sdk/emulator/qemu.git
9 years agofw_cfg_mem: max access size and region size are the same for data register
Laszlo Ersek [Mon, 22 Dec 2014 12:11:37 +0000 (13:11 +0100)]
fw_cfg_mem: max access size and region size are the same for data register

Make it clear that the maximum access size to the MMIO data register
determines the full size of the memory region.

Currently the max access size is 1.

This patch doesn't change behavior.

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 1419250305-31062-4-git-send-email-pbonzini@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agofw_cfg: move boards to fw_cfg_init_io() / fw_cfg_init_mem()
Laszlo Ersek [Mon, 22 Dec 2014 12:11:36 +0000 (13:11 +0100)]
fw_cfg: move boards to fw_cfg_init_io() / fw_cfg_init_mem()

This allows us to drop the fw_cfg_init() shim and to enforce the possible
mappings at compile time.

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 1419250305-31062-3-git-send-email-pbonzini@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agofw_cfg: hard separation between the MMIO and I/O port mappings
Laszlo Ersek [Mon, 22 Dec 2014 12:11:35 +0000 (13:11 +0100)]
fw_cfg: hard separation between the MMIO and I/O port mappings

We are going to introduce a wide data register for fw_cfg, but only for
the MMIO mapped device. The wide data register will also require the
tightening of endiannesses.

However we don't want to touch the I/O port mapped fw_cfg device at all.

Currently QEMU provides a single fw_cfg device type that can handle both
I/O port and MMIO mapping. This flexibility is not actually exploited by
any board in the tree, but it renders restricting the above changes to
MMIO very hard.

Therefore, let's derive two classes from TYPE_FW_CFG: TYPE_FW_CFG_IO and
TYPE_FW_CFG_MEM.

TYPE_FW_CFG_IO incorporates the base I/O port and the related combined
MemoryRegion. (NB: all boards in the tree that use the I/O port mapped
flavor opt for the combined mapping; that is, when the data port overlays
the high address byte of the selector port. Therefore we can drop the
capability to map those I/O ports separately.)

TYPE_FW_CFG_MEM incorporates the base addresses for the MMIO selector and
data registers, and their respective MemoryRegions.

The "realize" and "props" class members are specific to each new derived
class, and become unused for the base class. The base class retains the
"reset" member and the "vmsd" member, because the reset functionality and
the set of migrated data are not specific to the mapping.

The new functions fw_cfg_init_io() and fw_cfg_init_mem() expose the
possible mappings in separation. For now fw_cfg_init() is retained as a
compatibility shim that enforces the above assumptions.

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 1419250305-31062-2-git-send-email-pbonzini@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: add cpu feature EL3 to CPUs with Security Extensions
Fabian Aggeler [Mon, 15 Dec 2014 23:09:52 +0000 (17:09 -0600)]
target-arm: add cpu feature EL3 to CPUs with Security Extensions

Set ARM_FEATURE_EL3 feature for CPUs that implement Security Extensions.

Signed-off-by: Fabian Aggeler <aggelerf@ethz.ch>
Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-16-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Disable EL3 on unsupported machines
Greg Bellows [Mon, 15 Dec 2014 23:09:51 +0000 (17:09 -0600)]
target-arm: Disable EL3 on unsupported machines

Disables the CPU ARM_FEATURE_EL3 featuere on machine models that can be
configured to use Cortex-A9, Cortex-A15, and ARM1176 but don't officially
support EL3.  This preserves backwards compatibility.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-15-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Breakout integratorcp and versatilepb cpu init
Greg Bellows [Mon, 15 Dec 2014 23:09:50 +0000 (17:09 -0600)]
target-arm: Breakout integratorcp and versatilepb cpu init

This commit changes the integratorcp and versatilepb CPU initialization from
using the generic ARM cpu_arm_init function to doing it inline.  This is
necessary in order to allow CPU configuration changes to occur between CPU
instance initialization and realization.  Specifically, this change is in
preparation for disabling CPU EL3 support.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-14-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Set CPU has_el3 prop during virt init
Greg Bellows [Mon, 15 Dec 2014 23:09:49 +0000 (17:09 -0600)]
target-arm: Set CPU has_el3 prop during virt init

Adds setting of the CPU has_el3 property based on the virt machine
secure state property during initialization.  This enables/disables EL3
state during start-up.  Changes include adding an additional secure state
boolean during virt CPU initialization.  Also disables the ARM secure boot
by default.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Message-id: 1418684992-8996-13-git-send-email-greg.bellows@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Enable CPU has_el3 prop during VE init
Greg Bellows [Mon, 15 Dec 2014 23:09:48 +0000 (17:09 -0600)]
target-arm: Enable CPU has_el3 prop during VE init

Adds setting of the CPU has_el3 property based on the vexpress machine
secure state property during initialization.  This enables/disables EL3
state during start-up.  Changes include adding an additional secure state
boolean during vexpress CPU initialization.  Also enables the ARM secure boot
by default.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Message-id: 1418684992-8996-12-git-send-email-greg.bellows@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Add arm_boot_info secure_boot control
Greg Bellows [Mon, 15 Dec 2014 23:09:47 +0000 (17:09 -0600)]
target-arm: Add arm_boot_info secure_boot control

Adds the secure_boot boolean field to the arm_boot_info descriptor.  This
fields is used to indicate whether Linux should boot into secure or non-secure
state if the ARM EL3 feature is enabled.  The default is to leave the CPU in an
unaltered reset state.  On EL3 enabled systems, the reset state is secure and
can be overridden by setting the added field to false.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-11-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Add ARMCPU secure property
Greg Bellows [Mon, 15 Dec 2014 23:09:46 +0000 (17:09 -0600)]
target-arm: Add ARMCPU secure property

Added a "has_el3" state property to the ARMCPU descriptor.  This property
indicates whether the ARMCPU has security extensions enabled (EL3) or not.
By default it is disabled at this time.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-10-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Add feature unset function
Greg Bellows [Mon, 15 Dec 2014 23:09:45 +0000 (17:09 -0600)]
target-arm: Add feature unset function

Add an unset_feature() function to compliment the set_feature() function.  This
will be used to disable functions after they have been enabled during
initialization.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-9-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Add virt machine secure property
Greg Bellows [Mon, 15 Dec 2014 23:09:44 +0000 (17:09 -0600)]
target-arm: Add virt machine secure property

Add "secure" virt machine specific property to allow override of the
default secure state configuration.  By default, when using the QEMU
-kernel command line argument, virt machines boot into NS/SVC.  When using
the QEMU -bios command line argument, virt machines boot into S/SVC.

The secure state can be changed from the default specifying the secure
state as a machine property.  For example, the below command line would disable
security extensions on a -kernel Linux boot:

    aarch64-softmmu/qemu-system-aarch64
        -machine type=virt,secure=off
        -kernel ...

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-8-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Add virt class and machine types
Greg Bellows [Mon, 15 Dec 2014 23:09:43 +0000 (17:09 -0600)]
target-arm: Add virt class and machine types

Switch virt qemu machine support to use the newer object type, class, and
instance model.  Added virt TypeInfo with static registration along with virt
specific class and machine structs.  Also added virt class initialization
method.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-7-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Change vexpress daughterboard init arg
Greg Bellows [Mon, 15 Dec 2014 23:09:42 +0000 (17:09 -0600)]
target-arm: Change vexpress daughterboard init arg

Change the Vexpress daughterboard initialization method to take a vexpress
machine state pointer instead of the daughterboard struct pointer.  The machine
state now contains the daughterboard pointer.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-6-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Add vexpress machine secure property
Greg Bellows [Mon, 15 Dec 2014 23:09:41 +0000 (17:09 -0600)]
target-arm: Add vexpress machine secure property

Add "secure" Vexpress machine specific property to allow override of the
default secure state configuration.  By default, when using the QEMU
-kernel command line argument, Vexpress machines boot into NS/SVC.  When using
the QEMU -bios command line argument, Vexpress machines boot into S/SVC.

The secure state can be changed from the default specifying the secure
state as a machine property.  For example, the below command line would disable
security extensions on a -kernel Linux boot:

    aarch64-softmmu/qemu-system-aarch64
        -machine type=vexpress-a15,secure=off
        -kernel ...

