Anthony Liguori [Mon, 10 Dec 2012 14:31:19 +0000 (08:31 -0600)]
Merge remote-tracking branch 'kraxel/usb.74' into staging
* kraxel/usb.74:
usb-tablet: Allow connecting to ehci
ehci: Lower timer freq when the periodic schedule is idle
usb: Allow overriding of usb_desc at the device level
usb: Don't allow USB_RET_ASYNC for interrupt packets
usb: Call wakeup when data becomes available for all devices with int eps
add pc-1.4
Peter Maydell [Fri, 7 Dec 2012 15:39:13 +0000 (15:39 +0000)]
configure: Default to 'cc', not 'gcc'
Default to 'cc' as our compiler, rather than 'gcc'. We used to have
to insist on gcc when we still kept the CPU env in a fixed global
register, but this is no longer necessary and we will now compile OK
on clang as well as gcc. Using 'cc' should generally result in us
using the most standard and maintained system compiler for the
platform. (For instance on newer MacOS X 'gcc' exists but is an
elderly compiler provided mostly for legacy reasons, and 'cc'
(which is clang) is definitely the better choice.) On Linux there
will generally be no user-visible change since cc will be gcc.
This changeover necessitates a slight reworking of how we set the
'cc' variable, because GNU cross toolchains generally provide a
'${cross_prefix}gcc' but not a '${cross_prefix}cc'.
Peter Maydell [Fri, 12 Oct 2012 03:55:51 +0000 (03:55 +0000)]
linux-user: Merge pread/pwrite into pread64/pwrite64
The Linux syscalls underlying pread() and pwrite() take a 64 bit
offset on all architectures, even if some of them name the syscall
"pread/pwrite" rather than "pread64/pwrite64" for historical reasons.
So move the four QEMU target architectures (arm, i386, sparc,
unicore32) which were defining TARGET_NR_pread/pwrite to define
TARGET_NR_pread64/pwrite64 instead, and drop the TARGET_NR_pread/pwrite
implementation code completely.
(Based on examination of the kernel sources for the four architectures
this patch affects.)
Max Filippov [Wed, 5 Dec 2012 03:15:24 +0000 (07:15 +0400)]
target-xtensa: implement MISC SR
The Miscellaneous Special Registers Option provides zero to four scratch
registers within the processor readable and writable by RSR, WSR, and
XSR. These registers are privileged. They may be useful for some
application-specific exception and interrupt processing tasks in the
kernel. The MISC registers are undefined after reset.
See ISA, 4.7.3 for details.
Max Filippov [Wed, 5 Dec 2012 03:15:23 +0000 (07:15 +0400)]
target-xtensa: better control rsr/wsr/xsr access to SRs
There are read-only (DEBUGCAUSE, PRID) and write-only (INTCLEAR) SRs,
and INTERRUPT/INTSET SR allows rsr/wsr, but not xsr. Raise illeagal
opcode exception on illegal access to these SRs.
Max Filippov [Wed, 5 Dec 2012 03:15:22 +0000 (07:15 +0400)]
target-xtensa: restrict available SRs by enabled options
Beginning with the RA-2004.1 release, SR access instructions (rsr, wsr,
xsr) are associated with their corresponding SR and raise illegal opcode
exception in case the register is not configured for the core.
Max Filippov [Wed, 5 Dec 2012 03:15:21 +0000 (07:15 +0400)]
target-xtensa: implement CACHEATTR SR
In XEA1, the Options for Memory Protection and Translation and the
corresponding TLB management instructions are not available. Instead,
functionality similar to the Region Protection Option is available
through the cache attribute register. See ISA, A.2.14 for details.
Blue Swirl [Sat, 8 Dec 2012 17:50:57 +0000 (17:50 +0000)]
Merge branch 'master' of git.qemu-project.org:/pub/git/qemu
* 'master' of git.qemu-project.org:/pub/git/qemu:
target-mips: Fix incorrect shift for SHILO and SHILOV
target-mips: Fix incorrect code and test for INSV
xilinx_uartlite: Accept input after rx FIFO pop
xilinx_uartlite: suppress "cannot receive message"
xilinx_axienet: Implement R_IS behaviour
Eduardo Habkost [Tue, 4 Dec 2012 18:32:58 +0000 (16:32 -0200)]
finally kill cpudef config section support
The external CPU models were removed on QEMU 1.2, and the support for
the "cpudef" config sections was documented as deprecated, but the
actual removal of the config section was pending.
