0xdroid:system_core.git
6 years agolibpixelflinger: Add ARM NEON optimized scanline_t32cb16 leb-0xdroid
Jim Huang [Thu, 9 Jun 2011 08:44:36 +0000 (16:44 +0800)]
libpixelflinger: Add ARM NEON optimized scanline_t32cb16

Reference benchmark results on Beagleboard (TI OMAP353x) at 500 MHz:
    scanline_t32cb16_c memory bandwidth: 31.63 MB/s
    scanline_t32cb16_neon memory bandwidth: 147.69 MB/s

It can dramatically improve the performance of boot animation.

6 years agoNeon implementation of t32cb16blend function 0x7-leb-gingerbread
Martyn Capewell [Sun, 29 May 2011 14:23:22 +0000 (22:23 +0800)]
Neon implementation of t32cb16blend function

16-pixel per pass SIMD implementation, little endian only. Based on arm
implementation, guards for selection only if TARGET_ARCH_VERSION is armv7-a,
__ARM_NEON__ is defined (this is supplied by GCC when -mfpu=neon is passed to
the compiler) and the system is little endian.

Tested on ZoomII using soft-keyboard and web browser applications.

It's possible for a system to support ARMv7-a, but not have Neon available -
this is the reason for 'optimistically' including the file at the Makefile
level but using __ARM_NEON__ to actually select the code for compilation.

Uses ARCH_ARM rather than TARGET_ARCH to select NEON features
at build on a per-capability rather than per-arch

Signed-off-by: Jim Huang <jserv@0xlab.org>
6 years agolibpixelflinger: Build NEON optimized routine per target configuration
Jim Huang [Thu, 7 Apr 2011 19:45:08 +0000 (03:45 +0800)]
libpixelflinger: Build NEON optimized routine per target configuration

The variable TARGET_ARCH_VERSION variable has been renamed to
TARGET_ARCH_VARIANT, and armv7-a-neon is the exact target enables
ARM NEON instructions.

Also, do include header <machine/cpu-features.h> to get ARM optimized
routines compiled.

Change-Id: Id72b3cf9bbd854e84d0111305a5d9e97dff573f1

Signed-off-by: Jim Huang <jserv@0xlab.org>
6 years agoAuto reconnects to remote board if ADBHOST is set
Jim Huang [Tue, 27 Oct 2009 09:51:30 +0000 (17:51 +0800)]
Auto reconnects to remote board if ADBHOST is set

commit 5557302556bd0e3c6607accfec9d145e332ee4f9
Author: Nikita Shulga <malfet@embeddedalley.com>
Date:   Mon Jul 27 04:34:02 2009 -0400

    Auto reconnects to remote board if ADBHOST is set

6 years agoAdded 0xlab text bootlogo
Jim Huang [Sat, 12 Sep 2009 18:25:04 +0000 (02:25 +0800)]
Added 0xlab text bootlogo

The text was generated by 'figlet' tool.

Signed-off-by: Tick Chen <tick@0xlab.org>
6 years agolibpixelflinger: Enable CLZ instruction generation
Jim Huang [Thu, 26 May 2011 03:29:45 +0000 (11:29 +0800)]
libpixelflinger: Enable CLZ instruction generation

6 years agoFix ADB daemon file descriptor leak
Teddie Stenvi [Mon, 15 Feb 2010 11:20:44 +0000 (12:20 +0100)]
Fix ADB daemon file descriptor leak

The ADB (Android Debug Bridge) daemon suffers from a file descriptor leak.
When transfering file descriptors, opened by the adbd daemon, to a JDWP
process for debugging through sendmsg, the file descriptor is duplicated
by the kernel and a copy is transfered to the JDWP process.
Adbd neglects to close it's copy of the descriptor after the sendmsg call
which causes a file descriptor leak in the process.
Eventually this causes the daemon to hit its maximum roof of opened file
descriptors thus making it impossible for the process to serve any new
commands from the adb server.
Jdwp_service.c has been modified to close the file descriptor after the
send message call has succeded.
Also, the socket used for transfering the file descriptor is set to blocking
mode before the sendmsg call since the adbd processes needs to know when the
duplication has been completed by the kernel.
The flags are reset to its original state after the sendmsg call.