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-5-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Switch to common vexpress machine init
Greg Bellows [Mon, 15 Dec 2014 23:09:40 +0000 (17:09 -0600)]
target-arm: Switch to common vexpress machine init

Switched the Vexpress machine initialization to use the common function with
the machine pointer to board info.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-4-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Add vexpress a9 & a15 machine objects
Greg Bellows [Mon, 15 Dec 2014 23:09:39 +0000 (17:09 -0600)]
target-arm: Add vexpress a9 & a15 machine objects

Add Vexpress machine objects for the the Cortex A9 & A15 variants.  The older
style QEMUMachine types were replaced with dedicated TypeInfo objects. The new
objects include dedicated class init functions that currently ustilze dedicated
machine init methods.  The previous qemu_register_machine calls were replaced
with the newer type_register_status calls.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-3-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Add vexpress class and machine types
Greg Bellows [Mon, 15 Dec 2014 23:09:38 +0000 (17:09 -0600)]
target-arm: Add vexpress class and machine types

Adds base Vexpress class and machine objects and infrastructure.  This is in
preparation for switching to the full QEMU object model.  The base vexpress
infrastructure is intended to handle common vexpress details.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418684992-8996-2-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agovl.c: add HMP help to machine
Marcel Apfelbaum [Tue, 16 Dec 2014 16:58:06 +0000 (16:58 +0000)]
vl.c: add HMP help to machine

The help is based on the actual machine properties
exposing only the relevant options.

Signed-off-by: Marcel Apfelbaum <marcel.a@redhat.com>
Reviewed-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Greg Bellows <greg.bellows@linaro.org>
Message-id: 1418217570-15517-4-git-send-email-marcel.a@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agovl.c: simplified machine_set_property
Marcel Apfelbaum [Tue, 16 Dec 2014 16:58:05 +0000 (16:58 +0000)]
vl.c: simplified machine_set_property

Refactored the code to re-use object_property_parse.

Signed-off-by: Marcel Apfelbaum <marcel.a@redhat.com>
Reviewed-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Greg Bellows <greg.bellows@linaro.org>
Message-id: 1418217570-15517-3-git-send-email-marcel.a@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agomachine: remove qemu_machine_opts global list
Marcel Apfelbaum [Tue, 16 Dec 2014 16:58:05 +0000 (16:58 +0000)]
machine: remove qemu_machine_opts global list

QEMU has support for options per machine, keeping
a global list of options is no longer necessary.

Signed-off-by: Marcel Apfelbaum <marcel.a@redhat.com>
Reviewed-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Greg Bellows <greg.bellows@linaro.org>
Message-id: 1418217570-15517-2-git-send-email-marcel.a@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Merge EL3 CP15 register lists
Greg Bellows [Tue, 16 Dec 2014 16:58:05 +0000 (16:58 +0000)]
target-arm: Merge EL3 CP15 register lists

Merge of the v8_el2_cp_reginfo and el3_cp_reginfo ARMCPRegInfo lists.
Previously, some EL3 registers were restricted to the ARMv8 list under the
impression that they were not needed on ARMv7.  However, this is not the case
as the ARMv7/32-bit variants rely on the ARMv8/64-bit variants to handle
migration and reset.  For this reason they must always exist.

Signed-off-by: Greg Bellows <greg.bellows@linaro.org>
Message-id: 1418406450-14961-1-git-send-email-greg.bellows@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoaudio: Don't free hw resources until after hw backend is stopped
Peter Maydell [Tue, 16 Dec 2014 16:58:05 +0000 (16:58 +0000)]
audio: Don't free hw resources until after hw backend is stopped

When stopping an audio voice, call the audio backend's fini
method before calling audio_pcm_hw_free_resources_ rather than
afterwards. This allows backends which use helper threads (like
pulseaudio) to terminate those threads before the conv_buf or
mix_buf are freed and avoids race conditions where the helper
may access a NULL pointer or freed memory.

Cc: qemu-stable@nongnu.org
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1418406239-9838-1-git-send-email-peter.maydell@linaro.org

9 years agoMerge remote-tracking branch 'remotes/gonglei/tags/bootdevice-next-20141222' into...
Peter Maydell [Mon, 22 Dec 2014 14:52:52 +0000 (14:52 +0000)]
Merge remote-tracking branch 'remotes/gonglei/tags/bootdevice-next-20141222' into staging

bootdevice: Refactor and improvement

# gpg: Signature made Mon 22 Dec 2014 06:44:08 GMT using RSA key ID DDE30FBB
# gpg: Can't check signature: public key not found

* remotes/gonglei/tags/bootdevice-next-20141222:
  bootdevice: add Error **errp argument for QEMUBootSetHandler
  bootdevice: add validate check for qemu_boot_set()
  bootdevice: add Error **errp argument for qemu_boot_set()
  bootdevice: add Error **errp argument for validate_bootdevices()
  bootdevice: move code about bootorder from vl.c to bootdevice.c

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/bkoppelmann/tags/pull-tricore-20141221' into...
Peter Maydell [Mon, 22 Dec 2014 13:15:52 +0000 (13:15 +0000)]
Merge remote-tracking branch 'remotes/bkoppelmann/tags/pull-tricore-20141221' into staging

TriCore RR, RR1 insn added and several bug fixes

# gpg: Signature made Sun 21 Dec 2014 18:39:11 GMT using RSA key ID 6B69CA14
# gpg: Good signature from "Bastian Koppelmann <kbastian@mail.uni-paderborn.de>"

* remotes/bkoppelmann/tags/pull-tricore-20141221:
  target-tricore: Add instructions of RR1 opcode format, that have 0xb3 as first opcode
  target-tricore: Fix MFCR/MTCR insn and B format offset.
  target-tricore: Add missing 1.6 insn of BOL opcode format
  target-tricore: Add instructions of RR opcode format, that have 0x4b as the first opcode
  target-tricore: Add instructions of RR opcode format, that have 0x1 as the first opcode
  target-tricore: Add instructions of RR opcode format, that have 0xf as the first opcode
  target-tricore: Add instructions of RR opcode format, that have 0xb as the first opcode
  target-tricore: Change SSOV/SUOV makro name to SSOV32/SUOV32
  target-tricore: Fix mask handling JNZ.T being 7 bit long
  target-tricore: pretty-print register dump and show more status registers
  target-tricore: add missing 64-bit MOV in RLC format
  target-tricore: typo in BOL format
  target-tricore: fix offset masking in BOL format

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agobootdevice: add Error **errp argument for QEMUBootSetHandler
Gonglei [Wed, 3 Dec 2014 19:04:02 +0000 (19:04 +0000)]
bootdevice: add Error **errp argument for QEMUBootSetHandler

It will be useful for checking when we change traditional
boot order dynamically and propagate error message
to the monitor.
For x86 architecture, we pass &local_err to set_boot_dev()
when vm startup in pc_coms_init().

Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Alexander Graf <agraf@suse.de>
Cc: Blue Swirl <blauwirbel@gmail.com>
Cc: qemu-ppc@nongnu.org
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
9 years agobootdevice: add validate check for qemu_boot_set()
Gonglei [Wed, 3 Dec 2014 18:25:46 +0000 (18:25 +0000)]
bootdevice: add validate check for qemu_boot_set()

Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
9 years agobootdevice: add Error **errp argument for qemu_boot_set()
Gonglei [Wed, 3 Dec 2014 18:20:58 +0000 (18:20 +0000)]
bootdevice: add Error **errp argument for qemu_boot_set()

It will be useful for checking when we change traditional
boot order dynamically and propagate error message
to the monitor.

Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
9 years agobootdevice: add Error **errp argument for validate_bootdevices()
Gonglei [Wed, 3 Dec 2014 17:11:39 +0000 (17:11 +0000)]
bootdevice: add Error **errp argument for validate_bootdevices()

It will be useful for checking when we change traditional
boot order dynamically and propagate error message
to the monitor.

Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
9 years agobootdevice: move code about bootorder from vl.c to bootdevice.c
Gonglei [Wed, 3 Dec 2014 16:49:46 +0000 (16:49 +0000)]
bootdevice: move code about bootorder from vl.c to bootdevice.c

First, we can downsize vl.c, make it simpler by
little and little. Second, I can maintain those code
and make some improvement.

