and run fsck with -f on clean shutdown instead.
With -f, fsck.f2fs always performs a full scan of the /data
partition regardless of whether the partition is clean or not.
The full scan takes more than 2 seconds on volantis-userdebug
and delays the OS boot.
With -a, the command does almost nothing when the partition
is clean and finishes within 20-30ms on volantis-userdebug.
When the partition has an error or its check point has
CP_FSCK_FLAG (aka "need_fsck"), the command does exactly the
same full scan as -f to fix it.
Bug: 21853106
Change-Id: I126263caf34c0f5bb8f5e6794454d4e72526ce38
Using a getenv('OUT') in such a deep down function is a wrong design
choice. Replacing with explicit parameter that may be NULL in case
device specific files can be accessed from /.
Since TARGET_COPY_OUT_SYSTEM may be defined to something different than
system we also ensure that we use a path relative to TARGET_OUT to
compute path to fs_config_* files.
Bug: 21989305
Bug: 22048934
Change-Id: Id91bc183b29beac7379d1117ad83bd3346e6897b
Signed-off-by: Thierry Strudel <tstrudel@google.com>
When the ramdisk is built into the system image, we need to create
those folders during the build.
Change-Id: I40881940c75e379d7599dc1f743b35fbc7cc66e8
If a thread is created while the parent thread is "Background",
then the default timerslack value gets set to the current
timerslack value of the parent (40ms). The default value is
used when transitioning to "Foreground" -- so the effect is that
the timerslack value becomes 40ms regardless of foreground/background.
This does occur intermittently for systemui when creating its
render thread (pretty often on hammerhead and has been seen on
shamu). If this occurs, then some systemui animations like navbar
ripples can wait for up to 40ms to draw a frame when they intended
to wait 3ms -- jank.
This fix is to explicitly set the foreground timerslack to 50us.
A consequence of setting timerslack behind the process' back is
that any custom values for timerslack get lost whenever the thread
has transition between fg/bg.
See Bug: 19398120
Change-Id: Idc259717f62fa2255f8bafbbf88b68c0043f29cf
If /dev/kmsg already exists, it's unnecessary for klog_init()
to create it's own copy. This avoids needing to grant the mknod
permission to everyone who uses kmsg. Typically the only time
/dev/kmsg doesn't exist is before ueventd starts.
(cherrypicked from commit 4d32a486fe)
Bug: 21242418
Change-Id: I0c88d80feca6899fcdbc8c9f2f99448ee0a3422d
If /dev/kmsg already exists, it's unnecessary for klog_init()
to create it's own copy. This avoids needing to grant the mknod
permission to everyone who uses kmsg. Typically the only time
/dev/kmsg doesn't exist is before ueventd starts.
Bug: 21242418
Change-Id: I0c88d80feca6899fcdbc8c9f2f99448ee0a3422d
User specified values of numInts and numFds can overflow
and cause malloc to allocate less than we expect, causing
heap corruption in subsequent operations on the allocation.
Bug: 19334482
Change-Id: I43c75f536ea4c08f14ca12ca6288660fd2d1ec55
Also simplifies the Darwin implementation of gettid, because
apparently libbacktrace had a better way of doing it.
Bug: 19517541
(cherry picked from commit 23f750b068)
Change-Id: I2f888e8ed7a2f5719973786cbcbb399a81867ee9
Accidentally had this all hidden by an #ifndef _WIN32 when I wrote it.
Bug: 19517541
(cherry picked from commit b3a36ca5ee)
Change-Id: Ifbd5d19e506e7313700e2e29a2dae5736e049844
Also simplifies the Darwin implementation of gettid, because
apparently libbacktrace had a better way of doing it.
Change-Id: I9e14f9176b167405a5aaa5b8a0db0327c7126271
Fix broken mips64 build by replacing mips32r2-only android_memset.S.
Use HW-bonded pairs of 64-bit stores to fill 128 bits/cycle.
Rely on HW automatic cache prefetch optimizations.
Software cache prefetching is counterproductive on next mips cores.
New method is coded in C, and also works okay on non-Mips architectures.
Change-Id: Id7153a8fe11538fe25287e101375661b0e99e2a2
cutils/endian.h provides the helpers defined in endian(3), either by
pulling in the OS's built-in endian.h (where available) or recreating
them using GCC builtins.
Change-Id: Ic8965f67e1efdc03f884dbe6b7fe0276f840e4fc
Signed-off-by: Greg Hackmann <ghackmann@google.com>
Intention is to read from ${OUT}/system/etc/fs_config_dirs and
${OUT}/system/etc/fs_config_files on host. And
/system/etc/fs_config_dirs and /system/etc/config_files
on target systems.
Bug: 19908228
Change-Id: Ieaedd99e88c7f7f717878d9344c387ddf335ccbb
Intention is to read from ${OUT}/system/etc/fs_config_dirs and
${OUT}/system/etc/fs_config_files on host. And
/system/etc/fs_config_dirs and /system/etc/config_files
on target systems.
Bug: 19908228
Change-Id: I0966a94f79a3fae4f7325056c701ea355370f9df
Several people have been independently confused about when this gets
built and why over the past couple of days.
Change-Id: I2d4a02f66e24c0734327585b7d27e50a344b1e94
This also fixes the bug where we were always testing against the fake
strlcpy we provide for glibc/Windows rather than the Android one.
This also removes some unnecessary library dependencies.
This also builds all the cutils tests for the host (static and dynamic).
Change-Id: Icd129d5b025c0ca801be5acf31a54ecd88608df9
Also make important events in init's life NOTICE rather than INFO,
and ensure that NOTICE events actually make it to the kernel log.
Also fix the logging so that if you have a printf format string
error, the compiler now catches it.
