| Commit message (Collapse) | Author | Age | Files | Lines |
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The following changes are crucial for GPU H/W acceleration because
some graphics applications like games perform very frequent QEMU pipe
operations to send GLES commands to the host translator.
For each read/write buffer operation, currently QEMU pipe requires five
MMIO operations.
This causes significant overhead when running with hardware virtualization
enabled (e.g. HAXM) because each MMIO causes expensive transition from
the guest kernel to the kernel driver, and to the QEMU user space in the end.
Among such five MMIO accesses, four of them are required to just set up the
parameters for the access, like the buffer address, length etc. By passing
a buffer containing such parameters, we need only one MMIO operation to
send a GLES command to the host translator.
Update the qemu_pipe save version for pipe struct changes.
Change-Id: Idf6400f3c4c9c8473311312bb1d8f3ea92d68797
Signed-off-by: Xin, Xiaohui <xiaohui.xin@intel.com>
Signed-off-by: Jiang, Yunhong <yunhong.jiang@intel.com>
Signed-off-by: Nakajima, Jun <jun.nakajima@intel.com>
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: Icd617cb91fe87a6f88566dd57b6405b216ca3570
Signed-off-by: Chris Dearman <chris@mips.com>
|
|\ \ \
| |/ / |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
From v0.12.5 tag at git://git.sv.gnu.org/qemu.git
CommitID: 174f225e9d62e8f3002e274e4f718bd2a967fbf4
Change-Id: I35b49a4319cee4b69cf9da4e5af1f43327e21056
Signed-off-by: Chris Dearman <chris@mips.com>
|
|\ \ \ |
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Swap bytes in the display buffer if the host endianess does not match
the target endianess. This fixes a problem with the wrong colour
being displayed in the text framebuffer console.
Not sure if the is the correct place to fix it.. It might be better to
have the kernel goldfish framebuffer driver swap the bytes.
Change-Id: I8c7ada6e5cb6de7745511f3b83f07f0ac95276af
Signed-off-by: Chris Dearman <chris@mips.com>
|
| |/
|/|
| |
| |
| |
| |
| |
| |
| | |
Fix one issue because some field in nand device is not save/restore
Change-Id: I025addf4ddaea8420f6328b5e58c8514604b1719
Signed-off-by: Gao, Fengqian <fengqian.gao@intel.com>
Signed-off-by: Jiang, Yunhong <yunhong.jiang@intel.com>
Signed-off-by: Nakajima, Jun <jun.nakajiama@intel.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This is a continuation of an effort to simplify HW config, and make the UI for it clearer.
This CL gets rid of multiple camera emulation properties, combining everything into just
two properties:
- hw.camera_back that controls emulation of a camera facing back, and
- hw.camera_front that controls emulation of a camera facing front.
Change-Id: Ia114ae4caf2053685cbff00f39088e5b5be2952c
|
| |
| |
| |
| |
| |
| |
| | |
1. Remove unused hw.camera.maxHorizontalPixels, and hw.camera.maxVerticalPixels parameters.
2. Combine hw.touchScreen, and hw.multiTouch under one property: hw.screen
Change-Id: I3cdf64f7d7e46486110cbc0769f9d98a61f0bea5
|
|\ \ |
|
| |/
| |
| |
| | |
Change-Id: I311dc55fe10f41637775baa330a7c732ff317f51
|
|/
|
|
|
| |
Change-Id: I34e8e71f6ca03d58676ac350d6558fec32e6e448
Signed-off-by: SeongJae Park <sj38.park@gmail.com>
|
|
|
|
|
|
|
|
|
| |
The emulator crashes when loading a goldfish pipe from snapshot if it
was still in PipeConnection state when saved, since there is no
service associated with the pipe. Check for service existence before
dereferencing the service pointer.
Change-Id: I9093bdf403fd3532e9fb6584b98994d9d4dc686c
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Add snapshot load/save support to QEMU Pipes
This adds the ability to save and load QEMU Pipe connections
with snapshots. Note that by default, all loaded pipe client
connections are force-fully closed on load.