Now that QEMU 1.3 was released, we can finally kill the support for
cpudef config sections, and support only the built-in CPU models from
target-i386/cpu.c.
Peter Maydell [Wed, 31 Oct 2012 09:30:47 +0000 (09:30 +0000)]
HACKING: List areas where we may rely on impdef C behaviour
Add a section to HACKING saying which version of the C spec
we use and describing the bits of implementation defined C
compiler behaviour which C code in QEMU is allowed to rely on.
Brad Smith [Mon, 19 Nov 2012 09:22:12 +0000 (04:22 -0500)]
curses: Remove OpenBSD workaround
I removed the same sort of workaround for OpenBSD within the
configure script with commit 4dcc3f5876fa638d5c35bd47be3b717ea74cc2e7
but didn't bother to grep further to come across this same chunk
of code in the curses code itself. So the following diff removes
the same workaround chunk within the curses code.
Petar Jovanovic [Tue, 4 Dec 2012 23:29:10 +0000 (00:29 +0100)]
target-mips: Fix incorrect shift for SHILO and SHILOV
helper_shilo has not been shifting an accumulator value correctly for negative
values in 'shift' field. Minor optimization for shift=0 case.
This change also adds tests that will trigger issue and check for regressions.
The device return false from the can receive function when the FIFO is
full. This mean the device should check for buffered input whenever a byte is
popped from the FIFO.
The interrupt status register R_IS is the standard clear-on-write behaviour.
This was unimplemented and defaulting to updating the register to the written
value. Implemented clear-on-write.
Hans de Goede [Sat, 17 Nov 2012 11:47:18 +0000 (12:47 +0100)]
usb-tablet: Allow connecting to ehci
Our ehci code has is capable of significantly lowering the wakeup rate
for the hcd emulation while the device is idle. It is possible to add
similar code ot the uhci emulation, but that simply is not there atm,
and there is no reason why a (virtual) usb-tablet can not be a USB-2 device.
Making usb-hid devices connect to the emulated ehci controller instead
of the emulated uhci controller on vms which have both lowers the cpuload
for a fully idle vm from 20% to 2-3% (on my laptop).
An alternative implementation to using a property to select the tablet
type, would be simply making it a new device type, ie usb-tablet2, but the
downside of that is that this will require libvirt changes to be available
through libvirt at all, and then management tools changes to become the
default for new vms, where as using a property will automatically get
any pc-1.3 type vms the lower cpuload.
[ kraxel: adapt compat property for post-1.3 merge ]
Hans de Goede [Sat, 17 Nov 2012 11:47:17 +0000 (12:47 +0100)]
ehci: Lower timer freq when the periodic schedule is idle
Lower the timer freq if no iso schedule packets complete for 64 frames in
a row.
We can safely do this, without adding latency, because:
1) If there is isoc traffic this will never trigger
2) For async handled interrupt packets (only usb-host), the completion handler
will immediately schedule the frame_timer from a bh
3) All devices using NAK to signal no data for interrupt endpoints now use
wakeup, which will immediately schedule the frame_timer from a bh
The advantage of this is that when we only have interrupt packets in the
periodic schedule, async_stepdown can do its work and significantly lower
the frequency at which the frame_timer runs.
Hans de Goede [Sat, 17 Nov 2012 11:47:15 +0000 (12:47 +0100)]
usb: Don't allow USB_RET_ASYNC for interrupt packets
It is tempting to use USB_RET_ASYNC for interrupt packets, rather then the
current NAK + polling approach, but this causes issues for migration, as
an async completed packet will not getting written back to guest memory until
the next poll time, and if a migration happens in between it will get lost!
Make an exception for host devices, because:
1) host-linux actually uses async completion for interrupt endpoints
2) host devices don't migrate anyways
Ideally we would convert host-linux.c to handle (input) interrupt endpoints in
a buffered manner like it does for isoc endpoints, keeping multiple urbs
submitted to ensure the devices timing requirements are met, as well as making
its interrupt ep handling the same as other usb-devices.
Hans de Goede [Sat, 17 Nov 2012 11:47:14 +0000 (12:47 +0100)]
usb: Call wakeup when data becomes available for all devices with int eps
This is necessary for proper interaction with the xhci controller, and it
will allow other hcds to lower there frame timer while waiting for interrupt
data.
e1000: Discard packets that are too long if !SBP and !LPE
The e1000_receive function for the e1000 needs to discard packets longer than
1522 bytes if the SBP and LPE flags are disabled. The linux driver assumes
this behavior and allocates memory based on this assumption.