Cc: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
9 years agoMerge remote-tracking branch 'remotes/kraxel/tags/pull-sdl-20141219-1' into staging
Peter Maydell [Sun, 21 Dec 2014 23:17:00 +0000 (23:17 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-sdl-20141219-1' into staging

sdl2: fixes, cleanups and opengl preparation.

# gpg: Signature made Fri 19 Dec 2014 09:06:07 GMT using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"

* remotes/kraxel/tags/pull-sdl-20141219-1:
  sdl2: Work around SDL2 SDL_ShowWindow() bug
  sdl2: Use correct sdl2_console for window events
  sdl2: move sdl2_2d_refresh to sdl2-2d.c
  sdl2: factor out sdl2_poll_events
  sdl2: add+use sdl2_2d_redraw function.
  sdl2: move sdl_switch to sdl2-2d.c
  sdl2: overhaul window size handling
  sdl2: move sdl_update to new sdl2-2d.c
  sdl2: turn on keyboard grabs
  sdl2: move keyboard input code to new sdl2-input.c
  sdl2: rename sdl2_state to sdl2_console, move to header file
  sdl: move version logic from source code to makefile

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-tricore: Add instructions of RR1 opcode format, that have 0xb3 as first opcode
Bastian Koppelmann [Fri, 12 Dec 2014 16:55:34 +0000 (16:55 +0000)]
target-tricore: Add instructions of RR1 opcode format, that have 0xb3 as first opcode

Add instructions of RR1 opcode format, that have 0xb3 as first opcode.
Add helper functions mulh, mulmh and mulrh, that compute multiplication,
with multiprecision (mulmh) or rounding (mulrh) of 4 halfwords, being either low or high parts
of two 32 bit regs.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
9 years agotarget-tricore: Fix MFCR/MTCR insn and B format offset.
Bastian Koppelmann [Wed, 3 Dec 2014 17:40:21 +0000 (17:40 +0000)]
target-tricore: Fix MFCR/MTCR insn and B format offset.

Fix gen_mtcr using wrong register.
Fix gen_mtcr/mfcr using sign extended offsets.
Fix B format insn using not sign extendend offsets.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
9 years agotarget-tricore: Add missing 1.6 insn of BOL opcode format
Bastian Koppelmann [Fri, 5 Dec 2014 14:37:17 +0000 (14:37 +0000)]
target-tricore: Add missing 1.6 insn of BOL opcode format

Some of the 1.6 ISA instructions were still missing. So let's add them.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
9 years agotarget-tricore: Add instructions of RR opcode format, that have 0x4b as the first...
Bastian Koppelmann [Tue, 2 Dec 2014 17:22:27 +0000 (17:22 +0000)]
target-tricore: Add instructions of RR opcode format, that have 0x4b as the first opcode

Add instructions of RR opcode format, that have 0x4b as the first opcode.
Add helper functions:
    * parity: Calculates the parity bits for every byte of a 32 int.
    * bmerge/bsplit: Merges two regs into one bitwise/Splits one reg into two bitwise.
    * unpack: unpack a IEEE 754 single precision floating point number as exponent and mantissa.
    * dvinit_b_13/131: (ISA v1.3/v1.31)Prepare operands for a divide operation,
                       where the quotient result is guaranteed to fit into 8 bit.
    * dvinit_h_13/131: (ISA v1.3/v1.31)Prepare operands for a divide operation,
                       where the quotient result is guaranteed to fit into 16 bit.
OPCM_32_RR_FLOAT -> OPCM_32_RR_DIVIDE.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
9 years agotarget-tricore: Add instructions of RR opcode format, that have 0x1 as the first...
Bastian Koppelmann [Tue, 2 Dec 2014 17:15:09 +0000 (17:15 +0000)]
target-tricore: Add instructions of RR opcode format, that have 0x1 as the first opcode

Add instructions of RR opcode format, that have 0x1 as the first opcode.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
9 years agotarget-tricore: Add instructions of RR opcode format, that have 0xf as the first...
Bastian Koppelmann [Fri, 28 Nov 2014 17:07:26 +0000 (17:07 +0000)]
target-tricore: Add instructions of RR opcode format, that have 0xf as the first opcode

Add instructions of RR opcode format, that have 0xf as the first opcode.
Add helper functions:
    * clo/z/s: Counts leading ones/zeros/signs.
    * clo/z/s_h: Count leading ones/zeros/signs in two haflwords.
    * sh/_h: Shifts one/two word/hwords.
    * sha/_h: Shifts one/two word/hwords arithmeticly.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
9 years agotarget-tricore: Add instructions of RR opcode format, that have 0xb as the first...
Bastian Koppelmann [Thu, 27 Nov 2014 14:30:33 +0000 (14:30 +0000)]
target-tricore: Add instructions of RR opcode format, that have 0xb as the first opcode

Add instructions of RR opcode format, that have 0xb as the first opcode.
Add helper functions, for hword and byte arithmetics:
    * add_h_ssov/suov: Add two halfword and saturate on overflow.
    * sub_h_ssov/suov: Sub two halfword and saturate on overflow.
    * absdif_h_ssov: Compute absolute difference for halfwords and saturate on overflow.
    * abs_h_ssov/suov: Compute absolute value for two halfwords and saturate on overflow.
    * abs_b/h: Compute absolute value for four/two bytes/halfwords
    * absdif_b/h: Compute absolute difference for four/two bytes/halfwords
    * add_b/h: Add four/two bytes/halfwords.
    * sub_b/h: Sub four/two bytes/halfwords.
    * eq_b/h: Compare four/two bytes/halfwords with four/two bytes/halfwords on
              equality and set all bits of to either one ore zero.
    * eqany_b/h: Compare four/two bytes/halfwords with four/two bytes/halfwords on equality.
    * lt_b/bu/h/hu: Compare four/two bytes/halfwords with four/two bytes/halfwords
                    on less than signed and unsigned.
    * max_b/bu/h/hu: Calculate max for four/two bytes/halfwords signed and unsigned.
    * min_b/bu/h/hu: Calculate min for four/two bytes/halfwords signed and unsigned.
Add helper function abs_ssov, that computes the absolute value for a 32 bit integer and saturates on overflow.
Add microcode generator functions:
    * gen_sub_CC: Caluclates sub and sets the carry bit.
    * gen_subc_CC: Caluclates sub and carry and sets the carry bit
    * gen_abs: Compute absolute value for a 32 bit integer.
    * gen_cond_w: Compares two 32 bit values on cond and sets result either zero or all bits one.

OPC2_32_RR_MIN switched with OPC2_32_RR_MIN_U.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
9 years agotarget-tricore: Change SSOV/SUOV makro name to SSOV32/SUOV32
Bastian Koppelmann [Thu, 27 Nov 2014 13:45:09 +0000 (13:45 +0000)]
target-tricore: Change SSOV/SUOV makro name to SSOV32/SUOV32

Those makros are exclusively used for 32 bit arithmetics and won't work for
16 bit with two halfwords. So lets get rid of the len parameter and make them
always use 32 bit. Now no token pasting is needed anymore and they can be
regular functions.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
9 years agotarget-tricore: Fix mask handling JNZ.T being 7 bit long
Bastian Koppelmann [Mon, 15 Dec 2014 21:36:31 +0000 (21:36 +0000)]
target-tricore: Fix mask handling JNZ.T being 7 bit long

The mask is actually 7 bit long, instead of 6, so the expression checking
for JNZ.T is always false. Let's make the mask 1 bit wider.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
9 years agotarget-tricore: pretty-print register dump and show more status registers
Alex Zuepke [Fri, 19 Dec 2014 11:34:34 +0000 (12:34 +0100)]
target-tricore: pretty-print register dump and show more status registers

Now using psw_read() to retrieve the status bits correctly.