Change-Id: I76b15460301ac01e432f5ce8678ff58c658c16f2
Signed-off-by: Christian Bejram <christian.bejram@stericsson.com>
6 years agoBuild ash as ash with shell_ash tag; install sh symlink
Thorsten Glaser [Mon, 7 Jun 2010 08:26:37 +0000 (10:26 +0200)]
Build ash as ash with shell_ash tag; install sh symlink

Signed-off-by: Thorsten Glaser <tg@mirbsd.org>
6 years agoDon't alias 'stop' to 'kill -STOP'
Jeff Hamilton [Fri, 10 Sep 2010 15:46:06 +0000 (10:46 -0500)]
Don't alias 'stop' to 'kill -STOP'

Android has already has a stop command used
to stop the main runtime and the alias
interferes with testing tools that expect
stop to kill the runtime.

Change-Id: I02b7efb9203dc39e97f63eb702a54ff79935b316

6 years agoAdd mksh from CVS 2010/08/24 as system/core/mksh module
Thorsten Glaser [Tue, 24 Aug 2010 16:21:37 +0000 (18:21 +0200)]
Add mksh from CVS 2010/08/24 as system/core/mksh module

Both shells (ash from system/core/sh, and mksh) are built by
default but only the one where $(TARGET_SHELL) is set to is
actually installed (the shell and the mkshrc configuration
file are tagged shell_mksh for this to work).

Signed-off-by: Thorsten Glaser <tg@mirbsd.org>
6 years agolibpixelflinger: Apply expensive-optimizations to cflags
Jim Huang [Thu, 26 May 2011 03:29:44 +0000 (11:29 +0800)]
libpixelflinger: Apply expensive-optimizations to cflags

GCC's -fexpensive-optimizations is known to improve the Draw Canvas
bench slightly.

6 years agocodeflinger: Spill some register for filter32
Kan-Ru Chen [Thu, 26 May 2011 03:29:44 +0000 (11:29 +0800)]
codeflinger: Spill some register for filter32

6 years agoinit: Let console applications have a controlling tty
Jim Huang [Thu, 26 May 2011 03:29:44 +0000 (11:29 +0800)]
init: Let console applications have a controlling tty

While boot paramter "console" was not assigned, this change still
allows terminating console application within a shell by Ctrl-C.

6 years agodebuggerd: Enable ARM_HAVE_VFP and fix vfp.S for arm{v5,v6}-vfp
Jim Huang [Thu, 26 May 2011 03:29:44 +0000 (11:29 +0800)]
debuggerd: Enable ARM_HAVE_VFP and fix vfp.S for arm{v5,v6}-vfp

6 years agocodeflinger: Correct misleading comment
Kan-Ru Chen [Thu, 26 May 2011 03:29:43 +0000 (11:29 +0800)]
codeflinger: Correct misleading comment

6 years agolibpixelflinger: Don't saturate when alpha blending mode was used
Jim Huang [Thu, 26 May 2011 03:29:43 +0000 (11:29 +0800)]
libpixelflinger: Don't saturate when alpha blending mode was used

For performance on the hardware lacking of GPU we would like to avoid
saturating the color component.

6 years agolibpixelflinger: Sacrifice premultilied blending correctness for performance
Jim Huang [Thu, 26 May 2011 03:29:43 +0000 (11:29 +0800)]
libpixelflinger: Sacrifice premultilied blending correctness for performance

In theory this is not correct, we should saturate here.  However, this
mode is often used for displaying alpha-premultiplied graphics,
In which case, saturation is not necessary.

This is a case, where we sacrifice correctness for performance. we
should probably have some heuristics.

6 years agolibpixelflinger: ARMv6 specific objects are not used. Removed
Jim Huang [Thu, 26 May 2011 03:29:43 +0000 (11:29 +0800)]
libpixelflinger: ARMv6 specific objects are not used. Removed

6 years agolibpixelfliner: Use GCC visibility to reduce code size by 28%
Jim Huang [Thu, 26 May 2011 03:29:42 +0000 (11:29 +0800)]
libpixelfliner: Use GCC visibility to reduce code size by 28%

6 years agopixelflinger: Fix function naming typo: gglBitBlti
Jim Huang [Thu, 26 May 2011 03:29:42 +0000 (11:29 +0800)]
pixelflinger: Fix function naming typo: gglBitBlti

From the functionality, the function gglBitBlti() implemented bit-block
(image) transfer operation in which several bitmaps are combined into
one using a raster operator.  Renaming to gglBitBlit() should be
explicit.

6 years agolibpixelflinger: Move codeflinger test function to test-opengl-codegen
Jim Huang [Thu, 26 May 2011 03:29:42 +0000 (11:29 +0800)]
libpixelflinger: Move codeflinger test function to test-opengl-codegen

There is no need to include codeflinger test function in
libpixelflinger.  Instead, it should exist in test-opengl-codegen.