Anthony Liguori [Fri, 30 Nov 2012 16:24:24 +0000 (10:24 -0600)]
Merge remote-tracking branch 'kwolf/for-anthony' into staging
* kwolf/for-anthony:
coroutine-sigaltstack.c: Use stack_t, not struct sigaltstack
stream: fix ratelimit_set_speed
atapi: make change media detection for guests easier
Documentation: Update image format information
Documentation: Update block cache mode information
Anthony Liguori [Fri, 30 Nov 2012 16:20:21 +0000 (10:20 -0600)]
Merge remote-tracking branch 'kraxel/usb.73' into staging
* kraxel/usb.73:
ehci-sysbus: Attach DMA context.
usb: fail usbdevice_create() when there is no USB bus
usb: tag usb host adapters as not hotpluggable.
Most important change is that seabios sets the busmaster bit
in the pci config space for the lsi and esp scsi host adapters.
Since commit 1c380f9460522f32c8dd2577b2a53d518ec91c6d qemu
is strict here and disallows any dma access when the bit is
clear.
Paolo Bonzini [Thu, 29 Nov 2012 17:11:50 +0000 (18:11 +0100)]
multiboot: fix e801 memory map
The e801 memory sizes in the multiboot structures hard-code the available
low memory to 640. However, the value should not include the size of the
EBDA. Fill the value in the option ROM, getting the size of low memory
from the BIOS.
Peter Maydell [Sat, 10 Nov 2012 21:47:52 +0000 (21:47 +0000)]
coroutine-sigaltstack.c: Use stack_t, not struct sigaltstack
Use the POSIX-specified stack_t type as the argument to sigaltstack()
rather than the legacy struct sigaltstack. This allows us to compile
on MacOSX with --with-coroutine=sigaltstack.
Kevin Wolf [Wed, 21 Nov 2012 13:21:47 +0000 (14:21 +0100)]
Documentation: Update image format information
Document new and yet undocumented options and image formats. The
qemu-img man page contains information only for raw and qcow2 now and
references the HTML documentation for a more detailed description of
other formats.
Kevin Wolf [Wed, 21 Nov 2012 11:26:56 +0000 (12:26 +0100)]
Documentation: Update block cache mode information
Somehow we forgot to update this when cache=writeback became the
default. While changing the information on the default, also make the
description of all caches modes a bit more accurate.
Commit 0d8d7690850eb0cf2b2b60933cf47669a6b6f18f introduced
a regression in virtio-net performance because it looks
into the ring aggressively while we really only care
about a single packet worth of buffers.
Reported as bugzilla 1066055 in launchpad.
To fix, add parameters limiting lookahead, and
use in virtqueue_avail_bytes.
Anthony Liguori [Thu, 29 Nov 2012 13:46:23 +0000 (07:46 -0600)]
qdev: relax bus type check in qdev_device_add() (v2)
We are currently checking for an exact type match. Use QOM dynamic_cast to
check for a compatible type instead.
Cc: Konrad Frederic <[email protected]> Cc: Peter Maydell <[email protected]> Signed-off-by: Anthony Liguori <[email protected]>
---
v1 -> v2:
- also add cast to qbus_find_recursive (Peter)
- simplify by doing object_dynamic_cast instead of messing with classes
Bruce Rogers [Tue, 27 Nov 2012 20:11:25 +0000 (13:11 -0700)]
qapi: fix qapi_dealloc_type_size parameter type
The second parameter to qapi_dealloc_type_size should be a uint64_t *,
not a size_t *. This was causing our 32 bit x86 build to fail, since
warnings are treated as errors.
This was left as NULL on the initial merge due to debate on the mailing list on
how to handle DMA contexts for sysbus devices. Patch 9e11908f12f92e31ea94dc2a4c962c836cba9f2a was later merged to fix OHCI. This is the,
equivalent fix for sysbus EHCI.
Gerd Hoffmann [Thu, 22 Nov 2012 13:44:50 +0000 (14:44 +0100)]
usb: tag usb host adapters as not hotpluggable.
Hotplugging them simply doesn't work, so tag them accordingly to
avoid users trying and then crashing qemu.
For xhci there is nothing fundamental which prevents hotplug from
working, we'll "only" need a exit() function which cleans up
everything properly. That isn't for 1.3 though.
