Alexander Graf [Tue, 1 Feb 2011 14:51:31 +0000 (15:51 +0100)]
ahci: make number of ports runtime determined
Different AHCI controllers have a different number of ports, so the core
shouldn't care about the amount of ports available.
This patch makes the number of ports available to the AHCI core runtime
configurable, allowing us to have multiple different AHCI implementations
with different amounts of ports.
Alexander Graf [Tue, 1 Feb 2011 14:51:30 +0000 (15:51 +0100)]
ahci: Implement HBA reset
The ahci code was missing its soft reset functionality. This wasn't really an
issue for Linux guests, but Windows gets confused when the controller doesn't
reset when it tells it so.
Using this patch I can now successfully boot Windows 7 from AHCI using AHCI
enabled SeaBIOS.
Alexander Graf [Tue, 1 Feb 2011 14:51:28 +0000 (15:51 +0100)]
ahci: split ICH and AHCI even more
Sebastian's patch already did a pretty good job at splitting up ICH-9
AHCI code and the AHCI core. We need some more though. Copyright was missing,
the lspci dump belongs to ICH-9, we don't need the AHCI core to have its
own qdev device duplicate.
So let's split them a bit more in this patch, making things easier to
read an understand.
There are multiple ahci devices out there. The currently implemented ich-9
is only one of the many. So let's split that one out into a separate file
to stress the difference.
Stefan Weil [Fri, 4 Feb 2011 20:01:16 +0000 (21:01 +0100)]
block/vdi: Fix wrong size in conditionally used memset, memcmp
Error report from cppcheck:
block/vdi.c:122: error: Using sizeof for array given as function argument returns the size of pointer.
block/vdi.c:128: error: Using sizeof for array given as function argument returns the size of pointer.
Fix both by setting the correct size.
The buggy code is only used when QEMU is build without uuid support.
The bug is not critical, so there is no urgent need to apply it to
old versions of QEMU.
Peter Maydell [Thu, 3 Feb 2011 19:43:25 +0000 (19:43 +0000)]
target-arm: Fix decoding of Thumb preload and hint space
Refine the decoding of the Thumb preload and hint space, so we
UNDEF on the patterns that are supposed to UNDEF rather than NOP.
We also move the tests for this space earlier, so we don't emit
harmless but unnecessary address generation code for preload
hints (which by their nature are likely to be in hot code paths).
Peter Maydell [Thu, 3 Feb 2011 19:43:24 +0000 (19:43 +0000)]
target-arm: Fix decoding of preload and memory hint space
Correct the decoding of the ARM preload and memory hint space,
by adding decoding of PLI, PLDW and the v7MP unallocated hint
space. This commit also corrects a slightly overexuberant
decoding of PLD(register) which was not checking that bit 4
was one.
Jan Kiszka [Thu, 3 Feb 2011 21:54:13 +0000 (22:54 +0100)]
ioapic: Add support for qemu-kvm's vmstate v2
qemu-kvm carries the IOAPIC base address in its v2 vmstate. We only
support the default base address so far, and saving even that in the
device state was rejected.
Add a padding field to be able to read qemu-kvm's old state, but
increase our version to 3, indicating that we are not saving a valid
address. This also gives downstream the chance to change to stop
evaluating the base_address and move to v3 as well.
Jan Kiszka [Thu, 3 Feb 2011 21:54:11 +0000 (22:54 +0100)]
ioapic: Implement EOI handling for level-triggered IRQs
Add the missing EOI broadcast from local APIC to the IOAPICs on
completion of level-triggered IRQs. This ensures that a still asserted
IRQ source properly re-triggers an APIC IRQ.
Corentin Chary [Fri, 4 Feb 2011 08:05:53 +0000 (09:05 +0100)]
vnc: qemu can die if the client is disconnected while updating screen
agraf reported that qemu_mutex_destroy(vs->output_mutex) while failing
in vnc_disconnect_finish().
It's because vnc_worker_thread_loop() tries to unlock the mutex while
not locked. The unlocking call doesn't fail (pthread bug ?), but
the destroy call does.
Paolo Bonzini [Thu, 3 Feb 2011 13:49:01 +0000 (14:49 +0100)]
Unify alarm deadline computation
This patch shows how using the correct formula for
qemu_next_deadline_dyntick can simplify the code of
host_alarm_handler and eliminate useless duplication.