Signed-off-by: Alex Zuepke <alexander.zuepke@hs-rm.de>
Reviewed-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
9 years agotarget-tricore: add missing 64-bit MOV in RLC format
Alex Zuepke [Fri, 12 Dec 2014 14:10:29 +0000 (15:10 +0100)]
target-tricore: add missing 64-bit MOV in RLC format

Signed-off-by: Alex Zuepke <alexander.zuepke@hs-rm.de>
Reviewed-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
9 years agotarget-tricore: typo in BOL format
Alex Zuepke [Fri, 12 Dec 2014 14:10:28 +0000 (15:10 +0100)]
target-tricore: typo in BOL format

Signed-off-by: Alex Zuepke <alexander.zuepke@hs-rm.de>
Reviewed-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
9 years agotarget-tricore: fix offset masking in BOL format
Alex Zuepke [Fri, 12 Dec 2014 14:10:27 +0000 (15:10 +0100)]
target-tricore: fix offset masking in BOL format

Signed-off-by: Alex Zuepke <alexander.zuepke@hs-rm.de>
Reviewed-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
9 years agoMerge remote-tracking branch 'remotes/stefanha/tags/net-pull-request' into staging
Peter Maydell [Sat, 20 Dec 2014 22:04:13 +0000 (22:04 +0000)]
Merge remote-tracking branch 'remotes/stefanha/tags/net-pull-request' into staging

# gpg: Signature made Fri 19 Dec 2014 13:18:18 GMT using RSA key ID 81AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>"

* remotes/stefanha/tags/net-pull-request:
  e1000: defer packets until BM enabled
  net: Use g_new() & friends where that makes obvious sense
  net: Fuse g_malloc(); memset() into g_new0()
  net: don't use set/get_pointer() in set/get_netdev()
  tap: fix vcpu long time io blocking on tap

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/kraxel/tags/pull-roms-20141217-1' into staging
Peter Maydell [Sat, 20 Dec 2014 21:28:53 +0000 (21:28 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-roms-20141217-1' into staging

update ipxe from 69313ed to 35c5379

# gpg: Signature made Wed 17 Dec 2014 14:45:04 GMT using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"

* remotes/kraxel/tags/pull-roms-20141217-1:
  update ipxe from 69313ed to 35c5379

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoexec: change default exception_index value for migration to -1
Paolo Bonzini [Fri, 19 Dec 2014 11:53:14 +0000 (12:53 +0100)]
exec: change default exception_index value for migration to -1

In QEMU 2.2 the exception_index value was added to the migration stream
through a subsection.  The default was set to 0, which is wrong and
should have been -1.

However, 2.2 does not have commit e511b4d (cpu-exec: reset exception_index
correctly, 2014-11-26), hence in 2.2 the exception_index is never used
and is set to -1 on the next call to cpu_exec.  So we can change the
migration stream to make the default -1.  The effects are:

- 2.2.1 -> 2.2.0: cpu->exception_index set incorrectly to 0 if it
were -1 on the source; then reset to -1 in cpu_exec.  This is TCG
only; KVM does not use exception_index.

- 2.2.0 -> 2.2.1: cpu->exception_index set incorrectly to -1 if it
were 0 on the source; but it would be reset to -1 in cpu_exec anyway.
This is TCG only; KVM does not use exception_index.

- 2.2.1 -> 2.1: two bugs fixed: 1) can migrate backwards if
cpu->exception_index is set to -1; 2) should not migrate backwards
(but 2.2.0 allows it) if cpu->exception_index is set to 0

- 2.2.0 -> 2.3.0: 2.2.0 will send the subsection unnecessarily if
exception_index is -1, but that is not a problem.  2.3.0 will set
cpu->exception_index to -1 if it is 0 on the source, but this would
be anyway a problem for 2.2.0 -> 2.2.x migration (due to lack of
commit e511b4d in 2.2.x) so we can ignore it

- 2.2.1 -> 2.3.0: everything works.

In addition, play it safe and never send the subsection unless TCG
is in use.  KVM does not use exception_index (PPC KVM stores values
in it for use in the subsequent call to ppc_cpu_do_interrupt, but
does not need it as soon as kvm_handle_debug returns).  Xen and
qtest do not run any code for the CPU at all.

Reported-by: Igor Mammedov <imammedo@redhat.com>
Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com>
Tested-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 1418989994-17244-3-git-send-email-pbonzini@redhat.com
Cc: qemu-stable@nongnu.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agocpu: initialize cpu->exception_index on reset
Paolo Bonzini [Fri, 19 Dec 2014 11:53:13 +0000 (12:53 +0100)]
cpu: initialize cpu->exception_index on reset

This unbreaks linux-user (broken by e511b4d, cpu-exec: reset exception_index
correctly, 2014-11-26).

Reported-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com>
Tested-by: Eduardo Habkost <ehabkost@redhat.com>
Message-id: 1418989994-17244-2-git-send-email-pbonzini@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoe1000: defer packets until BM enabled
Michael S. Tsirkin [Mon, 1 Dec 2014 18:06:52 +0000 (20:06 +0200)]
e1000: defer packets until BM enabled

Some guests seem to set BM for e1000 after
enabling RX.
If packets arrive in the window, device is wedged.
Probably works by luck on real hardware, work around
this by making can_receive depend on BM.

Tested-by: Gabriel Somlo <somlo@cmu.edu>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
9 years agonet: Use g_new() & friends where that makes obvious sense
Markus Armbruster [Thu, 4 Dec 2014 13:28:17 +0000 (14:28 +0100)]
net: Use g_new() & friends where that makes obvious sense

g_new(T, n) is neater than g_malloc(sizeof(T) * n).  It's also safer,
for two reasons.  One, it catches multiplication overflowing size_t.
Two, it returns T * rather than void *, which lets the compiler catch
more type errors.

This commit only touches allocations with size arguments of the form
sizeof(T).

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
9 years agonet: Fuse g_malloc(); memset() into g_new0()
Markus Armbruster [Thu, 4 Dec 2014 13:28:16 +0000 (14:28 +0100)]
net: Fuse g_malloc(); memset() into g_new0()

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
9 years agonet: don't use set/get_pointer() in set/get_netdev()
Jason Wang [Mon, 13 Oct 2014 05:16:37 +0000 (13:16 +0800)]
net: don't use set/get_pointer() in set/get_netdev()

Commit 1ceef9f27359cbe92ef124bf74de6f792e71f6fb (net: multiqueue
support) tries to use set_pointer() and get_pointer() to set and get
NICPeers which is not a pointer defined in DEFINE_PROP_NETDEV. This
trick works but result a unclean and fragile implementation (e.g
print_netdev and parse_netdev).

This patch solves this issue by not using set/get_pinter() and set and
get netdev directly in set_netdev() and get_netdev(). After this the
parse_netdev() and print_netdev() were no longer used and dropped from
the source.

[Renamed 'err' label to 'out' as suggested by Markus Armbruster.
--Stefan]

Cc: Markus Armbruster <armbru@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Cc: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
9 years agotap: fix vcpu long time io blocking on tap
Wangkai (Kevin,C) [Fri, 18 Jul 2014 09:33:42 +0000 (09:33 +0000)]
tap: fix vcpu long time io blocking on tap

[Adjusted doc comment for grammar.
--Stefan]

Signed-off-by: Wangkai <wangkai86@huawei.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
9 years agoMerge remote-tracking branch 'remotes/kraxel/tags/pull-vga-20141216-1' into staging
Peter Maydell [Wed, 17 Dec 2014 19:22:41 +0000 (19:22 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-vga-20141216-1' into staging

cirrus hwcursor fixes.
set secondary-vga category.

# gpg: Signature made Tue 16 Dec 2014 14:44:09 GMT using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"

* remotes/kraxel/tags/pull-vga-20141216-1:
  vga: set catagory bit for secondary vga device
  move hw cursor pos from cirrus to vga
  cirrus: Force use of shadow pixmap when HW cursor is enabled
  vga: Add mechanism to force the use of a shadow surface

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/xtensa/tags/20141217-xtensa' into staging
Peter Maydell [Wed, 17 Dec 2014 17:31:26 +0000 (17:31 +0000)]
Merge remote-tracking branch 'remotes/xtensa/tags/20141217-xtensa' into staging

Xtensa updates for 2.3:

- fix cross-page opcode handling;
- move window overflow exception generation decision to translation phase;
- don't generate dead code after privilege, window overflow or coprocessor
  exception;
- add monitor command 'info opcount' for dumping TCG opcode counters.