We don't have a good way to save the state of network
connections to persistent storage.
Properly implements snapshot save / load for qemu pipe clients.
Change-Id: Ie5767f8ce40c8341b958cc5844e724dd4fc1ed2b
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is a temporary fix that enables GoogleTV emulation on tools R15+ The permanent
solution would be to increase number of available IRQs from 16 to something more
appropriate. For now just:
1. Use IRQ3 for the events device.
2. Decrease number of TTY ports from 4 to 2.
These two measures would free up 3 IRQs, so the chances of running out of available
IRQs are significantly decreased.
Change-Id: Ie788880935d6e15f1f02f61e862c8727c8690b4a
|
|
|
|
|
|
|
|
|
| |
Forces overwriting (truncating) of the userimage-qemu.img file on condition that
it is being initialized from userimage.img file. If *-qemu file is not truncated
the system will pick up on artifacts located behind the overwritten area of
the *-qemu.img file.
Change-Id: I091dff0069d7049588b11c4654fe134947db101b
|
|
|
|
| |
Change-Id: Ib3891704ece2cea0c5a80b468b2514e6c4d13216
|
|
|
|
|
|
|
|
|
| |
There were two issues fixed here:
1. IRQ allocation (for a device) has never been checked on going out of bounds.
2. In x86 platform some IRQs were reserved (for kbd, mouse, and exception), but IRQ allocation
for goldfish devices didn't check for those reserved IRQs being assigned to a device.
Change-Id: I9d48485d973bcc8fd8a3dd1b78fbfa6a05aeea22
|
|
|
|
| |
Change-Id: I62a2d54562ac043a5513585a2e86aa01038fdc29
|
|
|
|
|
|
|
|
|
| |
The issue was that CPU registers (CR3 in particular) in QEMU were out of sync with KVM
at the time when virtual address to physical address translation was performed. This
caused translation failure, and the subsequent crash. The fix was to explicitly sync
QEMU registers with KVM registers just before calling VA->PA translation.
Change-Id: I1ff4ed2cfddd77e6889bb645f08db442f119049a
|
|
|
|
|
|
|
|
| |
This is necessary ot initialize qemu pipes when running an
emulated x86 system. Required to try OpenGLES emulation
for x86.
Change-Id: I4fa38648dcdd4e85d1794b131238b27b4dd4b91e
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently NAND device driver need 8 MMIO access for goldfish nand device access.
MMIO access is expensive in virtualization environment because each MMIO access
will cause a context switch from in-kernel module to qemu user space process.
Through introducing a new batch command, these multiple MMIO access can be
combined into one, thus improve virtualization perofrmance. The booting time in
KVM environment is reduced from 25~26 seconds to 20~21 seconds.
A kernel patch is required to utilize this new batch command.
The patched kernel will detect the batch support through a new flags in
goldfish nand device, while current kernel is not impacted.
Change-Id: Ia460b1ba3c6fde5b1fc6101bb66f8f58e46e6a78
Signed-off-by: Xin Xiaohui <xiaohui.xin@intel.com>
Signed-off-by: Zhang Xiantao <xiantao.zhang@intel.com>
Signed-off-by: Jiang Yunhong <yunhong.jiang@intel.com>
Signed-off-by: Nakajima Jun <jun.nakajima@intel.com>
|
|
|
|
|
|
|
|
|
| |
Throughout emulator's code pipe's 'poll' callback was returning PIPE_WAKE_XXX,
instead of PIPE_POLL_XXX flags. This created whole sort of issues with the
qemu pipe service <-> client communications.
This is also a fix for http://b/issue?id=5196348
Change-Id: I92202cf4ef4554559eb022c4410ee93923edec1b
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
properties"..."
This reverts commit cccd28ca5064a7bd14282cd774ee5bb7126d5845.
c80340 add a CL to tools_r13 that broken the build. It was therefore reverted
with cccd28. This revert was merged into tools_r13 but shouldn't have been.
This reverts the revert to put back the CL in master.