Paolo Bonzini [Thu, 3 Feb 2011 13:49:00 +0000 (14:49 +0100)]
Correct alarm deadline computation
When the QEMU_CLOCK_HOST clock was added, computation of its
deadline was added to qemu_next_deadline, which is correct but
incomplete.
I noticed this by reading the very convoluted rules whereby
qemu_next_deadline_dyntick is computed, which miss QEMU_CLOCK_HOST
when use_icount is true. This patch inlines qemu_next_deadline
into qemu_next_deadline_dyntick, and then corrects the logic to skip
only QEMU_CLOCK_VIRTUAL when use_icount is true.
Yoshiaki Tamura [Thu, 3 Feb 2011 04:34:08 +0000 (13:34 +0900)]
savevm: fix corruption in vmstate_subsection_load().
Although it's rare to happen in live migration, when the head of a
byte stream contains 0x05 which is the marker of subsection, the
loader gets corrupted because vmstate_subsection_load() continues even
the device doesn't require it. This patch adds a checker whether
subsection is needed, and skips following routines if not needed.
Anthony Liguori [Tue, 1 Feb 2011 22:57:00 +0000 (16:57 -0600)]
Update SeaBIOS to 0.6.1.2
- 06d0bdd Minor build fixes.
- 33abfc0 Update version to 0.6.1.2.
- 484dd56 fix virtio-blk failure after reboot
- dd9c0d3 Update version to 0.6.1.1.
- 50ecfa8 mark irq9 active high in DSDT
When MSI is off, each interrupt needs to be bounced through the io
thread when it's set/cleared, so vhost-net causes more context switches and
higher CPU utilization than userspace virtio which handles networking in
the same thread.
We'll need to fix this by adding level irq support in kvm irqfd,
for now disable vhost-net in these configurations.
Added a vhostforce flag to force vhost-net back on.
With current sndbuf default value, a blocked
target guest can prevent another guest from
transmitting any packets. While current
sndbuf value (1M) is reported to help some
UDP based workloads, the default should
be safe (0).
Anthony Liguori [Mon, 31 Jan 2011 20:27:36 +0000 (14:27 -0600)]
vnc: Fix password expiration through 'change vnc ""' (v2)
commit 52c18be9e99dabe295321153fda7fce9f76647ac introduced a regression in the
change vnc password command that changed the behavior of setting the VNC
password to an empty string from disabling login to disabling authentication.
This commit refactors the code to eliminate this overloaded semantics in
vnc_display_password and instead introduces the vnc_display_disable_login. The
monitor implementation then determines the behavior of an empty or missing
string.
Recently, a set_password command was added that allows both the Spice and VNC
password to be set. This command has not shown up in a release yet so the
behavior is not yet defined.
This patch proposes that an empty password be treated as an empty password with
no special handling. For specifically disabling login, I believe a new command
should be introduced instead of overloading semantics.
I'm not sure how Spice handles this but I would recommend that we have Spice
and VNC have consistent semantics here for the 0.14.0 release.
Reported-by: Neil Wilson <[email protected]> Signed-off-by: Anthony Liguori <[email protected]>
---
v1 -> v2
- Add a proper return to make sure that login is really disabled instead of
relying on the VNC server to treat empty passwords specially
Peter Maydell [Tue, 1 Feb 2011 15:54:52 +0000 (15:54 +0000)]
linux-user: avoid gcc array overrun warning for sparc
Suppress a gcc array bounds overrun warning when filling in the SPARC
signal frame by adjusting our definition of the structure so that the
fp and callers_pc membes are part of the ins[] array rather than
separate fields; since qemu has no need to access the fields individually
there is no need to follow the kernel's structure field naming exactly.
Peter Maydell [Mon, 31 Jan 2011 10:42:26 +0000 (10:42 +0000)]
hw/slavio_intctl.c: fix gcc warning about array bounds overrun
The Ubuntu 10.10 gcc for ARM complains that we might be overrunning
the cpu_irqs[][] array: silence this by correcting the bounds on the
loop. (In fact we would not have overrun the array because bit
MAX_PILS in pil_pending and irl_out will always be 0.)