For ehci+uhci+ohci hotplug can't be supported until qemu gains the
capability to hotplug multifunction pci devices.
Yonit Halperin [Wed, 28 Nov 2012 15:08:22 +0000 (10:08 -0500)]
qxl: reload memslots after migration, when qxl is in UNDEFINED mode
The devram memslot stays active when qxl enters UNDEFINED mode (i.e, no
primary surface). If migration has occurred while the device is in
UNDEFINED stae, the memslots have to be reloaded at the destination.
Anthony Liguori [Wed, 28 Nov 2012 19:19:24 +0000 (13:19 -0600)]
Merge remote-tracking branch 'bonzini/scsi-next' into staging
* bonzini/scsi-next:
virtio-scsi: Fix subtle (guest) endian bug
virtio-scsi: Fix some endian bugs with virtio-scsi
iscsi: do not assume device is zero initialized
iscsi: fix deadlock during login
iscsi: fix segfault in url parsing
Paolo Bonzini [Wed, 28 Nov 2012 08:40:23 +0000 (09:40 +0100)]
build: compile translate.o with -fno-gcse option on GCC 4.6.x and 4.7.[012]
These versions of GCC require insane (>2GB) amounts of memory to compile
translate.o. As a countermeasure, disable the culprit optimization pass.
This should fix the buildbot failure for default_x86_64_fedora16. Anyway
this is a good thing to do because people will try to compile 1.3 with
less than 2GB of memory and complain.
David Gibson [Mon, 26 Nov 2012 01:33:52 +0000 (12:33 +1100)]
virtio-scsi: Fix subtle (guest) endian bug
The virtio-scsi config space is, by specification, in guest endian (which
is ill-defined, but there you go). In virtio_scsi_get_config() we set up
all the fields in there, using stl_raw(). Which is a problem for the
max_channel and max_target fields, which are 16-bit, not 32-bit. For
little-endian targets we get away with it by accident, since the first
two bytes will still be correct, and the extra two bytes written (with
zeroes) will be overwritten correctly by the next store.
But for big-endian guests, this means the max_target field ends up as zero,
which means the guest will only recognize a single disk on the virtio-scsi
bus. This patch fixes the problem.
David Gibson [Fri, 23 Nov 2012 05:08:44 +0000 (16:08 +1100)]
virtio-scsi: Fix some endian bugs with virtio-scsi
The virtio-scsi specification does not specify the correct endianness for
fields in the request structure. It's therefore best to assume that it is
"guest native" endian since that's the (stupid and poorly defined) norm in
virtio.
However, the qemu device for virtio-scsi has no byteswaps at all, and so
will break if the guest has different endianness from the host. This patch
fixes it by adding tswap() calls for the sense_len and resid fields in
the request structure. In theory status_qualifier needs swaps as well,
but that field is never actually touched. The tag field is a uint64_t, but
since its value is completely arbitrary, it might as well be uint8_t[8]
and so it does not need swapping.
Peter Lieven [Sat, 17 Nov 2012 13:37:39 +0000 (14:37 +0100)]
iscsi: fix deadlock during login
If the connection is interrupted before the first login is successfully
completed qemu-kvm is waiting forever in qemu_aio_wait().
This is fixed by performing an sync login to the target. If the
connection breaks after the first successful login errors are
handled internally by libiscsi.
Michal Privoznik [Wed, 28 Nov 2012 10:46:39 +0000 (11:46 +0100)]
nbd-server-add: Fix the default for 'writable'
The documentation to this monitor command tells, that 'writable'
argument is optional and defaults to false. However, the code sets
true as the default. But since some applications may already been
using this, it's safer to fix the code and not documentation which
would break those applications.
Alex Horn [Mon, 26 Nov 2012 16:32:54 +0000 (17:32 +0100)]
rtc: Only call rtc_set_cmos when Register B SET flag is disabled.
This bug occurs when the SET flag of Register B is enabled. When an RTC
data register (i.e. any of the ten time/calender CMOS bytes) is set, the
data is (as expected) correctly stored in the cmos_data array. However,
since the SET flag is enabled, the function rtc_set_time is not invoked.
As a result, the field base_rtc in RTCState remains uninitialized. This
causes a problem on subsequent writes which can end up overwriting data.