Change-Id: I95fcefedb77596cfdf76269f248b2ae80c85c830
|
|\ |
|
| |
| |
| |
| | |
Change-Id: I7130b7f82dec493fb893acf2527755398104cde3
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The name of the emulated charmap must be passed to the guest
kernel because it is used by the guest input handling code to
load the proper charmap files from the system partition.
This mandates that the name becomes a hardware property for
the sake of supporting proper snapshotting and ui/core separation.
From now on, the charmap name found in a skin will be ignored
completely. This shouldn't be a problem in practice because all
skins used the default name (qwerty2) anyway.
The only reason to change the default value would be if emulating
a guest system image that depends on a different charmap during
emulation.
Change-Id: If571684000b159ae9e9849661c9ff18e19b4cc75
|
| |
| |
| |
| |
| |
| |
| |
| | |
properties". do not merge.
This reverts commit f490aee025b3bab5bcdb065e5fc1b1abecf16b97.
Change-Id: I59a0b2aebc8ab2b788a820a206c6eb4fc6e6e6d2
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The name of the emulated charmap must be passed to the guest
kernel because it is used by the guest input handling code to
load the proper charmap files from the system partition.
This mandates that the name becomes a hardware property for
the sake of supporting proper snapshotting and ui/core separation.
From now on, the charmap name found in a skin will be ignored
completely. This shouldn't be a problem in practice because all
skins used the default name (qwerty2) anyway.
The only reason to change the default value would be if emulating
a guest system image that depends on a different charmap during
emulation.
Change-Id: If571684000b159ae9e9849661c9ff18e19b4cc75
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The new translator has the following benefits:
- faster emulation of ARMv5TE code (through improved JIT)
- proper support for ARMv7 and NEON
- rebuilding the full-eng platform images for ARMv7-A results
in additionnal speed increases (a.k.a. Thumb-2 rocks!).
Note that, as an interesting side effect, NEON machine code is generally
slower than the equivalent C code it is supposed to replace when run inside
the emulator. This can be explained by the fact that for now the translator
simply translates each NEON instruction into a series of sequential host
instructions (and also requires over-head for packing/unpacking/saturation/
etc...).
This change has been tested by running the "full-eng" platform image
rebuilt for ARMv7-A and Neon and using an appropriate kernel image
(prebuilt/android-arm/kernel/kernel-qemu-armv7). The system could boot
and seems to work perfectly. Not a single issue has been experienced
during testing. On a 2.4 GHz Xeon CPU, the image boots in about 25 seconds
(compared to 40 seconds for a vanilla one without this emulator patch).
Thanks to Peter Maydell at Linaro and ARM with his hard work to make this
happen (first in upstream, and now on Android).
This integration is based on the Meego git repository
(git://gitorious.org/qemu-maemo/qemu.git) using the following hash:
7e2d65b0c95c865b1fa6d3d4948e8e822b9ac2fd
On top of which, the following upstream patch has been applied
(with recommendation from Peter):
b7fa9214d8d4f57992c9acd0ccb125c54a095f00
(We chose this repository because it was the closest to the previous
integrate. We will probably use the Linaro ones for future work on this
part of the emulator).
Change-Id: I54837e3d2e908b2380d158411d7a9813630e7e4e
|
|
|
|
|
|
|
| |
This is done to allow for future integration of the
upstream auto-generated "trace.h" header file.
Change-Id: I9d1bba3320d9e8acfd7359540d3ed6695773061e
|
|
|
|
| |
Change-Id: I49e96e2d5ae571849b0b6fef0a30b41ecdee8d23
|
|
|
|
|
|
| |
+ convert to QemuOpts* initialization functions.
Change-Id: I5a245ed2de068af49fdc92de8e3e834d0f49aa27
|
|
|
|
| |
Change-Id: I16103c65ac7b15b2dc58dcc7dd6b3052004aa31a
|
|
|
|
|
|
|
|
|
|
|
| |
Get rid of qemu_timer_new() implementation, and update all
callers to use qemu_timer_new_ms() or qemu_timer_new_ns()
instead.
