Hans de Goede [Wed, 31 Oct 2012 12:47:10 +0000 (13:47 +0100)]
combined-packet: Add a workaround for Linux usbfs + live migration
Older versions (anything but the latest) of Linux usbfs + libusb(x),
will submit larger (bulk) transfers split into multiple 16k submissions,
which means that rather then all tds getting linked into the queue in
one atomic operarion they get linked in a bunch at a time, which could
cause problems if:
1) We scan the queue while libusb is in the middle of submitting a split
bulk transfer
2) While this bulk transfer is pending we migrate to another host.
The problem is that after 2, the new host will rescan the queue and
combine the packets in one large transfer, where as 1) has caused the
original host to see them as 2 transfers. This patch fixes this by stopping
combinging if we detect a 16k transfer with its int_req flag set.
This should not adversely effect performance for other cases as:
1) Linux never sets the interrupt flag on packets other then the last
2) Windows does set the in_req flag on each td, but will submit large
transfers in 20k tds thus never triggering the check
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Hans de Goede [Wed, 31 Oct 2012 12:47:09 +0000 (13:47 +0100)]
usb: Add packet combining functions
Currently we only do pipelining for output endpoints, since to properly
support short-not-ok semantics we can only have one outstanding input
packet. Since the ehci and uhci controllers have a limited per td packet
size guests will split large input transfers to into multiple packets,
and since we don't pipeline these, this comes with a serious performance
penalty.
This patch adds helper functions to (re-)combine packets which belong to 1
transfer at the guest device-driver level into 1 large transger. This can be
used by (redirection) usb-devices to enable pipelining for input endpoints.
This patch will combine packets together until a transfer terminating packet
is encountered. A terminating packet is a packet which meets one or more of
the following conditions:
1) The packet size is *not* a multiple of the endpoint max packet size
2) The packet does *not* have its short-not-ok flag set
3) The packet has its interrupt-on-complete flag set
The short-not-ok flag of the combined packet is that of the terminating packet.
Multiple combined packets may be submitted to the device, if the combined
packets do not have their short-not-ok flag set, enabling true pipelining.
If a combined packet does have its short-not-ok flag set the queue will
wait with submitting further packets to the device until that packet has
completed.
Once enabled in the usb-redir and ehci code, this improves the speed (MB/s)
of a Linux guest reading from a USB mass storage device by a factor of
1.2 - 1.5.
And the main reason why I started working on this, when reading from a pl2303
USB<->serial converter, it combines the previous 4 packets submitted per
device-driver level read into 1 big read, reducing the number of packets / sec
by a factor 4, and it allows to have multiple reads outstanding. This allows
for much better latency tolerance without the pl2303's internal buffer
overflowing (which was happening at 115200 bps, without serial flow control).
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Hans de Goede [Wed, 31 Oct 2012 11:54:37 +0000 (12:54 +0100)]
uhci: Don't crash on device disconnect
My recent uhci cleanup series has introduced a regression, where
qemu sometimes crashes on a device disconnect. The problem is that
the uhci code never checked for a device not / no longer existing, instead
it was relying on usb_handle_packet accepting a NULL device.
But since we now pass usb_handle_packet q->ep->dev, rather then just
a local dev variable, we crash as q->ep == NULL due to the device no longer
existing.
This patch fixes this. Note that this patch also improves over
the old behavior were we would:
1) create a queue for the device
2) create an async for the packet
3) have usb_handle_packet fail
4) destroy the async
5) wait for the queue to be idle for 32 frames
6) destroy the queue
Which was rather sub-optimal.
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Guard against re-definition of EHCI_DEBUG. Allows for turning on of debug info
from configure (using --qemu-extra-cflags="-DEHCI_DEBUG=1") rather than source
code hacking.
Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Seperate the PCI stuff from the EHCI components. Extracted the PCIDevice
out into a new wrapper struct to make EHCIState non-PCI-specific. Seperated
tho non PCI init component out into a seperate "common" init function.
Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Pull the DMAContext for the PCI DMA out at device init time and put it into
the device state. Use dma_memory_read/write() instead of pci specific versions.
Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Gerd Hoffmann [Fri, 26 Oct 2012 08:10:54 +0000 (10:10 +0200)]
xhci: add {get,set}_field macros & enum for pls
Add {get,set}_field macros (simliar to ehci) to read and update
some bits of a word. Put them into use for updating pls (port
link state) values. Also add a enum for pls values.
Aurelien Jarno [Tue, 9 Oct 2012 19:53:21 +0000 (21:53 +0200)]
target-mips: don't flush extra TLB on permissions upgrade
If the guest uses a TLBWI instruction for upgrading permissions, we
don't need to flush the extra TLBs. This improve boot time performance
by about 10%.
Aurelien Jarno [Tue, 9 Oct 2012 19:53:21 +0000 (21:53 +0200)]
target-mips: use deposit instead of hardcoded version
Use the deposit op instead of and hardcoded bit field insertion. It
allows the host to emit the corresponding instruction if available.
At the same time remove the (lsb > msb) test. The MIPS64R2 instruction
set manual says "Because of the instruction format, lsb can never be
greater than msb, so there is no UNPREDICATABLE case for this
instruction."
(Bug reported as LP:1071149.) Cc: Никита Канунников <n.kanunnikov@sbtcom.ru> Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Aurelien Jarno [Tue, 9 Oct 2012 19:53:21 +0000 (21:53 +0200)]
target-mips: optimize ddiv/ddivu/div/divu with movcond
The result of a division by 0, or a division of INT_MIN by -1 in the
signed case, is unpredictable. Just replace 0 by 1 in that case so that
it doesn't trigger a floating point exception on the host.
Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Aurelien Jarno [Tue, 9 Oct 2012 19:53:21 +0000 (21:53 +0200)]
target-mips: don't use local temps for store conditional
Store conditional operations only need local temps in user mode. Fix
the code to use temp local only in user mode, this spares two memory
stores in system mode.
At the same time remove a wrong a wrong copied & pasted comment,
store operations don't have a register destination.
Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Aurelien Jarno [Tue, 9 Oct 2012 19:53:21 +0000 (21:53 +0200)]
target-mips: implement unaligned loads using TCG
Load/store from helpers should be avoided as they are quite
inefficient. Rewrite unaligned loads instructions using TCG and
aligned loads. The number of actual loads operations to implement
an unaligned load instruction is reduced from up to 8 to 1.
Note: As we can't rely on shift by 32 or 64 undefined behaviour,
the code loads already shift by one constants.
Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
load/store microMIPS helpers are reinventing the wheel. Call do_lw,
do_ll, do_sw and do_sl instead of using a macro calling the cpu_*
load/store functions.
Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Aurelien Jarno [Tue, 23 Oct 2012 08:12:00 +0000 (10:12 +0200)]
target-mips: use softfloat constants when possible
softfloat already has a few constants defined, use them instead of
redefining them in target-mips.
Rename FLOAT_SNAN32 and FLOAT_SNAN64 to FP_TO_INT32_OVERFLOW and
FP_TO_INT64_OVERFLOW as even if they have the same value, they are
technically different (and defined differently in the MIPS ISA).
Remove the unused constants.
Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Aurelien Jarno [Tue, 9 Oct 2012 19:53:20 +0000 (21:53 +0200)]
target-mips: fix FPU exceptions
For each FPU instruction that can trigger an FPU exception, to call
call update_fcr31() after.
Remove the manual NaN assignment in case of float to float operation, as
softfloat is already taking care of that. However for float to int
operation, the value has to be changed to the MIPS one. In the cvtpw_ps
case, the two registers have to be handled separately to guarantee
a correct final value in both registers.
Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Aurelien Jarno [Sun, 28 Oct 2012 17:08:27 +0000 (18:08 +0100)]
target-mips: keep softfloat exception set to 0 between instructions
Instead of clearing the softfloat exception flags before each floating
point instruction, reset them to 0 in update_fcr31() when an exception
is detected.
Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Aurelien Jarno [Tue, 9 Oct 2012 19:53:20 +0000 (21:53 +0200)]
target-mips: use the softfloat floatXX_muladd functions
Use the new softfloat floatXX_muladd() functions to implement the madd,
msub, nmadd and nmsub instructions. At the same time replace the name of
the helpers by the name of the instruction, as the only reason for the
previous names was to keep the macros simple.
Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Jan Kiszka [Fri, 12 Oct 2012 07:52:49 +0000 (09:52 +0200)]
chardev: Use timer instead of bottom-half to postpone open event
As the block layer may decide to flush bottom-halfs while the machine is
still initializing (e.g. to read geometry data from the disk), our
postponed open event may be processed before the last frontend
registered with a muxed chardev.
Until the semantics of BHs have been clarified, use an expired timer to
achieve the same effect (suggested by Paolo Bonzini). This requires to
perform the alarm timer initialization earlier as otherwise timer
subsystem can be used before being ready.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Aurelien Jarno [Wed, 31 Oct 2012 21:14:46 +0000 (22:14 +0100)]
target-mips: remove #if defined(TARGET_MIPS64) in opcode enums
All switch() decoding instruction have a default entry, so it is possible
to have unused enum entries. Remove conditional definitions of MIPS64
opcode enums, as it only makes the code less readable.
Peter Maydell [Sat, 27 Oct 2012 21:19:07 +0000 (22:19 +0100)]
configure: use -Wwombat to test whether gcc recognizes -Wno-wombat
gcc will silently accept unrecognized -Wno-wombat warning suppression
options (it only mentions them if it has to print a compiler warning
for some other reason). Since we already run a check for whether gcc
recognizes the warning options we use, we can easily make this use
the positive sense of the option when checking for support for the
suppression option. This doesn't have any effect except that it avoids
gcc emitting extra messages about unrecognized command line options
when it is printing other warning messages.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Stefan Weil <sw@weilnetz.de> Reviewed-by: Igor Mitsyanko <i.mitsyanko@samsung.com> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
target-sparc: Revert setting cpu_dst to gen_dest_gpr
There is some read-after-write error within the OP=2 insns which
prevents setting cpu_dst to the real output register. Until this
is found and fixed, always write to a temporary first.
Cc: Blue Swirl <blauwirbel@gmail.com> Cc: Aurelien Jarno <aurelien@aurel32.net> Signed-off-by: Richard Henderson <rth@twiddle.net> Tested-by: Aurelien Jarno <aurelien@aurel32.net> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Blue Swirl [Tue, 30 Oct 2012 18:35:18 +0000 (18:35 +0000)]
Merge branch 'qspi.2' of git://developer.petalogix.com/public/qemu
* 'qspi.2' of git://developer.petalogix.com/public/qemu:
xilinx_zynq: added QSPI controller
xilinx_spips: Generalised to model QSPI
m25p80: Support for Quad SPI
Aurelien Jarno [Mon, 29 Oct 2012 23:35:43 +0000 (00:35 +0100)]
Merge branch 's390-for-upstream' of git://repo.or.cz/qemu/agraf
* 's390-for-upstream' of git://repo.or.cz/qemu/agraf:
s390: sclp ascii console support
s390: sclp signal quiesce support
s390: sclp event support
s390: sclp base support
s390: use sync regs for register transfer
s390/kvm_stat: correct sys_perf_event_open syscall number
s390x: fix -initrd in virtio machine
Aurelien Jarno [Fri, 19 Oct 2012 21:48:13 +0000 (23:48 +0200)]
tcg/mips: use MUL instead of MULT on MIPS32 and above
MIPS32 and later instruction sets have a multiplication instruction
directly operating on GPRs. It only produces a 32-bit result but
it is exactly what is needed by QEMU.
Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Avi Kivity [Mon, 29 Oct 2012 16:22:36 +0000 (18:22 +0200)]
memory: fix rendering of a region obscured by another
The memory core drops regions that are hidden by another region (for example,
during BAR sizing), but it doesn't do so correctly if the lower address of the
existing range is below the lower address of the new range.