# gpg: Signature made Wed 17 Dec 2014 02:57:01 GMT using RSA key ID F83FA044
# gpg: Good signature from "Max Filippov <max.filippov@cogentembedded.com>"
# gpg:                 aka "Max Filippov <jcmvbkbc@gmail.com>"

* remotes/xtensa/tags/20141217-xtensa:
  target-xtensa: don't generate dead code
  target-xtensa: record available window in TB flags
  target-xtensa: test cross-page opcode
  target-xtensa: fix translation for opcodes crossing page boundary
  tcg: add separate monitor command to dump opcode counters

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/lalrae/tags/mips-20141216' into staging
Peter Maydell [Wed, 17 Dec 2014 16:25:21 +0000 (16:25 +0000)]
Merge remote-tracking branch 'remotes/lalrae/tags/mips-20141216' into staging

* remotes/lalrae/tags/mips-20141216: (30 commits)
  target-mips: remove excp_names[] from linux-user as it is unused
  disas/mips: disable unused mips16_to_32_reg_map[]
  disas/mips: remove unused mips_msa_control_names_numeric[32]
  target-mips: convert single case switch into if statement
  target-mips: Fix DisasContext's ulri member initialization
  target-mips: Use local float status pointer across MSA macros
  target-mips: Add missing calls to synchronise SoftFloat status
  linux-user: Use the 5KEf processor for 64-bit emulation
  target-mips: Also apply the CP0.Status mask to MTTC0
  target-mips: gdbstub: Clean up FPU register handling
  target-mips: Correct 32-bit address space wrapping
  target-mips: Tighten ISA level checks
  target-mips: Fix CP0.Config3.ISAOnExc write accesses
  target-mips: Output CP0.Config2-5 in the register dump
  target-mips: Fix the 64-bit case for microMIPS MOVE16 and MOVEP
  target-mips: Correct the writes to Status and Cause registers via gdbstub
  target-mips: Correct the handling of writes to CP0.Status for MIPSr6
  target-mips: Correct MIPS16/microMIPS branch size calculation
  target-mips: Restore the order of helpers
  target-mips: Remove unused `FLOAT_OP' macro
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoupdate ipxe from 69313ed to 35c5379
Gerd Hoffmann [Wed, 17 Dec 2014 12:56:50 +0000 (13:56 +0100)]
update ipxe from 69313ed to 35c5379

Anton D. Kachalov (1):
      [intel] Add 8086:1557 card (Intel 82599 10G ethernet mezz)

Christian Hesse (1):
      [build] Merge util/geniso and util/genliso

Curtis Larsen (3):
      [efi] Use EFI_CONSOLE_CONTROL_PROTOCOL to set text mode if available
      [efi] Report errors from attempting to disconnect existing drivers
      [efi] Try various possible SNP receive filters

Dale Hamel (1):
      [smbios] Expose board serial number as ${board-serial}

Florian Schmaus (1):
      [build] Set GITVERSION only if there is a git repository

Hannes Reinecke (3):
      [ethernet] Provide eth_random_addr() to generate random Ethernet addresses
      [igbvf] Assign random MAC address if none is set
      [igbvf] Allow changing of MAC address

Jan Kiszka (1):
      [intel] Add I217-LM PCI ID

Marin Hannache (4):
      [nfs] Fix an invalid free() when loading a symlink
      [nfs] Fix an invalid free() when loading a regular (non-symlink) file
      [nfs] Rewrite NFS URI handling
      [readline] Add CTRL-W shortcut to remove a word

Michael Brown (144):
      [profile] Allow interrupts to be excluded from profiling results
      [intel] Exclude time spent in hypervisor from profiling
      [build] Fix version.o dependency upon git index
      [tcp] Defer sending ACKs until all received packets have been processed
      [lkrnprefix] Function as a bzImage kernel
      [build] Avoid errors when build directory is mounted via NFS
      [undi] Apply quota only to number of complete received packets
      [lkrnprefix] Make real-mode setup code relocatable
      [intel] Increase receive ring fill level
      [syslog] Strip invalid characters from hostname
      [test] Add self-tests for strdup()
      [libc] Prevent strndup() from reading beyond the end of the string
      [efi] Allow for optional protocols
      [efi] Make EFI_DEVICE_PATH_TO_TEXT_PROTOCOL optional
      [efi] Make EFI_HII_DATABASE_PROTOCOL optional
      [efi] Do not try to fetch loaded image device path protocol
      [ipv6] Fix definition of IN6_IS_ADDR_LINKLOCAL()
      [dhcpv6] Do not set sin6_scope_id on the unspecified client socket address
      [ipv6] Do not set sin6_scope_id on source address
      [ipv6] Include network device when transcribing multicast addresses
      [ipv6] Avoid potentially copying from a NULL pointer in ipv6_tx()
      [librm] Allow for the PIC interrupt vector offset to be changed
      [ifmgmt] Do not sleep CPU while configuring network devices
      [scsi] Improve sense code parsing
      [iscsi] Read IPv4 settings only from the relevant network device
      [iscsi] Include IP address origin in iBFT
      [debug] Allow debug message colours to be customised via DBGCOL=...
      [build] Expose build timestamp, build name, and product names
      [efi] Allow device paths to be easily included in debug messages
      [efi] Provide a meaningful EFI SNP device name
      [efi] Restructure EFI driver model
      [build] Fix erroneous object name in version object
      [build] Add yet another potential location for isolinux.bin
      [efi] Allow network devices to be created on top of arbitrary SNP devices
      [autoboot] Allow autoboot device to be identified by link-layer address
      [efi] Identify autoboot device by MAC address when chainloading
      [efi] Attempt to start only drivers claiming support for a device
      [efi] Rewrite SNP NIC driver
      [efi] Include SNP NIC driver within the all-drivers target
      [crypto] Add support for iPAddress subject alternative names
      [crypto] Fix debug message
      [netdevice] Reset network device index when last device is unregistered
      [efi] Update EDK2 headers
      [efi] Install our own disk I/O protocol and claim exclusive use of it
      [efi] Allow for interception of boot services calls by loaded image
      [efi] Print well-known GUIDs by name in debug messages
      [efi] Include EFI_CONSOLE_CONTROL_PROTOCOL header
      [ioapi] Fail ioremap() when attempting to map a zero bus address
      [intel] Check for ioremap() failures
      [realtek] Check for ioremap() failures
      [vmxnet3] Check for ioremap() failures
      [skel] Check for ioremap() failures
      [myson] Check for ioremap() failures
      [natsemi] Check for ioremap() failures
      [i386] Add functions to read and write model-specific registers
      [x86_64] Add functions to read and write model-specific registers
      [efi] Show more diagnostic information when building with DEBUG=efi_wrap
      [ioapi] Centralise notion of PAGE_SIZE
      [lotest] Discard packets arriving on the incorrect network device
      [xen] Import selected public headers
      [xen] Add basic support for PV-HVM domains
      [xen] Add support for Xen netfront virtual NICs
      [efi] Default to releasing network devices for use via SNP
      [efi] Unload started images only on failure
      [efi] Fill in loaded image's DeviceHandle if firmware fails to do so
      [efi] Fix incorrect debug message level when device has no device path
      [efi] Report exact failure when unable to open the device path
      [netdevice] Avoid registering duplicate network devices
      [efi] Ignore failures when attempting to install SNP HII protocol
      [efi] Expand the range of well-known EFI GUIDs in debug messages
      [efi] Provide efi_handle_name() for debugging
      [efi] Add ability to dump all openers of a given protocol on a handle
      [efi] Use efi_handle_name() instead of efi_handle_devpath_text()
      [efi] Use efi_handle_name() instead of efi_devpath_text() where applicable
      [efi] Allow compiler to perform type checks on EFI_HANDLE
      [efi] Avoid unnecessarily passing pointers to EFI_HANDLEs
      [efi] Dump existing openers when we are unable to open a protocol
      [efi] Dump handle information around connect/disconnect attempts
      [efi] Improve debugging of the debugging facilities
      [efi] Add excessive sanity checks into efi_debug functions
      [efi] Also try original ComponentName protocol for retrieving driver names
      [efi] Print raw device path when we have no DevicePathToTextProtocol
      [efi] Add ability to dump SNP device mode information
      [efi] Reset multicast filter list when setting SNP receive filters
      [efi] Provide centralised definitions of commonly-used GUIDs
      [efi] Open device path protocol only at point of use
      [efi] Move abstract device path and handle functions to efi_utils.c
      [efi] Generalise snpnet_pci_info() to efi_locate_device()
      [bios] Support displaying and hiding cursor
      [efi] Support displaying and hiding cursor
      [readline] Ensure cursor is visible when prompting for input
      [xen] Accept alternative Xen platform PCI device ID 5853:0002
      [xen] Use version 1 grant tables by default
      [xen] Cope with unexpected initial backend states
      [smc9000] Avoid using CONFIG as a preprocessor macro
      [build] Allow for named configurations at build time
      [intel] Display PBS value when applying ICH errata workaround
      [intel] Display before and after values for both PBS and PBA
      [intel] Apply PBS/PBA errata workaround only to ICH8 PCI device IDs
      [efi] Add definitions of GUIDs observed during Windows boot
      [efi] Dump details of any calls to our dummy block and disk I/O protocols
      [romprefix] Do not preserve unused register %di
      [build] Remove obsolete references to .zrom build targets
      [build] Allow ISA ROMs to be built
      [build] Avoid deleting config header files if build is interrupted
      [prefix] Halt system without burning CPU if we cannot access the payload
      [prefix] Report both %esi and %ecx when opening payload fails
      [util] Use PCI length field to obtain length of individual images
      [mromprefix] Use PCI length field to obtain length of individual images
      [mromprefix] Allow for .mrom images larger than 128kB
      [efi] Show details of intercepted LoadImage() calls
      [efi] Make our virtual file system case insensitive
      [efi] Wrap any images loaded by our wrapped image
      [efi] Use the SNP protocol instance to match the SNP chainloading device
      [efi] Avoid returning uninitialised data from PCI configuration space reads
      [efi] Make EFI_PCI_ROOT_BRIDGE_IO_PROTOCOL optional
      [efi] Allow for non-PCI snpnet devices
      [build] Clean up all binary directories on "make [very]clean"
      [efi] Add efifatbin utility
      [efi] Provide dummy device path in efi_image_probe()
      [dhcp] Check for matching chaddr in received DHCP packets
      [dhcp] Remove obsolete dhcp_chaddr() function
      [build] Use -malign-double to build 32-bit UEFI binaries
      [efi] Centralise definitions of more protocol GUIDs
      [efi] Add definitions of GUIDs observed when chainloading from Intel driver
      [efi] Free transmit ring entry before calling netdev_tx_complete()
      [efi] Generalise snpnet_dev_info() to efi_device_info()
      [efi] Update to current EDK2 headers
      [efi] Add NII / UNDI driver
      [efi] Check for presence of UNDI in NII protocol
      [efi] Include NII driver within "snp" and "snponly" build targets
      [ping] Report timed-out pings via the callback function
      [ping] Allow termination after a specified number of packets
      [ping] Allow "ping" command output to be inhibited
      [intel] Use autoloaded MAC address instead of EEPROM MAC address
      [crypto] Fix parsing of OCSP responder ID key hash
      [vmxnet3] Add profiling code to exclude time spent in the hypervisor
      [netdevice] Fix erroneous use of free(iobuf) instead of free_iob(iobuf)
      [libc] Add ASSERTED macro to test if any assertion has triggered
      [list] Add sanity checks after list-adding functions
      [malloc] Tidy up debug output
      [malloc] Sanity check parameters to alloc_memblock() and free_memblock()
      [malloc] Check integrity of free list
      [malloc] Report caller address as soon as memory corruption is detected