Also give messages from init, ueventd, and watchdogd distinct tags.
(Previously they'd all call themselves "init", and dmesg doesn't
include pids, so you couldn't untangle them.)
Also include the tag in SELinux messages.
Bug: 19544788
Change-Id: Ica6daea065bfdb80155c52c0b06f346a7df208fe
This change defines atrace_* functions for the host that act as
no-ops, which makes it easier to add tracing in ART.
Change-Id: I89397e83986686a2b6a6f245c25017eb379081b1
This was supposedly used to enable logging when !HAVE_ANDROID_OS
but it's only used in a file that's target specific.
Change-Id: Id83f2597e48a66b4821fc3b1237e212872b909fb
On 64 bit systems, calls to dump_backtrace_to_file or dump_tombstone
try and directly contact the correct debuggerd (32 bit vs 64 bit)
by reading the elf information for the executable.
Unfortunately, system_server makes a call to dump_backtrace_to_file
and it doesn't have permissions to read the executable data, so it
defaults to always contacting the 64 bit debuggerd.
This CL changes the code so that all dump requests go to the 64 bit
debuggerd, which reads the elf information and redirects requests for
32 bit processes to the 32 bit debuggerd.
Testing:
- Forced the watchdog code in system_server to dump stacks and
verified that all native stacks are dumped correctly.
- Verified that dumpstate and bugreport still properly dump the native
processes on a 64 bit and 32 bit system.
- Intentionally forced the 64 bit to 32 bit redirect to write only a
byte at a time and verified there are no errors, and no dropped data.
- Used debuggerd and debuggerd64 to dump 32 bit and 64 bit processes
seemlessly.
- Used debuggerd on a 32 bit system to dump native stacks.
Bug: https://code.google.com/p/android/issues/detail?id=97024
Change-Id: Ie01945153bdc1c4ded696c7334b61d58575314d1
Under some unknown circumstances, debuggerd could become unresponsive.
If you try and take a bugreport during this time, it will hang forever.
Adding functions that have a timeout will allow dumpstate to stop if
dumping is taking too long.
Bug: 18766581
(cherry picked from commit 5f2ff6a910)
Change-Id: I39e8e9c60209e3ef9efac795fedb8e1edce2bd3e
Packets captured and logged by the NFLOG target are unicast, so
extend to catch and decode them. To avoid escaping issues, the raw
contents are passed around as hex strings.
Bug: 18335678
Change-Id: Ib7299500baa00080a1f000f9da843eb527363353
Under some unknown circumstances, debuggerd could become unresponsive.
If you try and take a bugreport during this time, it will hang forever.
Adding functions that have a timeout will allow dumpstate to stop if
dumping is taking too long.
Bug: 18766581
Change-Id: I85053b8dcfe6224e2b64b4d8f7f2ef448b3cda34
This means that code that uses libcutils no longer has to ensure that
it's set ANDROID_SMP in the calling code's Android.mk for this to
function correctly.
Change-Id: I80c7ff170cd621106f34d6b74689d6b4f03e4eb7
The static libcutils tests cannot be built when using libc++ because
there are multiple symbol definition errors between libc++ and libgcc.
Bug: 18389856
Change-Id: I6f41c561f97b3a37477f844e9abf6551524a0fa3
This should not be committed until win_sdk and aarch64 builds are
fixed in the presence of this CL.
This reverts commit 2789faabfa.
We additionally remove uniprocessor support from the earlier CL,
thus avoiding a potential compiler code reordering issue.
Change-Id: I7207a5ca2efa907a6f757f172d7090a62b2311fe
The build system is still linking both libgcc and libcompiler-rt,
which is causing duplicate symbol errors on some architectures. This
requires a fix in the build system, so reverting for now.
This reverts commit e3cccbfd47.
When atrace_* functions are inlined,
the rarely used 1024-byte buffers are allocated on stack.
BUG: 17444504
Change-Id: I773512aeb70e8b79f3803c6d59cba064d2aa65b6
All kernel services will now be in the same cgroup as
foreground applications. This will now make kernel threads
not implicitly higher priority than android foreground
services.
Bug 17681097
Change-Id: I28e81c7aade50428d5395df86f00ce01c1e7af02
Replace atomic-inl.h with a file that just includes atomic.h.
Remove platform specific implementations.
Change-Id: If16d74fbe0af7836ed8c1296c17e13a2d0d20f64
On 64 bit systems, calling dump_backtrace_to_file will automatically
call debuggerd64. If the process to dump is actually 32 bit, this
creates an unrecognizable dump backtrace. Modify the code to check the
type of the process and connect to the appropriate debuggerd process.
This change refactors both the tombstone and backtrace functionality to
allow both to work properly on 64 bit systems when dealing with mixed
processes.
Bug: 17487122
(cherry picked from commit a9fa7b87f1)
Change-Id: I3c9e0212c8720877a6af092071a3695df2a36df8
On 64 bit systems, calling dump_backtrace_to_file will automatically
call debuggerd64. If the process to dump is actually 32 bit, this
creates an unrecognizable dump backtrace. Modify the code to check the
type of the process and connect to the appropriate debuggerd process.
This change refactors both the tombstone and backtrace functionality to
allow both to work properly on 64 bit systems when dealing with mixed
processes.
Bug: 17487122
Change-Id: Icf123a6f4508b1aeec073663aa1a0ceae5380aa1
Use expected inline behavior with clang.
GCC defaults to -std=gnu90, giving C89 inline semantics with GNU
extensions. Clang defaults to C99. Explicitly use gnu90.
Mark an unused parameter as __unused.
Fix some incorrect casts.
Change-Id: I05b95585d5e3688eda71769b63b6b8a9237bcaf4