7 years agoFix infinite loop in init when debugging is turned off
Colin Cross [Wed, 5 Jan 2011 02:18:45 +0000 (18:18 -0800)]
Fix infinite loop in init when debugging is turned off

If a keychord is pressed when debugging is turned off, the main
event poll in init will return an event on the keychord fd,
but handle_keychord never reads the data.  Once this happens,
the poll will always return immediately, and init enters an
infinite loop.  Fix it by always reading from the keychord fd,
but only handling the returned keychord if debugging is
enabled.

Change-Id: Ie4efa98247d3cc978d275dc8a4516b32aa710278

7 years agodisable console in user builds beagle-gingerbread
Brian Swetland [Wed, 27 Oct 2010 22:40:23 +0000 (15:40 -0700)]
disable console in user builds

Also, run console as shell/log instead of root/root.

Change-Id: I6104f1edddfd5e11843ff41d7839d0c296e6a6f9

7 years agoadd rules to set perms for input device sysfs attributes
Brian Swetland [Tue, 26 Oct 2010 22:38:00 +0000 (15:38 -0700)]
add rules to set perms for input device sysfs attributes

Change-Id: I28ab7f1ac5cc07ade7830e87f5e70c96464bb686

7 years agoinit: support owner/permission setting for sysfs attributes of devices
Brian Swetland [Tue, 26 Oct 2010 22:09:43 +0000 (15:09 -0700)]
init: support owner/permission setting for sysfs attributes of devices

This should be much nicer than peppering init.rc with chown/chmod
directives.

Also, remove some dead code and obsolete comments.

Change-Id: I10895f10a9cf2f1226c8d12976cd3db3743da9ec

7 years agoadd sdcard access permission to ril daemon
Brian Swetland [Sat, 23 Oct 2010 05:17:17 +0000 (22:17 -0700)]
add sdcard access permission to ril daemon

Some RIL daemons write crashdumps to sdcard.

Change-Id: I49a8d2cb342fd7f8bd23dbb1bcc378726e10a579

7 years agoadb: Add support for Philips's USB vendor ID
Mike Lockwood [Fri, 15 Oct 2010 22:14:22 +0000 (18:14 -0400)]
adb: Add support for Philips's USB vendor ID

Change-Id: Iaf2e3f6f111198740703b5c3581d95756f797224
Signed-off-by: Mike Lockwood <lockwood@google.com>
7 years agoMerge "Add AID_NFC for use by the NFC service." into gingerbread
Nick Pelly [Fri, 15 Oct 2010 17:56:17 +0000 (10:56 -0700)]
Merge "Add AID_NFC for use by the NFC service." into gingerbread

7 years agoAdd Asus USB vendor ID to adb.
Xavier Ducrohet [Fri, 15 Oct 2010 17:25:54 +0000 (10:25 -0700)]
Add Asus USB vendor ID to adb.

Bug: 3099385

Change-Id: Ic068fbf58abd23f4bddfe13ae33854a8b4053cf8

7 years agoAdd AID_NFC for use by the NFC service.
Nick Pelly [Fri, 15 Oct 2010 06:30:51 +0000 (23:30 -0700)]
Add AID_NFC for use by the NFC service.

Change-Id: I5b92f87dbacb9d855d99e10b3811f290fb64a144
Signed-off-by: Nick Pelly <npelly@google.com>
7 years agoSet PROT_EXEC on the whole pixelflinger code cache.
Jean-Baptiste Queru [Thu, 14 Oct 2010 21:29:00 +0000 (14:29 -0700)]
Set PROT_EXEC on the whole pixelflinger code cache.

The pointer difference between word pointers is a number
of words, and it needs to be multiplied by the size of a word
to get a proper byte size.

Without this, we tend to see crashes when the code crosses
a page boundary.

Bug: 3026204
Bug: 3097482
Change-Id: I37776d26d5afcdb1da71680de02fbb95e6548371

7 years agoMerge "Define HAVE_STAT_ST_MTIM and disable it under MinGW/Cygwin." into gingerbread
Raphael Moll [Thu, 14 Oct 2010 20:01:06 +0000 (13:01 -0700)]
Merge "Define HAVE_STAT_ST_MTIM and disable it under MinGW/Cygwin." into gingerbread

7 years agoFix KT Tech's usb vendor ID.
Xavier Ducrohet [Thu, 14 Oct 2010 17:06:13 +0000 (10:06 -0700)]
Fix KT Tech's usb vendor ID.