Also add a comment about why the loop's lower bound is OK.
add_init_drive() is confused about drive_init()'s failure modes, and
cleans up when it shouldn't. This leaves the DriveInfo with member
opts dangling. drive_del attempts to free it, and dies.
drive_init() behaves as follows:
* If it created a drive with media, it returns its DriveInfo.
* If it created a drive without media, it clears *fatal_error and
returns NULL.
* If it couldn't create a drive, it sets *fatal_error and returns
NULL.
Of its three callers:
* drive_init_func() is correct.
* usb_msd_init() assumes drive_init() failed when it returns NULL.
This is correct only because it always passes option "file", and
"drive without media" can't happen then.
* add_init_drive() assumes drive_init() failed when it returns NULL.
This is incorrect.
Clean up drive_init() to return NULL on failure and only on failure.
Drop its parameter fatal_error.
The second -device fails, because it refers to drive zwei, which got
silently ignored.
Make multiple drive definitions fail cleanly.
Unfortunately, there's code that relies on multiple drive definitions
being silently ignored: main() merrily adds default drives even when
the user already defined these drives. Fix that up.
blockdev: Make drive_add() take explicit type, index parameters
Before, type & index were hidden in printf-like fmt, ... parameters,
which get expanded into an option string. Rather inconvenient for
uses later in this series.
New IF_DEFAULT to ask for the machine's default interface. Before,
that was done by having no option "if" in the option string.
blockdev: Fix regression in -drive if=scsi,index=N
Before commit 622b520f, index=12 meant bus=1,unit=5.
Since the commit, it means bus=0,unit=12. The drive is created, but
not the guest device. That's because the controllers we use with
if=scsi drives (lsi53c895a and esp) support only 7 units, and
scsi_bus_legacy_handle_cmdline() ignores drives with unit numbers
exceeding that limit.
Changing the mapping of index to bus, unit is a regression. Breaking
-drive invocations that used to work just makes it worse.
Revert the part of commit 622b520f that causes this, and clean up
some.
Note that the fix only affects if=scsi. You can still put more than 7
units on a SCSI bus with -device & friends.
blockdev: New drive_get_next(), replacing qdev_init_bdrv()
qdev_init_bdrv() doesn't belong into qdev.c; it's about drives, not
qdevs. Rename to drive_get_next, move to blockdev.c, drop the bogus
DeviceState argument, and return DriveInfo instead of
BlockDriverState.
Blue Swirl [Sat, 29 Jan 2011 08:19:15 +0000 (08:19 +0000)]
qcow2-refcount: remove write-only variables
Variables l2_modified and l2_size are not really used, remove them.
Spotted by GCC 4.6.0:
CC block/qcow2-refcount.o
/src/qemu/block/qcow2-refcount.c: In function 'qcow2_update_snapshot_refcount':
/src/qemu/block/qcow2-refcount.c:708:37: error: variable 'l2_modified' set but not used [-Werror=unused-but-set-variable]
/src/qemu/block/qcow2-refcount.c:708:9: error: variable 'l2_size' set but not used [-Werror=unused-but-set-variable]
drive_init() picks the first free bus and unit number, unless the user
specifies them.
This isn't a good fit for the drive_add monitor command, because there
we specify the controller by PCI address instead of using bus number
set by drive_init().
scsi_hot_add() takes care to replace the unit number set by
drive_init() by the real one, but it neglects to replace the bus
number. Thus, bus/unit in DriveInfo may be bogus. Affects
drive_get() and drive_get_max_bus(). I'm not aware of anything bad
happening because of that; looks like by the time we're hot-plugging,
the two functions aren't used anymore. Fix it anyway.
Stefan Hajnoczi [Fri, 28 Jan 2011 17:11:59 +0000 (17:11 +0000)]
qed: Images with backing file do not require QED_F_NEED_CHECK
The consistency check on open is necessary in order to fix inconsistent
table offsets left as a result of a crash mid-operation. Images with a
backing file actually flush before updating table offsets and are
therefore guaranteed to be consistent. Do not mark these images dirty.
Kevin Wolf [Wed, 26 Jan 2011 15:56:48 +0000 (16:56 +0100)]
qcow2: Add bdrv_discard support
This adds a bdrv_discard function to qcow2 that frees the discarded clusters.