Example (qemu-system-mips -M malta -kernel vmlinux-2.6.32-5-4kc-malta
-append "console=ttyS0" -nographic -vga cirrus):
Anthony Liguori [Mon, 29 Oct 2012 19:31:47 +0000 (14:31 -0500)]
Merge remote-tracking branch 'mst/tags/for_anthony' into staging
virtio,pci infrastructure
This includes infrastructure patches that don't do much by themselves
but should help vfio and q35 make progress.
Also included is rework of virtio-net to use iovec APIs
for vector access - helpful to make it more secure
and in preparation for a new feature that will allow
arbitrary s/g layout for guests.
Also included is a pci bridge bugfix by Avi.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
* mst/tags/for_anthony: (25 commits)
pci: avoid destroying bridge address space windows in a transaction
virtio-net: enable mrg buf header in tap on linux
virtio-net: test peer header support at init time
virtio-net: minor code simplification
virtio-net: simplify rx code
virtio-net: switch tx to safe iov functions
virtio-net: first s/g is always at start of buf
virtio-net: refactor receive_hdr
virtio-net: use safe iov operations for rx
virtio-net: avoid sg copy
iov: add iov_cpy
virtio-net: track host/guest header length
pcie: Convert PCIExpressHost to use the QOM.
pcie: pass pcie window size to pcie_host_mmcfg_update()
pci: Add class 0xc05 as 'SMBus'
pci: introduce pci_swizzle_map_irq_fn() for standardized interrupt pin swizzle
pci_ids: add intel 82801BA pci-to-pci bridge id
pci: pci capability must be in PCI space
pci: make each capability DWORD aligned
qemu: enable PV EOI for qemu 1.3
...
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Heinz Graalfs [Mon, 29 Oct 2012 02:13:25 +0000 (02:13 +0000)]
s390: sclp ascii console support
This code adds console support by implementing SCLP's ASCII Console
Data event. This is the same console as LPARs ASCII console or z/VMs
sysascii.
The console can be specified manually with something like
-chardev stdio,id=charconsole0 -device sclpconsole,chardev=charconsole0,id=console0
Newer kernels will autodetect that console and prefer that over virtio
console.
When data is received from the character layer it creates a service
interrupt to trigger a Read Event Data command from the guest that will
pick up the received character byte-stream.
When characters are echo'ed by the linux guest a Write Event Data occurs
which is forwarded by the Event Facility to the console that supports
a corresponding mask value.
Console resizing is not supported.
The character layer byte-stream is buffered using a fixed size iov
buffer.
Signed-off-by: Heinz Graalfs <graalfs@linux.vnet.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com> Signed-off-by: Alexander Graf <agraf@suse.de>
Heinz Graalfs [Mon, 29 Oct 2012 02:13:24 +0000 (02:13 +0000)]
s390: sclp signal quiesce support
This implements the sclp signal quiesce event via the SCLP Event
Facility.
This allows to gracefully shutdown a guest by using system_powerdown
notifiers. It creates a service interrupt that will trigger a
Read Event Data command from the guest. This code will then add an
event that is interpreted by linux guests as ctrl-alt-del.
Signed-off-by: Heinz Graalfs <graalfs@linux.vnet.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com> Signed-off-by: Alexander Graf <agraf@suse.de>
Heinz Graalfs [Mon, 29 Oct 2012 02:13:23 +0000 (02:13 +0000)]
s390: sclp event support
Several SCLP features are considered to be events. Those events don't
provide SCLP commands on their own, instead they are all based on
Read Event Data, Write Event Data, Write Event Mask and the service
interrupt. Follow-on patches will provide SCLP's Signal Quiesce (via
system_powerdown) and the ASCII console.
Further down the road the sclp line mode console and configuration
change events (e.g. cpu hotplug) can be implemented.