Peter Lemenkov (1):
      [build] Check if git index actually exists

Robin Smidsrød (2):
      [build] Add named configuration for VirtualBox
      [build] Avoid using embedded script in VirtualBox named configuration

Sven Ulland (1):
      [lacp] Set "aggregatable" flag in response LACPDU

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agosdl2: Work around SDL2 SDL_ShowWindow() bug
Max Reitz [Fri, 12 Dec 2014 09:52:52 +0000 (10:52 +0100)]
sdl2: Work around SDL2 SDL_ShowWindow() bug

Apparently it is possible for X to send an event to a hidden SDL2
window, leading to SDL2 believing it is now shown. SDL2 will pass the
SDL_WINDOWEVENT_SHOWN message to the application without actually
showing the window; the problem is that the next SDL_ShowWindow() will
be a no-op because SDL2 assumes the window is already shown.

The correct way to react to SDL_WINDOWEVENT_SHOWN would be to clear
scon->hidden (analogous for SDL_WINDOWEVENT_HIDDEN). However, due to the
window not actually being shown, this will somehow not be correct after
all.

Therefore, just hide the window on SDL_WINDOWEVENT_SHOWN if it is
supposed to be hidden (and analogous for SDL_WINDOWEVENT_HIDDEN).

Signed-off-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agosdl2: Use correct sdl2_console for window events
Max Reitz [Fri, 12 Dec 2014 09:52:51 +0000 (10:52 +0100)]
sdl2: Use correct sdl2_console for window events

SDL_PollEvent() polls events for all windows; therefore,
sdl2_poll_events() will poll the events for all windows and not only for
the one identified by the given sdl2_console.

This should be considered in handle_windowevent(): The window affected
by the event is not necessarily the one identified by the sdl2_console
object given to sdl2_poll_events(), but the one identified by
ev->window.windowID.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agosdl2: move sdl2_2d_refresh to sdl2-2d.c
Gerd Hoffmann [Wed, 12 Nov 2014 07:03:34 +0000 (08:03 +0100)]
sdl2: move sdl2_2d_refresh to sdl2-2d.c

Now that common event handling code is split off, we can move
over sdl_refresh to sdl2-2d.c, and rename it to sdl2_2d_refresh.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
9 years agosdl2: factor out sdl2_poll_events
Gerd Hoffmann [Wed, 12 Nov 2014 07:01:27 +0000 (08:01 +0100)]
sdl2: factor out sdl2_poll_events

Create a new function to poll and handle sdl2 events,
which is then just called from the refresh timer.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
9 years agosdl2: add+use sdl2_2d_redraw function.
Gerd Hoffmann [Tue, 11 Nov 2014 12:31:08 +0000 (13:31 +0100)]
sdl2: add+use sdl2_2d_redraw function.

Add a new sdl2_2d_redraw function for a complete screen refresh,
so we can stop using graphic_hw_invalidate for that.  There is
no need to bother console / gfx emulation code if we are just
going to re-blit the screen after window resizes.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
9 years agosdl2: move sdl_switch to sdl2-2d.c
Gerd Hoffmann [Tue, 11 Nov 2014 12:22:49 +0000 (13:22 +0100)]
sdl2: move sdl_switch to sdl2-2d.c

Move sdl_switch to sdl2-2d.c file, rename to sdl2_2d_switch.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
9 years agosdl2: overhaul window size handling
Gerd Hoffmann [Tue, 11 Nov 2014 12:12:02 +0000 (13:12 +0100)]
sdl2: overhaul window size handling

Split do_sdl_resize function (which does alot more than just resizing)
into three: sdl2_window_{create,destroy,resize}.

Fix SDL_Renderer handling: must be guest display size not host window
size, and SDL2 will magically handle all scaling for us.

Make fullscreen actually enter fullscreen mode and simplify the code.
There is no need to store the original window size, the window manager
will do that for us.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
9 years agosdl2: move sdl_update to new sdl2-2d.c
Gerd Hoffmann [Tue, 11 Nov 2014 10:09:26 +0000 (11:09 +0100)]
sdl2: move sdl_update to new sdl2-2d.c

Create new sdl2-2d file for 2d display rendering.
Move over sdl_update code, and rename to sdl2_2d_update.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
9 years agosdl2: turn on keyboard grabs
Gerd Hoffmann [Tue, 11 Nov 2014 11:02:50 +0000 (12:02 +0100)]
sdl2: turn on keyboard grabs

Makes quite some keys actually go to the guest instead of
being captured by the host window manager.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
9 years agosdl2: move keyboard input code to new sdl2-input.c
Gerd Hoffmann [Tue, 11 Nov 2014 09:58:19 +0000 (10:58 +0100)]
sdl2: move keyboard input code to new sdl2-input.c

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
9 years agosdl2: rename sdl2_state to sdl2_console, move to header file
Gerd Hoffmann [Tue, 11 Nov 2014 09:21:24 +0000 (10:21 +0100)]
sdl2: rename sdl2_state to sdl2_console, move to header file

Create sdl2.h header file, in preparation for sdl2 code splitup.
Populate it with sdl2_console struct (renamed from sdl2_state).

