUBSYSTEM_VENDOR_ID / PCI_SUBVENDOR_ID,
PCI_SUBSYSTEM_ID / PCI_SUBDEVICE_ID).
I suggest to use the "standard" from linux/pci_regs.h and
replace the "old" Qemu ones. To facilitate the migration,
my patch does not remove the old defines but marks them
as obsolete. After a migration to the "standard" defines,
pci.h could use linux/pci_regs.h which is far more complete.
The patch is needed for an updated maintainer version of
hw/eepro100.c which I'd like to see in Qemu stable.
* Remove declarations already declared in header file from eepro100.c
* Add missing declarations from pci_regs.h to pci.h
* Mark "non-standard" declarations in pci.h as obsolete
aliguori [Sat, 28 Mar 2009 17:28:58 +0000 (17:28 +0000)]
Remove nodisk_ok machine feature (Jan Kiszka)
All archs have some kind of firmware to load and can be fine with it
already. So there is not much use in enforcing the presence of a disk.
If the system setup requires one, the user will notice it anyway once
the firmware/bios fails to boot from it.
aliguori [Sat, 28 Mar 2009 17:28:45 +0000 (17:28 +0000)]
get roms more room. (Glauber Costa)
This patch increases by 50 % the size available for option roms.
The main motivator is that some roms grew bigger than the 64k we
currently allocate for them (Hey, it's 2009!)
One example is the gpxe project, that produces some roms with 69k,
70k, etc. The space proposed by this patch actually makes it as
big as 84k. Probably still a fit for some time.
But there is no free lunch. This space must come from somewhere,
and we take it from vga rom space. Currently, our vga roms are
around 35k in size. With this patch, option rom space will begin
just after vga ends, aligned to the next 2k boundary.
Technicaly, we could do the same with the uper space (the bios itself),
but since bochs bios is already 128 k in size, I don't see an
urgent need to do it.
[ fix case for vgabioses smaller than 30k, by Carl-Daniel Hailfinger ]
aliguori [Sat, 28 Mar 2009 17:28:41 +0000 (17:28 +0000)]
new scsi-generic abstraction, use SG_IO (Christoph Hellwig)
Okay, I started looking into how to handle scsi-generic I/O in the
new world order.
I think the best is to use the SG_IO ioctl instead of the read/write
interface as that allows us to support scsi passthrough on disk/cdrom
devices, too. See Hannes patch on the kvm list from August for an
example.
Now that we always do ioctls we don't need another abstraction than
bdrv_ioctl for the synchronous requests for now, and for asynchronous
requests I've added a aio_ioctl abstraction keeping it simple.
Long-term we might want to move the ops to a higher-level abstraction
and let the low-level code fill out the request header, but I'm lazy
enough to leave that to the people trying to support scsi-passthrough
on a non-Linux OS.
Tested lightly by issuing various sg_ commands from sg3-utils in a guest
to a host CDROM device.
aliguori [Sat, 28 Mar 2009 16:11:25 +0000 (16:11 +0000)]
Fix DMA API when handling an immediate error from block layer (Avi Kivity)
The block layer may signal an immediate error on an asynchronous request
by returning NULL. The DMA API did not handle this correctly, returning
an AIO request which would never complete (and which would crash if
cancelled).
If a bounced vectored aio fails immediately (the inner aio submission
returning NULL) then the bounce handler erronously returns an aio
request which will never be completed (and which crashes when cancelled).
Fix by detecting that the inner request has failed and propagating the
error.
aliguori [Sat, 28 Mar 2009 15:58:58 +0000 (15:58 +0000)]
host_device_remove: remove incorrect check for device name (Eduardo Habkost)
There is no need to check for valid prefixes on the the device name
when removing it. If the device name is found on the vlan client list,
it can be removed, regardless of the prefix used on its name.
