Age | Commit message (Collapse) | Author |
|
There's just no way to do this piece correctly:
int dummy = 0;
printf("corrupting our stack...\n");
*(volatile long long*)&dummy = 0;
Change-Id: If02ff71b3c23c3d8118694f6edba0ecf904baf8a
Signed-off-by: Bernhard Rosenkraenzer <Bernhard.Rosenkranzer@linaro.org>
|
|
|
|
Change-Id: I2cd2f1a1e51f6e3d1ebbde4d5442be68a218b33d
|
|
|
|
Change-Id: Ieef12b36b8519ff689ab5a81447b1b9bf4806171
|
|
|
|
This change makes enabling clock tracing and bus bandwidth tracing depend on
whether the device's kernel supports these features. If it's not supported it
will be skipped.
Change-Id: I51d57299693635b0ecebb75de5dccd5b62d47be3
|
|
|
|
Change-Id: I4e1764b51225436fc2b3d61454ee822f511a1a7d
|
|
|
|
Bug report needs to be launched for the current user.
Change-Id: I2b8f812ede43fc36c76deab544bacf88dba8fefd
|
|
Bug: 7152673
Change-Id: Iae7b298721d2b6690ce9b57245400702fb5d0c4c
|
|
Bug: 7152673
Change-Id: Ibff4503ef21816bfe936eeba89183d1647c85043
|
|
Bug: 7152673
Change-Id: Iae7b298721d2b6690ce9b57245400702fb5d0c4c
|
|
Bug: 7152673
Change-Id: Ibff4503ef21816bfe936eeba89183d1647c85043
|
|
|
|
Add some null checks in case that SendBug is invoked while system
server is in fact not running, e.g. during a runtime restart. So
that we don't hit NPE in SendBug, looking suspiciously like the
cause rather than a consequence of the real issue.
Bug: 7097125
Change-Id: I82676ab936fe19a781c32f949e6e429eb4045d63
|
|
Since it forces user to make selection, not offering an option to
remember the choice. Remove the new task flag, also adds an
explicit default category
Change-Id: I3834a8fa27782d14989e510c0756d36aaeeb2bab
|
|
Since it forces user to make selection, not offering an option to
remember the choice. Remove the new task flag, also adds an
explicit default category
Change-Id: I3834a8fa27782d14989e510c0756d36aaeeb2bab
|
|
|
|
|
|
Drop caches before read tests.
Fix random read tests to read area big enough to prevent
the block cache from filling with the data being read.
Test with the read-ahead set to 4K, and at the default value.
Change-Id: Ifffa07fe4c89b6e103ce4fd71d0fa9881fed0dc7
|
|
|
|
|
|
This enables:
1. an explicit title for intent chooser, so that the user is
aware that chosen action is for sending out bugreport
2. the original intent has new task flag that causes
ResolveActivity to finish() when stop()ped. This means that
if intent is fired when screen is off (common scenario now
because bugreport takes a long time), the chooser will not be
shown
Bug: 6864596
Change-Id: I6cce2d1390c8ac24b15d26e9ba80a845222a1c18
|
|
This reverts commit 9ebc639a313d143ade1293cb5932b446d5e623d5.
|
|
* commit '7793c09ec42bf1a01a5f45f0e723f9a36fcfd9d7':
Support for multiple modules with the same name is going away.
Support for multiple modules with the same name is going away.
|
|
* changes:
Support for multiple modules with the same name is going away.
Support for multiple modules with the same name is going away.
|
|
|
|
When running on multi-user devices, we need to save bugreport data
to a write path, and launch intents using a different read path.
Bug: 6925012
Change-Id: I32773b733e8a4dd75a30f5b9b05424c4238243af
|
|
Change-Id: I4e94b8b9cc7c88e07e8a09abc26a3e7d495ac5be
|
|
Change-Id: Iae48d4890aa7266ca7b57b3b9a8712d2d33d4909
|
|
|
|
|
|
Also tweak the script a little bit. This is only for testing,
and doesn't ship in any product.
Change-Id: I9506d6babddef91fc6d9c0083064e0ef8a2d70a8
|
|
* use a property to indicate a preferred domain in case multiple
email addresses are detected; otherwise just pick the first
one that matches email address regex
* don't fire the intent if no email account is configured, there
have been reports on confusing behavior when capturing
bugreport for setup wizard
Bug: 6720467
Change-Id: Ib9b0b8a09e4e565b2531f3ec53723cd46186596f
|
|
* use a property to indicate a preferred domain in case multiple
email addresses are detected; otherwise just pick the first
one that matches email address regex
* don't fire the intent if no email account is configured, there
have been reports on confusing behavior when capturing
bugreport for setup wizard
Bug: 6720467
Change-Id: Ib9b0b8a09e4e565b2531f3ec53723cd46186596f
|
|
code."
* commit '332e8babe350433b320efc9484f15a5dbdd8c68f':
Avoid dereferencing NULL pointer with memcmp in the test code.
|
|
|
|
|
|
Generates a directory tree with requested depth, also writing files
at each leaf to add width. Then it times traversal of entire tree,
which can reveal impact of VFS caching inside kernel.
Also specify mode_t when creating files.
Bug: 6925012
Change-Id: I65370d6fd0b9777e959be3a0867f27ce22ee4644
|
|
- send_bug requires platform API support which is
enabled in PDK fusion build with platform.zip from platform-java build
Change-Id: I3285f595eb2a310a34770d40fa0b7c606669f1bf
|
|
* commit '983ce5738cee0bf78b87d6274fe1387129768302':
Fix SELinux labeling for system.img
|
|
|
|
* commit 'fc3e87f70ba3c896ea5a85df3648623f207f88cc':
system/extras/ext4_utils: move libsparse to system/core
|
|
* commit 'cab036827411791e8f197a0194f95faad6f12cec':
make_ext4fs: allow sparse gzip files
libsparse: make API sane
Refactor sparse file support into libsparse
|
|
|
|
* changes:
make_ext4fs: allow sparse gzip files
libsparse: make API sane
Refactor sparse file support into libsparse
|
|
Delete libsparse from ext4_utils, move it to system/core/libsparse in
preparation for linking tools in system/core against it.
Change-Id: I3545c2e0ad7122f99bbc998ab0e85ca1e65d5add
|
|
libsparse can handle sparse gzipped files, allow -s and -z together.
Change-Id: I32c336269375135af6be5d00a55604f8bf7619a3
|