logd suffers major performance degradation when persistent (blocking)
client reader connects to it (e.g. logcat). The root cause of the
degradation is that each time when reader is notified of the arrival
of new log entries, the reader commences its search for the new entries
from the beginning of the linked list (oldest entries first).
This commit alters the search to start from the end of the linked list
and work backwards. This dramatically decreases logd CPU consumption
when blocking reader is connected, and increases the maximum logging
throughput (before the logs start getting lost) by a factor ~ 20.
Change-Id: Ib60955ce05544e52a8b24acc3dcf5863e1e39c5c
- create a structure to depict the private header
expected at logd end of socket.
- utilize this new structure instead of unscalable
byte stream technique used to unpack in logd.
Change-Id: I2d0e5c3531c279f2dc1fbd74807210ff8d804de0
When mounting fstab entries it is practical to be able to specify
for example LABEL=data instead of a specific block device.
This is particularly important for the new Android Emulator code base,
which uses virtio block devices for the various partitions
(system,data,cache), because there is no defined ordering between the
way the filesystems are specified to the emulator and how they are
enumerated by the android kernel as /dev/vdX.
Change-Id: I7aef95f12e8f7b02ac2e33804ba7897fdcb9ad7f
Signed-off-by: Christoffer Dall <christoffer.dall@linaro.org>
Due to a typo, the --no-rebind option never worked (it always failed).
The root of the problem was that the client was sending on the wire
a command like:
host:forward:norebind::tcp:<port>;tcp:<port>
^^
Instead of:
host:forward:norebind:tcp:<port>;tcp:<port>
^
Note the erroneous double-column.
The fix is local to the adb client and thus doesn't require a new
version of the server or guest adbd on the device-side.
This also fixes 'adb reverse --no-rebind'.
See https://code.google.com/p/chromium/issues/detail?id=451109
Change-Id: I680fd432b5470072f6a9968ca32a7f90c600ac68
LOCAL_FORCE_STATIC_EXECUTABLE was not working properly for clang
because -Bstatic was ignored by clang. We can now enable clang
for init after the -static flag is added to link static executable
files in build/core/definitions.mk.
BUG: 18008984
Change-Id: I3f361b83c1e0b313914603dff33fd090cd3b116a
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
When mounting fstab entries it is practical to be able to specify
for example LABEL=data instead of a specific block device.
This is particularly important for the new Android Emulator code base,
which uses virtio block devices for the various partitions
(system,data,cache), because there is no defined ordering between the
way the filesystems are specified to the emulator and how they are
enumerated by the android kernel as /dev/vdX.
Change-Id: I12c3db0ba6a515dc8e917e0349afd257888d3aef
Signed-off-by: Christoffer Dall <christoffer.dall@linaro.org>
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
This works around a bug on on 64 bit kernels + sdcard daemons
where we were using memory addresses as inode numbers.
bug: 19012244
(cherry picked from commit faa0935ffb)
Change-Id: Idbf9e285e507e702e04e7461a10153df68ef2322
In a scenario in which an on-line (blocking) client is running and
a clean is attempted (logcat -c), the following can be observed:
1) the on-line logger seems to freeze
2) any other clear attempt will have no effect
What is actually happening:
In this case prune function will "instruct" the oldest timeEntry
to skip a huge number (very close to ULONG_MAX) of messages, this
being the cause of 1.
Since the consumer thread will skip all the log entries, mStart
updating will also be skipped. So a new cleaning attempt will have
the same oldest entry, nothing will be done.
Fix description:
a. keep a separated skipAhead count for individual log buffers (log_id_t)
LogTimeEntry::LogTimeEntry
LogTimeEntry::FilterSecondPass
LogTimeEntry::skipAhead
LogTimeEntry::riggerSkip_Locked
b. update LogTimeEntry::mStart even if the current message is skipped
LogTimeEntry::FilterSecondPass
c. while pruning, only take into account the LogTimeEntrys that are monitoring
the log_id in question, and provide a public method of checking this.
LogTimeEntry::isWatching
LogTimeEntry::FilterFirstPass
LogTimeEntry::FilterSecondPass
d. Reset the skip cont befor the client thtread starts to sleep, at this point
we should be up to date.
LogTimeEntry::cleanSkip_Locked
LogTimeEntry::threadStart
Change-Id: I1b369dc5b02476e633e52578266a644e37e188a5
Signed-off-by: TraianX Schiau <traianx.schiau@intel.com>