To reproduce the bug, just run this on the monitor:
blueswir1 [Sat, 28 Mar 2009 08:37:13 +0000 (08:37 +0000)]
FreeBSD host physical cdrom fixes
This improves physical cdrom support on FreeBSD hosts to be almost as
good as on Linux, with the only notable exception that you still need to
either have the guest itself eject the disc if you want to take it
out/change it, or do a change command in the monitor after taking out
a disc in case a guest cannot eject it itself - otherwise the guest may
continue using state (like size) of the old disc.
blueswir1 [Sat, 28 Mar 2009 06:44:27 +0000 (06:44 +0000)]
Syncing documentation vs. -help vs. qemu_options table
Try to keep documentation about command line switches, -help text and
qemu_options table synchronized.
In true Qemu tradition, an include file is generated from single .hx file
containing all relevant information in one place. The include file is
parsed once for getting the enums, another time for getopt tables and
hird time for help messages. Texi documentation for the options is
generated from the same .hx file.
balrog [Sat, 21 Mar 2009 01:09:16 +0000 (01:09 +0000)]
Fix absolute mouse events generated by SDL frontend.
In r6839 ("DisplayAllocator interface") the "width" and "height" globals
stopped ever being assigned. Note that last time absolute input stopped
working was for the same reason.
aliguori [Fri, 20 Mar 2009 18:26:12 +0000 (18:26 +0000)]
Use vectored aiocb storage to store vector translation state (Avi Kivity)
Now that we have a dedicated acb pool for vector translation acbs, we can
store the vector translation state in the acbs instead of in an external
structure.
aliguori [Fri, 20 Mar 2009 18:26:07 +0000 (18:26 +0000)]
Implement cancellation method for dma async I/O (Avi Kivity)
Move the dma helpers to a private aio pool, and implement a cancellation
method for them. Should prevent issues when cancelling I/O while dma is
in progress.
aliguori [Fri, 20 Mar 2009 18:25:59 +0000 (18:25 +0000)]
Refactor aio callback allocation to use an aiocb pool (Avi Kivity)
Move the AIOCB allocation code to use a dedicate structure, AIOPool. AIOCB
specific information, such as the AIOCB size and cancellation routine, is
moved into the pool.
At present, there is exactly one pool per block format driver, maintaining
the status quo.
aliguori [Fri, 20 Mar 2009 16:13:50 +0000 (16:13 +0000)]
virtio: Allow guest to defer VIRTIO_F_NOTIFY_ON_EMPTY (Alex Williamson)
There may be cases where the guest does not want the avail queue
interrupt, even when it's empty. For the virtio-net case, the
guest may use a different buffering scheme or decide polling for
used buffers is more efficient. This can be accomplished by simply
checking for whether the guest has acknowledged the existing notify
on empty flag.
The RXDMT0 interrupt is supposed to fire when the number of free
RX descriptors drops to some fraction of the total descriptors.
However in practice, it seems like we're adding this interrupt
cause on every RX. Fix the logic to treat (tail - head) as the
number of free entries rather than the number of used entries.
aliguori [Fri, 20 Mar 2009 16:13:41 +0000 (16:13 +0000)]
x86: Add NULL check to lsl (Jan Kiszka)
According to the Intel specs, lsl performs a check against NULL for the
provided selector, just like lar does. helper_lar() includes the
corresponding code, helper_lsl() was lacking it so far.
aliguori [Fri, 20 Mar 2009 15:59:24 +0000 (15:59 +0000)]
vnc: throttle screen updates. (Gerd Hoffmann)
This patch makes the vnc server code skip screen refreshes in case
there is data in the output buffer. This reduces the refresh rate to
throttle the bandwidth needed in case the network link is saturated.
This patch killes the old_data hack in the qemu server and replaces
it with a clean separation of the guest-visible display surface and
the vnc server display surface. Both guest and server surface have
their own dirty bitmap for tracking screen updates.
Workflow is this:
(1) The guest writes to the guest surface. With shared buffers being
active the guest writes are directly visible to the vnc server code.
Note that this may happen in parallel to the vnc server code running
(today only in xenfb, once we have vcpu threads in qemu also for
other display adapters).
(2) vnc_update() callback tags the specified area in the guest dirty
map.
(3) vnc_update_client() will first walk through the guest dirty map. It
will compare guest and server surface for all regions tagged dirty
and in case the screen content really did change the server surface
and dirty map are updated.
Note: old code used old_data in a simliar way, so this does *not*
introduce an extra memcpy.