To see this, consider writing data to Register A after having written
data to any of the RTC data registers; the following figure illustrates
the call stack for the Register A write operation:
Paolo Bonzini [Tue, 27 Nov 2012 08:16:24 +0000 (09:16 +0100)]
virtio-rng: do not use g_assert_cmpint
g_assert_cmpint is not available on glib 2.12, which is the minimum
version required to build QEMU (we only require 2.16 to run tests,
since that is the first version including GTester). Do not use it
in hardware models, use a normal assertion instead.
This fixes the buildbot failure for default_x86_64_rhel5.
Paolo Bonzini [Tue, 27 Nov 2012 08:51:48 +0000 (09:51 +0100)]
tests: make threadpool cancellation test looser
The cancellation test is failing on the buildbots. While the failure
merits a little more investigation to understand what is going on,
the logs show that the failure is not impacting the coverage
provided by the test. Hence, loosen a bit the assertions in a
way that should let the test proceed and hopefully pass.
Amit Shah [Wed, 21 Nov 2012 05:51:19 +0000 (11:21 +0530)]
virtio-rng: remove extra request for entropy
If we got fewer bytes from the backend than requested, don't poke the
backend for more bytes; the guest will ask for more (or if the guest has
already asked for more, the backend knows about it via handle_input()).
Amit Shah [Wed, 21 Nov 2012 05:51:18 +0000 (11:21 +0530)]
virtio-rng: use virtqueue_get_avail_bytes, fix migration
Popping an elem from the vq just to find out its length causes problems
with save/load later on. Use the new virtqueue_get_avail_bytes()
function instead, saves us the complexity in the migration code, as well
as makes the migration endian-safe.
Andreas Färber [Sun, 25 Nov 2012 19:44:33 +0000 (20:44 +0100)]
i8259: Fix PIC_COMMON() macro
It used a wrong struct type name since its introduction in 8f04ee0882aec9fe91fb70f767edf5dacff59835 (isa: pic: convert to QEMU
Object Model), apparently it is unused so far.
Eduardo Habkost [Thu, 22 Nov 2012 15:31:03 +0000 (13:31 -0200)]
target-i386: cpu: add missing flags to Haswell CPU model
When adding the Haswell CPU model, I intended to make it a superset of the
features present on the SandyBridge model, but I have removed the SEP and
RDTSCP features from the feature list by mistake. This patch adds the missing
SEP and RDTSCP features (that are present on SandyBridge) to Haswell.
Peter Maydell [Thu, 22 Nov 2012 16:48:45 +0000 (16:48 +0000)]
vl.c: Fix broken -usb option
Commit 094b287f0b accidentally broke the "-usb" command line
option, so it would have no effect if the user had not specified
any machine options at that point. (the return value from
'qemu_opts_find(qemu_find_opts("machine"), 0);' is NULL if there
are no user specified options, so it is only to be used for
looking up an option, not when trying to set one.) Similarly,
would '-usbdevice' no longer cause USB to default to enabled.
Fix this regression by using the same style of code for forcing
the usb=on machine option that we use for other aliases such as
'-enable-kvm'.
Paolo Bonzini [Fri, 23 Nov 2012 08:47:15 +0000 (09:47 +0100)]
qdev: simplify (de)allocation of buses
All conditional deallocation can now be done with object_delete.
Remove the @qom_allocated and @glib_allocated fields; replace the latter
with a direct assignment of the @free function pointer.
Paolo Bonzini [Fri, 23 Nov 2012 08:47:14 +0000 (09:47 +0100)]
qom: make object_delete usable for statically-allocated objects
Store in the object the freeing function that will be used at deletion
time. This makes it possible to use object_delete on statically-allocated
(embedded) objects. Dually, it makes it possible to use object_unparent
and object_unref without leaking memory, when the lifetime of object
might extend until after the call to object_delete.
Paolo Bonzini [Fri, 23 Nov 2012 08:47:13 +0000 (09:47 +0100)]
qdev: move bus removal to object_unparent
Add an ObjectClass method that is done at object_unparent time. It
should remove any backlinks to the object in the composition tree,
so that object_delete will be able to drop the last reference and
free the object.
Paolo Bonzini [Fri, 23 Nov 2012 08:47:12 +0000 (09:47 +0100)]
qom: fix refcount of non-heap-allocated objects
The reference count for embedded objects is always one too low, because
object_initialize_with_type returns with zero references to the object.
This causes premature finalization of the object (or an assertion failure)
after calling object_ref to add an extra reference and object_unref to
remove it.
The fix is to move the initial object_ref call from object_new_with_type
to object_initialize_with_type.
projects / qemu.git / log