Rename qemu_new_timer_scale() to qemu_new_timer() to follow
upstream conventions.
Change-Id: Id2c04f8597ec5026e02f87b3e2c5507920eb688e
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds a new virtual hardware device named "goldfish_pipe"
used to implement a very fast communication channel between the
guest system and the emulator.
IMPORTANT: This depends on a special kernel driver, see:
https://review.source.android.com/#change,22496
Usage from the guest is simply the following:
fd = open("/dev/qemu_pipe", O_RDWR);
const char* pipename = "pipe:<name>";
ret = write(fd, pipename, strlen(pipename)+1);
if (ret < 0) {
/* could not connect to service named <name> */
}
/* now you can read()/write()/close() as a normal
* file descriptor to exchange data with the service.
*/
In addition, this implements the following pipe services in the
emulator:
tcp:<port>
tcp:<hostname>:<port>
unix:<path>
opengles
The 'tcp:' and 'unix:' services simply redirect to a TCP or Unix
socket on the host with minimal
The 'opengles' service simply connects to tcp:locahost:22468 for now.
We may change this to be more configurable in the future, but that's
the port number used by the current experimental OpenGL ES hardware
emulation host libraries / programs.
Benchmarking with a simple ping-pong program shows that the
guest <-> emulator can achieve a roundtrip bandwidth of 192 MB/s
(on a 2.7 Ghz Xeon PC).
Using the tcp: service to talk to a ping-pong server listening
on localhost reaches 102 MB/s on the same machine, using a Unix
socket reaches 140 MB/s.
By contrast, using standard sockets in the guest reaches only
3.8 MB/s on the same machine (and requires special privileges
from the application anyway).
|
|\
| |
| |
| | |
warnings with some compilers."
|
| |
| |
| |
| |
| |
| | |
Removes compiler warnings with some compilers.
Change-Id: Idfb14557181e744d07bef6b5d0da734012d2ff42
|
| |
| |
| |
| | |
Change-Id: I7f7f46b6e9526d207f53ad6a9e6baa315bd6af7b
|
| |
| |
| |
| |
| |
| |
| |
| | |
Note that this is also forces tracing support during machine
initialization. Measurements show that this doesn't have any
significant impact on emulation performance.
Change-Id: Ie32352659611531ce65ba2c95c5896e2421f509d
|
|/
|
|
| |
Change-Id: I338334d53fa9bc52c87e9da18341d0cb94fd4269
|
|
|
|
|
|
|
|
|
| |
This reverts commit 2facea6512c84af9285682909f6ac6edcd4130ab
which was not supposed to merge down from tools_R10 to master
since the original reverted commit was a cherry-pick from master
to tools_r10
Change-Id: Ide87af693a0f3239bacaa7313b9262e6e8c46800
|
|\
| |
| |
| | |
Change-Id: I70d9ccafb5912c85baa0711f7e9010bec223efaa
|
| |
| |
| |
| | |
This reverts commit e36a1cc2561b2ce27d3c0a5b3a7b79d4e8531f0e.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Fix touchscreen emulation to return correct min/max coordinate bounds.
This is needed by future changes in the input framework that depend on
touchscreen devices to properly return the min/max bounds of their
absolute pointer coordinates.
Fixes bug 4126574
Change-Id: I1ace56d5433be5d97eb79895cd57b6e442245ca0
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Add hw.keyboard.lid
This is used to control the lid switch virtual device.
+ Add support for a "version <number>" field to the skin format.
This will later be used for skin format #3.
+ Fix a bug where string hw properties didn't work/compile
Change-Id: I27ff3293b439917a51c584a25139264ddb1aff12
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This is needed by future changes in the input framework that depend on
touchscreen devices to properly return the min/max bounds of their
absolute pointer coordinates.
Fixes bug 4126574
Change-Id: I982c0d8e78fc59912e964af958929d2b0718eaa4
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
file.
This only happened for fresh newly-created AVDs. Existing ones used for testing
the emulator already had a cache.img and booted properly.
Change-Id: I87901ea5fb1dc6aa7d2bb3dd191de58d972c386f
|