(4) Then vnc_update_cient() will send the updates to the vnc client
using the server surface and dirty map.
Note: old code used the guest-visible surface instead, causing
screen corruption in case of guest screen updates running in
parallel.
The separate dirty bitmap also has the nice effect that forced screen
updates can be done cleanly by simply tagging the area in both guest and
server dirty map. The old, hackish way was memset(old_data, 42, size)
to trick the code checking for screen changes.
blueswir1 [Fri, 13 Mar 2009 21:16:24 +0000 (21:16 +0000)]
Make the ELF loader aware of backwards compatibility
Most 64 bit architectures I'm aware of support running 32 bit code
of the same architecture as well.
So x86_64 can run i386 code easily and ppc64 can run ppc code.
Unfortunately, the current checks are pretty strict. So you can only
load e.g. an x86_64 elf binary on qemu-system-x86_64, but no i386 one.
This can get really annoying. I first encountered this issue with
my multiboot patch, where qemu-system-x86_64 was unable to load an
i386 elf binary because the elf loader rejected it.
The same thing happened again on PPC64 now. The firmware we're loading
is a PPC32 elf binary, as it's shared with PPC32. But the platform is
PPC64.
Right now there is a hack for this in the ppc cpu.h definition, that
simply sets the type to PPC32 in system emulation mode. While that
works fine for the firmware, it's no good if you also want to load a
PPC64 kernel with -kernel.
So in order to solve this mess, I figured the easiest way is to make
the elf loader aware of platforms that are backwards compatible. For
now I was only sure that x86_64 does i386 and ppc64 does ppc32, but
maybe there are other combinations too.
This patch is a prerequisite for having a working -kernel option on
PPC64.
aliguori [Fri, 13 Mar 2009 15:02:23 +0000 (15:02 +0000)]
Add and use remaining #defines for PCI device IDs (Stuart Brady)
This patch adds and uses #defines for the remaining hardcoded PCI
device IDs. It also moves definitions taken from linux/pci_ids.h
into a separate header (hw/pci_ids.h), removes the 'RTL' from
PCI_DEVICE_ID_REALTEK_RTL8029, and renames PCI_DEVICE_ID_FSL_E500
to PCI_DEVICE_ID_MPC8533E to match Linux's definition.
Changes in v2:
* Don't use C99-style comments
* Move definitions from linux/pci_ids.h into a separate header
* Rename PCI_DEVICE_ID_FSL_E500 to PCI_DEVICE_ID_MPC8533E
aliguori [Fri, 13 Mar 2009 15:02:18 +0000 (15:02 +0000)]
remove is_graphic_console from vga.c (Stefano Stabellini)
Hi all,
since vga_draw_graphic is only called by vga_hw_update when the console
associated with the graphic card is active, we don't need to check if
the current console is active using is_graphic_console.
I suspect I introduced these checks when the console switching mechanism
didn't work as it does now.
aliguori [Fri, 13 Mar 2009 15:02:13 +0000 (15:02 +0000)]
DisplayAllocator interface (Stefano Stabellini)
Hi all,
this patch adds a DisplayAllocator interface that allows display
frontends (sdl in particular) to provide a preallocated display buffer
for the graphical backend to use.
Whenever a graphical backend cannot use
qemu_create_displaysurface_from because its own internal pixel format
cannot be exported directly (text mode or graphical mode with color
depth 8 or 24), it creates another display buffer in memory using
qemu_create_displaysurface and does the conversion.
This new buffer needs to be blitted into the sdl surface buffer every time
we need to update portions of the screen.
We can avoid this using the DisplayAllocator interace: sdl provides its
own implementation of qemu_create_displaysurface, giving back the sdl
surface buffer directly (as we used to do before the DisplayState
changes).
Since the buffer returned by sdl could be in bgr format we need to put
back in the handlers of that case.
This approach is good if the two following conditions are true:
1) the sdl surface is a software surface that resides in main memory;
2) the host display color depth is either 16 or 32 bpp.
If first condition is false we can have bad performances using sdl
and vnc together.
If the second condition is false performances are certainly not going to
improve but they shouldn't get worse either.
