| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch saves the current state of the SMSC 91C111 ethernet
emulation. The restore function will restore all buffers, thus
preventing segfaults when guest code tries to access data that was
there when state was saved. However, no effort is made to
re-establish connections. This is probably impossible anyway, since
we have no way to generically make the remote end pick up where we
left when state was suspended.
From the perspective of the guest, resuming is therefore equivalent
to a temporary network failure, in which all open connections are
lost. After the restore is complete, new connections can be made
as though nothing happened.
Change-Id: Ibbbabc685ef296e675d4018996f703e07d9aad6a
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This brings the code for saving and restoring the vm clock up to
date with the current codebase. It also includes a workaround for an
issue that occurred when the state of a goldfish timer was restored.
For unexplained reasons, the compiler generated code to copy the
least significant 32 bits only when reading a 64 bit number,
dropping the remaining 32 bits.
In order to facilitate debugging, the variables in the goldfish
timer code have been suffixed with their units ("ns" for nanoseconds
or "tks" for ticks).
|
|
|
|
| |
Change-Id: Ide3914fa52f62f618d39ac20f02bce8185a6805a
|
|
|
|
|
|
| |
+ new document under docs/DISPLAY-STATE.TXT to explain what's happening.
Change-Id: Ia0d233377266212da49af932c7528f46f5feb92d
|
|
|
|
| |
Change-Id: Iea5edfbae0caff45161c814d631d35d2e6d61d1c
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Preparation for future UI frontend/backend separation.
This is done to ensure that the code under android/skin/ does not depend
on any QEMU-specific header. We achieve this by adding a new abstract header
"user-events.h" and one QEMU-specific implementations for the functions
defined here.
This also modifies console.h and vl-android.c to make them closer to
upstream (by removing Android-specific changes).
+ fix Makefile.android to always build SDL from sources in standalone mode.
Change-Id: I0d152741e7bb2c9cd283f5c35bd054385c7c1eb3
|
|
|
|
| |
Change-Id: I24acbdebe58d207352548f54dda1abf5be01e7d4
|
|
|
|
| |
Change-Id: I41aa30fff127338665a6a32fa66dcd43dd3bfe8b
|
|
|
|
| |
Change-Id: I5c51f54a7fe2ea702420429bbf0c789ed6d8c534
|
|
|
|
| |
Change-Id: Ica9022695d83fb48a8c25fdb1e1f0dc1c63747ff
|
|
|
|
|
|
|
|
|
|
|
| |
The previous patch modifies SD Card hardware emulation in a way that is
not backwards compatible with previous SDK platform releases. This has
the unfortunate effect of making SD Card emulation not work properly
when running existing AVDs.
Reverting the patch until we get a better one.
Change-Id: I82dac87709d88aa3bd292bf2ed0008ae74d95853
|
| |
|
|
|
|
| |
Change-id: Ibce845d0
|
|
|
|
|
|
|
|
| |
Signed-off-by: San Mehat <san@google.com>
qemu: goldfish: Fix compile error introduced in refactoring
Signed-off-by: San Mehat <san@google.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add commandline support for an additional SD card, and plumb support to
the MMC emulation layer. This patch also reworks the way the hw emulation
layer registers devices. Previously, the mmc virtual hardware was only
created if the sdcard image existed. Now, two virtual MMC controllers are
*always* created as on real hardware. When a request is made to one
of our controllers which has no image file bound to it, a new interrupt
status indicating a command timeout is sent to the guest driver, as
is standard with MMC controllers. This patch also sets the stage for
supporting hot-add / hot-remove.
Signed-off-by: San Mehat <san@google.com>
qemu: android: Integrate card detect into the virtual mmc hardware and
wire it up to the STATE_CHANGE irq status
Signed-off-by: San Mehat <san@google.com>
fixups from review - mmc now works with legacy
drivers
Signed-off-by: San Mehat <san@google.com>
|
|
|
|
|
| |
The minimum size is now 9MB. This appears to be a limitation
of the emulated MMC hardware code and/or kernel device driver.
|
|
|
|
| |
Change-Id: I67152761977b330e3f4c1bbe3d8eb99e48e2e7e7
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This change integrates many changes from the upstream QEMU sources.
Its main purpose is to enable correct ARMv6 and ARMv7 support to the
Android emulator. Due to the nature of the upstream code base, this
unfortunately also required changes to many other parts of the source.
Note that to ensure easier integrations in the future, some source files
and directories that have heavy Android-specific customization have been
renamed with an -android suffix. The original files are still there for
easier integration tracking, but *never* compiled. For example:
net.c net-android.c
qemu-char.c qemu-char-android.c
slirp/ slirp-android/
etc...
Tested on linux-x86, darwin-x86 and windows host machines.
|
|\
| |
| |
| |
| | |
* changes:
Add support for new trace records for native (JNI) calls.
|
| |
| |
| |
| |
| |
| | |
This change adds new trace record types for entering (and exiting)
native code from Java code. There is a corresponding change in the
Dalvik interpreter to generate these trace records.
|
|/ |
|
|
|
|
|
|
|
|
|
|
|
| |
This is part of a larger change to add support for tracing Java methods.
There is also a kernel change and a small change to the Dalvik interpreter
that will be checked in separately.
There used to be support for tracing Java methods but it relied on trapping
every store and checking if the store address matched a special "magic"
region (and that stopped working because we can no longer trap on loads
and stores). The new approach uses a memory-mapped page to catch stores to
just that page.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|