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
9 years agotarget-xtensa: don't generate dead code
Max Filippov [Sat, 8 Nov 2014 16:00:55 +0000 (19:00 +0300)]
target-xtensa: don't generate dead code

Don't generate TCG operations when privilege, register window or
coprocessor checks fail.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
9 years agotarget-xtensa: record available window in TB flags
Max Filippov [Thu, 30 Oct 2014 15:07:47 +0000 (18:07 +0300)]
target-xtensa: record available window in TB flags

Record last valid 4-register window pane number in TB flags so that a
window overflow exception throw point is known at the translation time.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
9 years agotarget-xtensa: test cross-page opcode
Max Filippov [Sun, 14 Dec 2014 05:23:52 +0000 (08:23 +0300)]
target-xtensa: test cross-page opcode

Alter cross-page TB test to also test cross-page opcode.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
9 years agotarget-xtensa: fix translation for opcodes crossing page boundary
Max Filippov [Sun, 14 Dec 2014 04:50:55 +0000 (07:50 +0300)]
target-xtensa: fix translation for opcodes crossing page boundary

If TB ends with an opcode that crosses page boundary and the following
page is not executable then EPC1 for the code fetch exception wrongly
points at the beginning of the TB. Always treat instruction that crosses
page boundary as a separate TB.

Cc: qemu-stable@nongnu.org
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
9 years agotcg: add separate monitor command to dump opcode counters
Max Filippov [Sun, 2 Nov 2014 08:04:18 +0000 (11:04 +0300)]
tcg: add separate monitor command to dump opcode counters

Currently 'info jit' outputs half of the information to monitor and the
rest to qemu log. Dumping opcode counts to monitor as a part of 'info
jit' command doesn't sound useful. Add new monitor command 'info
opcount' that only dumps opcode counters.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
9 years agoqemu-log: add log category for MMU info
Antony Pavlov [Sat, 13 Dec 2014 16:48:18 +0000 (19:48 +0300)]
qemu-log: add log category for MMU info

Running barebox on qemu-system-mips* with '-d unimp' overloads
stderr by very very many mips_cpu_handle_mmu_fault() messages:

  mips_cpu_handle_mmu_fault address=b80003fd ret 0 physical 00000000180003fd prot 3
  mips_cpu_handle_mmu_fault address=a0800884 ret 0 physical 0000000000800884 prot 3
  mips_cpu_handle_mmu_fault pc a080cd80 ad b80003fd rw 0 mmu_idx 0

So it's very difficult to find LOG_UNIMP message.

The mips_cpu_handle_mmu_fault() messages appear on enabling ANY
logging! It's not very handy.

Adding separate log category for *_cpu_handle_mmu_fault()
logging fixes the problem.

Signed-off-by: Antony Pavlov <antonynpavlov@gmail.com>
Acked-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Message-id: 1418489298-1184-1-git-send-email-antonynpavlov@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/spice/tags/pull-spice-20141216-1' into staging
Peter Maydell [Tue, 16 Dec 2014 16:52:42 +0000 (16:52 +0000)]
Merge remote-tracking branch 'remotes/spice/tags/pull-spice-20141216-1' into staging

misc spice updates.

# gpg: Signature made Tue 16 Dec 2014 14:03:07 GMT using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"

* remotes/spice/tags/pull-spice-20141216-1:
  spice: fix memory leak
  spice: remove spice-experimental.h include
  spice: do not require TCP ports
  spice: rework mirror allocation, add no-resize fast path
  spice: reduce refresh rate in native mode
  spice: use bottom half instead of refresh timer for cursor updates

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/amit-migration/tags/for-2.3-2' into staging
Peter Maydell [Tue, 16 Dec 2014 14:53:23 +0000 (14:53 +0000)]
Merge remote-tracking branch 'remotes/amit-migration/tags/for-2.3-2' into staging

Migration pull for 2.3.  Mostly moving the code to the migration/
directory, and updating MAINTAINERS.

I've also folded my other MAINTAINERS update patches into this, as
they're small by themselves.

# gpg: Signature made Tue 16 Dec 2014 12:21:24 GMT using RSA key ID 854083B6
# gpg: Good signature from "Amit Shah <amit@amitshah.net>"
# gpg:                 aka "Amit Shah <amit@kernel.org>"
# gpg:                 aka "Amit Shah <amitshah@gmx.net>"

* remotes/amit-migration/tags/for-2.3-2:
  MAINTAINERS: Update for migrated migration code
  Split the QEMU buffered file code out
  Split struct QEMUFile out
  Remove migration- pre/post fixes off files in migration/ dir
  Start migrating migration code into a migration directory
  qmp-command.hx: add missing docs for migration capabilites
  cpu: verify that block->host is set
  cpu: assert host pointer offset within block
  exec: add wrapper for host pointer access
  MAINTAINERS: add include files to virtio-serial entry
  MAINTAINERS: add entry for virtio-rng
  MAINTAINERS: migration: add vmstate static checker files
  MAINTAINERS: Add myself to migration maintainers

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agovga: set catagory bit for secondary vga device
Gonglei [Fri, 5 Dec 2014 03:33:37 +0000 (11:33 +0800)]
vga: set catagory bit for secondary vga device

Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agomove hw cursor pos from cirrus to vga
Gerd Hoffmann [Thu, 16 Oct 2014 08:22:23 +0000 (10:22 +0200)]
move hw cursor pos from cirrus to vga

9 years agocirrus: Force use of shadow pixmap when HW cursor is enabled
Benjamin Herrenschmidt [Mon, 7 Jul 2014 00:28:39 +0000 (10:28 +1000)]
cirrus: Force use of shadow pixmap when HW cursor is enabled

The HW cursor cannot be painted on a shared surface. This fixes HW
cursor display in Windows NT 4.0 and Windows 98.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agovga: Add mechanism to force the use of a shadow surface
Benjamin Herrenschmidt [Mon, 7 Jul 2014 00:17:44 +0000 (10:17 +1000)]
vga: Add mechanism to force the use of a shadow surface

This prevents surface sharing which will be necessary to
fix cirrus HW cursor support.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agospice: fix memory leak
Gonglei [Fri, 5 Dec 2014 08:30:10 +0000 (16:30 +0800)]
spice: fix memory leak

If errors happen for middle items of channel_list,
qmp_query_spice_channels() returns NULL, and the variable
cur_item going out of scope leaks the storage it points to.

The flag is a compatibility thing for older spice-server
versions. Meanwhile our minimum spice version requirement is
new enough that we should never ever see this error, and if we
do something went very seriously wrong. Let's using assert()
instead of returning NULL to avoid a memory leak.

Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agospice: remove spice-experimental.h include
Marc-André Lureau [Mon, 17 Nov 2014 15:52:49 +0000 (16:52 +0100)]
spice: remove spice-experimental.h include

Nothing seems to be using functions from spice-experimental.h (better
that way). Let's remove its inclusion.

Signed-off-by: Marc-André Lureau <marcandre.lureau@gmail.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agospice: do not require TCP ports
Marc-André Lureau [Tue, 11 Nov 2014 12:39:19 +0000 (13:39 +0100)]
spice: do not require TCP ports

It is possible to use Spice server without TCP port.  On local VM,
qemu (and libvirt) can add new clients thanks to QMP add_client command.

Signed-off-by: Marc-André Lureau <marcandre.lureau@gmail.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agospice: rework mirror allocation, add no-resize fast path
Gerd Hoffmann [Sat, 8 Nov 2014 07:56:34 +0000 (08:56 +0100)]
spice: rework mirror allocation, add no-resize fast path

Add fast path to qemu_spice_display_switch in case old and new
displaysurface have identical size (happens with display panning
and page flipping).  We just swap the backing store then and don't
go through the whole process of deleting and creating the primary
surface.

To simplify the code a bit move mirror surface allocation to
qemu_spice_display_switch().

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agospice: reduce refresh rate in native mode
Gerd Hoffmann [Tue, 4 Nov 2014 13:16:12 +0000 (14:16 +0100)]
spice: reduce refresh rate in native mode

Now that cursor updates are out of the way qxl needs the refresh timer
only when when running in vga mode, for dirty bitmap checking.  In
native qxl mode the guest will notify us, so we don't need to poll and
can use the idle interval (one refresh wakeup every few seconds).