The first condition is always true, at least on linux/X11 systems; but I
believe is true also on other platforms.
The second condition is true in the vast majority of the cases.
This patch should also have the good side effect of solving the sdl
2D slowness malc was reporting on MacOS, because SDL_BlitSurface is not
going to be called anymore when the guest is in text mode or 24bpp.
However the root problem is still present so I suspect we may
still see some slowness on MacOS when the guest is in 32 or 16 bpp.
aliguori [Fri, 13 Mar 2009 03:12:03 +0000 (03:12 +0000)]
Fix regression introduced by r6824
The changes introduced by r6824 broke a subtle, and admittedly obscure, aspect
of the block API. While bdrv_{pread,pwrite} return the number of bytes read
or written upon success, bdrv_{read,write} returns a zero upon success.
When using bdrv_pread for bdrv_read, special care must be taken to handle this
case.
This fixes certain guest images (notably linux-0.2 provided on the qemu
website).
aliguori [Thu, 12 Mar 2009 20:12:48 +0000 (20:12 +0000)]
Guest debugging support for KVM (Jan Kiszka)
This is a backport of the guest debugging support for the KVM
accelerator that is now part of the KVM tree. It implements the reworked
KVM kernel API for guest debugging (KVM_CAP_SET_GUEST_DEBUG) which is
not yet part of any mainline kernel but will probably be 2.6.30 stuff.
So far supported is x86, but PPC is expected to catch up soon.
Core features are:
- unlimited soft-breakpoints via code patching
- hardware-assisted x86 breakpoints and watchpoints
Changes in this version:
- use generic hook cpu_synchronize_state to transfer registers between
user space and kvm
- push kvm_sw_breakpoints into KVMState
aliguori [Thu, 12 Mar 2009 19:57:16 +0000 (19:57 +0000)]
Drop internal bdrv_pread()/bdrv_pwrite() APIs (Avi Kivity)
Now that scsi generic no longer uses bdrv_pread() and bdrv_pwrite(), we can
drop the corresponding internal APIs, which overlap bdrv_read()/bdrv_write()
and, being byte oriented, are unnatural for a block device.
aliguori [Thu, 12 Mar 2009 19:57:12 +0000 (19:57 +0000)]
Add internal scsi generic block API (Avi Kivity)
Add an internal API for the generic block layer to send scsi generic commands
to block format driver. This means block format drivers no longer need
to consider overloaded nb_sectors parameters.
aliguori [Thu, 12 Mar 2009 19:57:08 +0000 (19:57 +0000)]
Add specialized block driver scsi generic API (Avi Kivity)
When a scsi device is backed by a scsi generic device instead of an
ordinary host block device, the block API is abused in a couple of annoying
ways:
- nb_sectors is negative, and specifies a byte count instead of a sector count
- offset is ignored, since scsi-generic is essentially a packet protocol
This overloading makes hacking the block layer difficult. Remove it by
introducing a new explicit API for scsi-generic devices. The new API
is still backed by the old implementation, but at least the users are
insulated.
aliguori [Wed, 11 Mar 2009 20:05:37 +0000 (20:05 +0000)]
Revert r6404
This series is broken by design as it requires expensive IO operations at
open time causing very long delays when starting a virtual machine for the
first time.
aliguori [Wed, 11 Mar 2009 20:05:33 +0000 (20:05 +0000)]
Revert r6405
This series is broken by design as it requires expensive IO operations at
open time causing very long delays when starting a virtual machine for the
first time.
aliguori [Wed, 11 Mar 2009 20:05:29 +0000 (20:05 +0000)]
Revert r6406
This series is broken by design as it requires expensive IO operations at
open time causing very long delays when starting a virtual machine for the
first time.
aliguori [Wed, 11 Mar 2009 20:05:25 +0000 (20:05 +0000)]
Revert r6407
This series is broken by design as it requires expensive IO operations at
open time causing very long delays when starting a virtual machine for the
first time.
aliguori [Wed, 11 Mar 2009 20:05:20 +0000 (20:05 +0000)]
Revert r6408
This series is broken by design as it requires expensive IO operations at
open time causing very long delays when starting a virtual machine for the
first time.
projects / qemu.git / log