It does not yet pass the discard on to the underlying file system driver, but
the space can be reused by future writes to the image.
Stefan Hajnoczi [Tue, 25 Jan 2011 16:17:14 +0000 (16:17 +0000)]
virtio-pci: Disable virtio-ioeventfd when !CONFIG_IOTHREAD
It is not possible to use virtio-ioeventfd when building without an I/O
thread. We rely on a signal to kick us out of vcpu execution. Timers
and AIO use SIGALRM and SIGUSR2 respectively. Unfortunately eventfd
does not support O_ASYNC (SIGIO) so eventfd cannot be used in a signal
driven manner.
Raise a config change interrupt when the size changed. This allows
virtio-blk guest drivers to read-read the information from the
config space once it got the config chaged interrupt.
Add a monitor command that allows resizing of block devices while
qemu is running. It uses the existing bdrv_truncate method already
used by qemu-img to do it's work. Compared to qemu-img the size
parsing is very simplicistic, but I think having a properly numering
object is more useful for non-humand monitor users than having
the units and relative resize parsing.
For SCSI devices the new size can be updated in Linux guests by
doing the following shell command:
For ATA devices I don't know of a way to update the block device
size in Linux system, and for virtio-blk the next two patches
will provide an automatic update of the size when this command
is issued on the host.
Blue Swirl [Sat, 29 Jan 2011 22:52:33 +0000 (22:52 +0000)]
sdl: remove unused variable
Variable rec is not used, remove it. Spotted by GCC 4.6.0:
CC ui/sdl.o
/src/qemu/ui/sdl.c: In function 'sdl_setdata':
/src/qemu/ui/sdl.c:90:14: error: variable 'rec' set but not used [-Werror=unused-but-set-variable]
Andreas Färber [Tue, 18 Jan 2011 21:43:55 +0000 (22:43 +0100)]
prep: Remove bogus BIOS size check
r3480 added this check to account for the entry vector 0xfff00100 to be
available for CPUs that need it. Today however, the NIP is not yet
initialized at this point (zero), so the check always triggers.
Moreover, BIOS size check is already done previously, so this part can
be removed too.
Aurelien Jarno [Tue, 25 Jan 2011 10:55:15 +0000 (11:55 +0100)]
mc146818rtc: update registers after a format change
For some unknown reason, the MIPS kernel briefly changes the RTC to
binary mode during boot, switch back to BCD mode and read the time. As
the registers are updated only every second, they may still be in the
old format when they are read.
This patch forces a register update immediately after a format change
(BCD/binary or 12/24H). This avoid long fsck during boot due to time
wrap.
Aurelien Jarno [Tue, 25 Jan 2011 10:55:14 +0000 (11:55 +0100)]
virtio-blk: fix cross-endianness targets
virtio-blk doesn't work on cross-endian configuration, as endianness is
not handled correctly.
This patch adds missing endianness conversions to make virtio-blk
working. Tested on the following configurations:
- i386 guest on x86_64 host
- ppc guest on x86_64 host
- i386 guest on mips host
- ppc guest on mips host
Aurelien Jarno [Tue, 25 Jan 2011 10:55:14 +0000 (11:55 +0100)]
virtio-net: fix cross-endianness support
virtio-net used to work on cross-endianness configurations, but doesn't
anymore with recent guest kernels, as the new features don't handle
endianness correctly.
This patch fixes wrong conversion, and add missing ones to make
virtio-net working. Tested on the following configurations:
- i386 guest on x86_64 host
- ppc guest on x86_64 host
- i386 guest on mips host
- ppc guest on mips host
Aurelien Jarno [Thu, 27 Jan 2011 07:21:35 +0000 (08:21 +0100)]
escc: fix interrupt flags
Recent PowerPC kernel end up in kernel panic during boot in -nographic
mode. In this mode the second serial port is used as the udbg console,
and thus a few characters are sent on this port. This activates the
tx interrupt flag, and later choke the Linux kernel, as it was not
expecting such a flag to be set.
The problem here comes from the fact that contrary to most devices the
interrupt flags are only set if the interrupt is enabled. Quoting the
datasheet: "If the corresponding IE bit is not set, the IP for that
source of interrupt will never be set."
This patch fixes that by enabling the interrupt flag only when the
corresponding interrupt is enabled.
projects / qemu.git / log