Cc: Marc-André Lureau <marcandre.lureau@gmail.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agospice: use bottom half instead of refresh timer for cursor updates
Gerd Hoffmann [Tue, 4 Nov 2014 12:59:59 +0000 (13:59 +0100)]
spice: use bottom half instead of refresh timer for cursor updates

Calling directly doesn't work due to the qxl-render code running in
spice server thread context.  Meanwhile bottom half scheduling is
thread-safe though, so we can use that to kick a cursor update in
main i/o thread context.

Cc: Marc-André Lureau <marcandre.lureau@gmail.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agosdl: move version logic from source code to makefile
Gerd Hoffmann [Tue, 11 Nov 2014 09:13:08 +0000 (10:13 +0100)]
sdl: move version logic from source code to makefile

Compile sdl.c / sdl2.c depending on CONFIG_SDLABI instead of
compiling both and have version #ifdefs in the source code.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
9 years agotarget-mips: remove excp_names[] from linux-user as it is unused
Leon Alrae [Wed, 10 Dec 2014 16:30:59 +0000 (16:30 +0000)]
target-mips: remove excp_names[] from linux-user as it is unused

Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
9 years agodisas/mips: disable unused mips16_to_32_reg_map[]
Leon Alrae [Fri, 12 Dec 2014 16:12:11 +0000 (16:12 +0000)]
disas/mips: disable unused mips16_to_32_reg_map[]

This array is used by print_mips16_insn_arg() which is guarded by #if 0.
Therefore doing the same with the array as it generates clang warnings.

Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
9 years agodisas/mips: remove unused mips_msa_control_names_numeric[32]
Leon Alrae [Wed, 10 Dec 2014 15:48:02 +0000 (15:48 +0000)]
disas/mips: remove unused mips_msa_control_names_numeric[32]

Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-mips: convert single case switch into if statement
Leon Alrae [Wed, 10 Dec 2014 15:39:42 +0000 (15:39 +0000)]
target-mips: convert single case switch into if statement

Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-mips: Fix DisasContext's ulri member initialization
Maciej W. Rozycki [Tue, 2 Dec 2014 22:31:33 +0000 (22:31 +0000)]
target-mips: Fix DisasContext's ulri member initialization

Set DisasContext's ulri member to 0 or 1 as with other bool members.

Signed-off-by: Maciej W. Rozycki <macro@codesourcery.com>
Reviewed-by: Leon Alrae <leon.alrae@imgtec.com>
Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
9 years agotarget-mips: Use local float status pointer across MSA macros
Maciej W. Rozycki [Tue, 2 Dec 2014 17:51:12 +0000 (17:51 +0000)]
target-mips: Use local float status pointer across MSA macros

Reduce line wrapping throughout MSA helper macros by using a local float
status pointer rather than referring to the float status through the
environment each time.  No functional change.

Signed-off-by: Maciej W. Rozycki <macro@codesourcery.com>
Reviewed-by: Leon Alrae <leon.alrae@imgtec.com>
Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
9 years agotarget-mips: Add missing calls to synchronise SoftFloat status
Maciej W. Rozycki [Tue, 2 Dec 2014 17:02:38 +0000 (17:02 +0000)]
target-mips: Add missing calls to synchronise SoftFloat status

Add missing calls to synchronise the SoftFloat status with the CP1.FSCR:

+ for the rounding and flush-to-zero modes upon processor reset,

+ for the flush-to-zero mode on FSCR updates through the GDB stub.

Refactor code accordingly and remove the redundant RESTORE_ROUNDING_MODE
macro.

Signed-off-by: Thomas Schwinge <thomas@codesourcery.com>
Signed-off-by: Maciej W. Rozycki <macro@codesourcery.com>
Reviewed-by: Leon Alrae <leon.alrae@imgtec.com>
Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
9 years agolinux-user: Use the 5KEf processor for 64-bit emulation
Maciej W. Rozycki [Thu, 20 Nov 2014 16:00:54 +0000 (16:00 +0000)]
linux-user: Use the 5KEf processor for 64-bit emulation

Replace the 20Kc original MIPS64 ISA processor used for 64-bit user
emulation with the 5KEf processor that implements the MIPS64r2 ISA,
complementing the choice of the 24Kf processor for 32-bit emulation.

Signed-off-by: Maciej W. Rozycki <macro@codesourcery.com>
Reviewed-by: Leon Alrae <leon.alrae@imgtec.com>
Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
9 years agotarget-mips: Also apply the CP0.Status mask to MTTC0
Maciej W. Rozycki [Thu, 20 Nov 2014 11:15:34 +0000 (11:15 +0000)]
target-mips: Also apply the CP0.Status mask to MTTC0

Make CP0.Status writes made with the MTTC0 instruction respect this
register's mask just like all the other places.  Also preserve the
current values of masked out bits.

Signed-off-by: Maciej W. Rozycki <macro@codesourcery.com>
Reviewed-by: Leon Alrae <leon.alrae@imgtec.com>
Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
9 years agotarget-mips: gdbstub: Clean up FPU register handling
Maciej W. Rozycki [Fri, 5 Dec 2014 18:46:04 +0000 (18:46 +0000)]
target-mips: gdbstub: Clean up FPU register handling

Rewrite the FPU register access parts of `mips_cpu_gdb_read_register'
and `mips_cpu_gdb_write_register' for consistency between each other.

Signed-off-by: Maciej W. Rozycki <macro@codesourcery.com>
Reviewed-by: Leon Alrae <leon.alrae@imgtec.com>
Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
9 years agotarget-mips: Correct 32-bit address space wrapping
Maciej W. Rozycki [Wed, 19 Nov 2014 17:29:00 +0000 (17:29 +0000)]
target-mips: Correct 32-bit address space wrapping

Make sure the address space is unconditionally wrapped on 32-bit
processors, that is ones that do not implement at least the MIPS III
ISA.

Also make MIPS16 SAVE and RESTORE instructions use address calculation
rather than plain arithmetic operations for stack pointer manipulation
so that their semantics for stack accesses follows the architecture
specification.  That in particular applies to user software run on
64-bit processors with the CP0.Status.UX bit clear where the address
space is wrapped to 32 bits.

Signed-off-by: Maciej W. Rozycki <macro@codesourcery.com>
Reviewed-by: Leon Alrae <leon.alrae@imgtec.com>
Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
9 years agotarget-mips: Tighten ISA level checks
Maciej W. Rozycki [Fri, 5 Dec 2014 18:47:58 +0000 (18:47 +0000)]
target-mips: Tighten ISA level checks

Tighten ISA level checks down to MIPS II that many of our instructions
are missing.  Also make sure any 64-bit instruction enables are only
applied to 64-bit processors, that is ones that implement at least the
MIPS III ISA.

Signed-off-by: Maciej W. Rozycki <macro@codesourcery.com>
Reviewed-by: Leon Alrae <leon.alrae@imgtec.com>
Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
9 years agotarget-mips: Fix CP0.Config3.ISAOnExc write accesses
Maciej W. Rozycki [Tue, 18 Nov 2014 03:59:07 +0000 (03:59 +0000)]
target-mips: Fix CP0.Config3.ISAOnExc write accesses

Fix CP0.Config3.ISAOnExc write accesses on microMIPS processors.  This
bit is mandatory for any processor that implements the microMIPS
instruction set.  This bit is r/w for processors that implement both the
standard MIPS and the microMIPS instruction set.  This bit is r/o and
hardwired to 1 if only the microMIPS instruction set is implemented.

There is no other bit ever writable in CP0.Config3 so defining a
corresponding `CP0_Config3_rw_bitmask' member in `CPUMIPSState' is I
think an overkill.  Therefore make the ability to write the bit rely on
the presence of ASE_MICROMIPS set in the instruction flags.

The read-only case of the microMIPS instruction set being implemented
only can be added when we add support for such a configuration.  We do
not currently have such support, we have no instruction flag that would
control the presence of the standard MIPS instruction set nor any
associated code in instruction decoding.

This change is needed to boot a microMIPS Linux kernel successfully,
otherwise it hangs early on as interrupts are enabled and then the
exception handler invoked loops as its first instruction is interpreted
in the wrong execution mode and triggers another exception right away.
And then over and over again.

We already check the current setting of the CP0.Config3.ISAOnExc in
`set_hflags_for_handler' to set the ISA bit correctly on the exception
handler entry so it is the ability to set it that is missing only.

Signed-off-by: Maciej W. Rozycki <macro@codesourcery.com>
Reviewed-by: Leon Alrae <leon.alrae@imgtec.com>
Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>