x86 just bit me. I did manage to build mips today, but I assume it
can't be long before we hit the limit there too if every other target
has hit it already.
Change-Id: I28dfe3b4f9565cb79e0bf6b0ffc55a9d6e64a9b0
Basically, allow access of qemu_device where gpu_device is allowed, for the
case when the emulator runs with OpenGL/ES emulation. Most noticably,
surfaceflinger crashes without qemu_device access.
Bug: 15052949
Change-Id: Ib891365a6d503309bced64e2512c4d8f29d9a07e
When starting the emulator, the system console writes entries
to /dev/ttyS2. We need to allow the writes, otherwise this generates
denials when you run "emulator -verbose -logcat '*:v' -show-kernel"
Addresses the following denial:
type=1400 audit(1395076594.320:446): avc: denied { read write } for pid=5600 comm="sh" path="/dev/ttyS2" dev="tmpfs" ino=1487 scontext=u:r:shell:s0 tcontext=u:object_r:serial_device:s0 tclass=chr_file
Bug: 13506702
Change-Id: I3729537cabb0bf8e8b2905d3def43a293bb1081f
The qemud and /dev/qemu_pipe policy bits copied to generic
and generic_x86 by I620d4aef84a5d4565abb1695db54ce1653612bce
are required for generic_mips as well. In testing, we
further saw other denials for generic_mips that correspond
exactly to what is already allowed in the generic sepolicy, so
just inherit the sepolicy files from generic for now.
We could do likewise for the generic_x86 sepolicy for the files that are
identical with generic if desired, but that is not done by this change.
The generic_x86 sepolicy was missing a rule for /sys/qemu_trace
moved to the generic sepolicy by the prior change, so fix that omission.
The generic*64 variants will need something similar, either by inheriting
from one of the existing sepolicy directories as in the MIPS
case or by forking their own copies as in the x86 case.
Change-Id: Iec7c8825734a3f96f7db8ae1d10dce1f30b22bdf
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
Following the new vibrator hardware module implementation, the
goldfish vibrator module is no longer integrated in the default
vibrator implementation and therefore needs to be explicitly built
for the emulator.
This patch is related to others in:
- hardware/libhardware
- hardware/libhardware_legacy
- frameworks/base
- device/generic/goldfish
Change-Id: I844279f5535289f079d412fdc44c5cb3c9c1130c
Signed-off-by: David Wagner <david.wagner@intel.com>
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>
Signed-off-by: Jack Ren <jack.ren@intel.com>
Author-tracking-BZ: 94611
Since 558477c0b2a5a7db679c1ad7551d43e6364d1df2 in
external/wpa_supplicant_8 we don't need those modules if
WPA_SUPPLICANT_VERSION is not set by the BoardConfig.mk.
Actually the generic devices don't need them.
Change-Id: Iff7c1fbf483bd16d5f898836c4e7401d56def80f
Conflicts:
target/product/core_tiny.mk
Since 558477c0b2a5a7db679c1ad7551d43e6364d1df2 in
external/wpa_supplicant_8 we don't have WPA_SUPPLICANT_VERSION default
to VER_0_8_X.
This fixes the SDK build.
Change-Id: Ib7a8e10705dc6b987b60b1308405e89eaa2a47c9
Change TARGET_ARCH_VARIANT from x86-atom to x86, for the reason
that atom-specific instruciton like "movbe" isn't "generic" to
x86 Android devices and x86 host running emulator-x86, which may
crash non-atom host in VT because "movbe" isn't supported natively.
Also revert previous fix in TARGET_linux-x86.mk which conservatively
disable "movbe" across the board
Change-Id: Ief93a3585566ffae558fcdc29741e6213a048d7d
That can gain better memory performance than 8-byte alignment
Change-Id: Ibcc5bf7b96ed10166038357dd0c86c9171768f3e
Signed-off-by: Jack Ren <jack.ren@intel.com>
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>
Add boot property enabling ADBD over QEMU
Emulator and system image now support ADBD communication over QEMUD pipe rather
than over a TCP port forwarding. However, emulator has to know ahead of time
(before system starts booting) whether or not the system image supports ADBD
over pipe to properly setup the communication. For that, we introduce a boot
property "ro.adb.qemu" that is readable by the emulator early enough for the
proper ADB communication setup.
Change-Id: I978489c5acf46177b520e775d745bcc78f469837
This patch re-enables dex preoptimization in emulator-specific
builds. This considerably speeds up the first-boot of SDK AVDs
since the dex optimization pass is performed during the build
instead.
This was enabled in a previous patch, then reverted due to a
bug in the build system that has been fixed since.
This patch doesn't affect any device build.
Change-Id: I3be8a9c7f51a86029fde30b7d37584cb40b4db69
The file is required by the camera framework to properly set it up to work with
emulated camera devices.
Change-Id: I39686841030d21ef23775972aa4ee0ecd7a91c02
This patch enables the build of OpenGLES emulation libraries
(both in the guest and the host) for any emulator-specific
product config.
+ update core/main.mk to ensure that the right path is parsed
when building the Windows SDK
Change-Id: Ifde047f8c0ef922504a7286c8c5b57f7c79c71f8
This enables the Dex pre-optimization for emulator-specific builds.
The idea is to speed up the initial boot sequence of an SDK AVD by
avoiding the costly on-device preopt pass that is performed otherwise
the first time the AVD is launched.
Benchmarking shows reductions in first boot time from 1m30s to 40s
on a 2.4GHz Xeon when emulating the AOSP master ToT.
Change-Id: I67a9ce9791220c202817864c6e83dc8fd9cbc027
Switch the emulator builds to ash to make sure it still
works, in case we find that we need to go back to using ash.
Change-Id: Id5641940b63cc4225011be1c375f4c86ec4d21cc
Change-Id: I1fc3cb491d593d3e28f90278cfc8de5d77852932
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
Change-Id: I1fc3cb491d593d3e28f90278cfc8de5d77852932
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
This patch removes support for prelinking from the build system. By now, the
prelinker has outlived its usefulness for several reasons. Firstly, the
speedup that it afforded in the early days of Android is now nullified by the
speed of hardware, as well as by the presence of Zygote. Secondly, the space
savings that come with prelinking (measued at 17MB on a recent honeycomb
stingray build) are no longer important either. Thirdly, prelinking reduces
the effectiveness of Address-Space-Layout Randomization. Finally, since it is
not part of the gcc suite, the prelinker needs to be maintained separately.
The patch deletes apriori, soslim, lsd, isprelinked, and iself from the source
tree. It also removes the prelink map.
LOCAL_PRELINK_MODULE becomes a no-op. Individual Android.mk will get cleaned
separately. Support for prelinking will have to be removed from the recovery
code and from the dynamic loader as well.
Change-Id: I5839c9c25f7772d5183eedfe20ab924f2a7cd411
It was defaulting to 128M which was slightly too small for all the
applications. Increasing to 150M fits the resulting image.
Change-Id: I55957c54fbdb77090487c08ed3948988621978b3
Add in Makefiles and support files for x86 builds
Based on changes by: wonjong.lee <wonjong.lee@windriver.com>
Additional changes by: Mark Gross <mark.gross@intel.com>
Additional changes by: Bruce Beare <brucex.j.beare@intel.com>
Change-Id: Id087411b282b1b68a0251e7f9cbbdde183549787
Signed-off-by: Bruce Beare <brucex.j.beare@intel.com>