Signed-off-by: Heinz Graalfs <graalfs@linux.vnet.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com> Signed-off-by: Alexander Graf <agraf@suse.de>
Heinz Graalfs [Mon, 29 Oct 2012 02:13:22 +0000 (02:13 +0000)]
s390: sclp base support
This adds a more generic infrastructure for handling Service-Call
requests on s390. Currently we only support a small subset of Read
SCP Info directly in target-s390x. This patch provides the base
infrastructure for supporting more commands and moves Read SCP
Info.
In the future we could add additional commands for hotplug, call
home and event handling.
Signed-off-by: Heinz Graalfs <graalfs@linux.vnet.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com> Signed-off-by: Alexander Graf <agraf@suse.de>
Newer kernels provide the guest registers in kvm_run. Lets use
those if available (i.e. the capability is set). This avoids
ioctls on cpu_synchronize_state making intercepts faster.
In addition, we have now the prefix register, the access registers
the control registers up to date. This helps in certain cases,
e.g. for resolving kernel module addresses with gdb on a guest.
On return, we update the registers according to the level statement,
i.e. we put all registers for KVM_PUT_FULL_STATE and _RESET_STATE.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com> Signed-off-by: Alexander Graf <agraf@suse.de>
Heinz Graalfs [Mon, 29 Oct 2012 02:13:20 +0000 (02:13 +0000)]
s390/kvm_stat: correct sys_perf_event_open syscall number
Correct sys_perf_event_open syscall number for s390 architecture
- the hardcoded syscall number 298 is for x86 but should
be different for other architectures.
In case we figure out via /proc/cpuinfo that we are running
on s390 the appropriate syscall number is used from map
syscall_numbers; other architectures can extend this.
Alexander Graf [Wed, 19 Sep 2012 15:24:46 +0000 (17:24 +0200)]
s390x: fix -initrd in virtio machine
When using -initrd in the virtio machine, we need to indicate the initrd
start and size inside the kernel image. These parameters need to be stored
in native endianness.
Signed-off-by: Alexander Graf <agraf@suse.de> Acked-by: Richard Henderson <rth@twiddle.net> Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Modern linux supports arbitrary header size,
which makes it possible to pass mrg buf header
to tap directly without iovec mangling.
Use this capability when it is there.
This removes the need to deal with it in
vhost-net as we do now.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
There's no reason to query header support at random
times: at load or feature query.
Driver also might not query functions.
Cleaner to do it at device init.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Remove code duplication using guest header length that we track.
Drop specific layout requirement for rx buffers: things work
using generic iovec functions in any case.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Avoid magling iov manually: use safe iov operations
for processing packets incoming to guest.
This also removes the requirement for virtio header to
fit the first s/g entry exactly.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Avoid tweaking iovec during receive. This removes
the need to copy the vector.
Note: we currently have an evil cast in work_around_broken_dhclient
and unfortunately this patch does not fix it - just
pushes the evil cast to another place.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Baron [Fri, 19 Oct 2012 20:43:33 +0000 (16:43 -0400)]
pcie: Convert PCIExpressHost to use the QOM.
Let's use PCIExpressHost with QOM.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Acked-by: Andreas Färber <afaerber@suse.de> Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jan Kiszka [Fri, 19 Oct 2012 20:43:31 +0000 (16:43 -0400)]
pci: Add class 0xc05 as 'SMBus'
[jbaron@redhat.com: add PCI_CLASS_SERIAL_SMBUS definition] Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Isaku Yamahata [Fri, 19 Oct 2012 20:43:28 +0000 (16:43 -0400)]
pci: introduce pci_swizzle_map_irq_fn() for standardized interrupt pin swizzle
Introduce pci_swizzle_map_irq_fn() for interrupt pin swizzle which is
standardized. PCI bridge swizzle is common logic, by introducing
this function duplicated swizzle logic will be avoided later.
[jbaron@redhat.com: drop opaque argument] Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp> Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp> Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Isaku Yamahata [Fri, 19 Oct 2012 20:43:28 +0000 (16:43 -0400)]
pci: pci capability must be in PCI space
pci capability must be in PCI space.
It can't lay in PCIe extended config space.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp> Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
projects / qemu-upstream-4.4-testing.git / log