Change-Id: I77a6de8c11d490e2427f704ff64e7c09874a5d14

7 years agoDefine HAVE_STAT_ST_MTIM and disable it under MinGW/Cygwin.
Raphael Moll [Wed, 13 Oct 2010 23:47:35 +0000 (16:47 -0700)]
Define HAVE_STAT_ST_MTIM and disable it under MinGW/Cygwin.

This is used by build\libs\host\CopyFile.c

Also define HAVE_PRINTF_ZD since printf %zd is not
supported on MinGW/Cygwin.
This is used by frmk/base/libs/utils/ZipFileRO.cpp

Change-Id: I570ac5dd0a5b256ee67ad62433f1226597da3ea4

7 years agoPrint out the signal code along with signal number and address.
Carl Shapiro [Sat, 9 Oct 2010 01:10:24 +0000 (18:10 -0700)]
Print out the signal code along with signal number and address.

Change-Id: I60a8b13f706714e80b11bc517101715dfe1c73c1

7 years agoAdd HAVE_PREAD definition to platforms
Kenny Root [Mon, 4 Oct 2010 21:43:15 +0000 (14:43 -0700)]
Add HAVE_PREAD definition to platforms

ZipFileRO uses pread() to guarantee atomic reads at an offset, so use
the HAVE_PREAD flag to trigger off of. The only platform that doesn't
implement it directly is Windows.

Change-Id: Ic927611dc45b75ff8f672ed59afee5be1ad07ab4

7 years agoinit: Add support for SOCK_SEQPACKET socket type
Mike Lockwood [Fri, 1 Oct 2010 12:20:36 +0000 (08:20 -0400)]
init: Add support for SOCK_SEQPACKET socket type

Change-Id: Ib264ecf9beb2685b070436d2bdec9655c7a31b47
Signed-off-by: Mike Lockwood <lockwood@android.com>
7 years agonetstat should include /proc/net/tcp6 and /proc/net/udp6 data
Brian Carlstrom [Tue, 28 Sep 2010 05:49:11 +0000 (22:49 -0700)]
netstat should include /proc/net/tcp6 and /proc/net/udp6 data

Update netstat.c to process /proc/net/tcp6 and /proc/net/udp6.
Example input and output and output follows.

cat /proc/net/tcp:
  sl  local_address rem_address   st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode
   0: 0100007F:13AD 00000000:0000 0A 00000000:00000000 00:00000000 00000000     0        0 669 1 c74dbb40 300 0 0 2 -1
   1: 0100007F:A432 0100007F:2253 01 00000000:00000000 00:00000000 00000000     0        0 15257 1 c74da040 22 4 16 2 -1

cat /proc/net/udp:
  sl  local_address rem_address   st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode ref pointer drops

cat /proc/net/tcp6:
  sl  local_address                         remote_address                        st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode
   0: 00000000000000000000000000000000:8A88 00000000000000000000000000000000:0000 0A 00000000:00000000 00:00000000 00000000     0        0 15461 1 c730aaa0 300 0 0 2 -1
   1: 00000000000000000000000000000000:2253 00000000000000000000000000000000:0000 0A 00000000:00000000 00:00000000 00000000     0        0 15255 1 c7393060 300 0 0 2 -1
   2: 00000000000000000000000000000000:C91B 00000000000000000000000000000000:0000 0A 00000000:00000000 00:00000000 00000000     0        0 15261 1 c6591080 300 0 0 2 -1
   3: 0000000000000000FFFF00000100007F:C91B 0000000000000000FFFF00000100007F:9238 01 00000000:00000000 00:00000000 00000000     0        0 15263 1 c7394aa0 22 4 29 4 -1
   4: 0000000000000000FFFF00000100007F:C0E9 0000000000000000FFFF00000100007F:D3A4 01 00000000:00000000 00:00000000 00000000     0        0 15274 1 c7393aa0 22 4 26 4 -1
   5: 0000000000000000FFFF00000100007F:C41A 0000000000000000FFFF00000100007F:B031 01 00000000:00000000 00:00000000 00000000     0        0 15269 1 c7126ac0 22 4 29 4 -1
   6: 0000000000000000FFFF00000100007F:2253 0000000000000000FFFF00000100007F:A432 01 00000000:00000000 00:00000000 00000000     0        0 15256 1 c65e9040 21 0 0 5 -1
   7: 0000000000000000FFFF00000100007F:9238 0000000000000000FFFF00000100007F:C91B 01 00000000:00000000 00:00000000 00000000     0        0 15262 1 c7394580 22 4 16 4 -1
   8: 0000000000000000FFFF00000100007F:B031 0000000000000000FFFF00000100007F:C41A 01 00000000:00000000 00:00000000 00000000     0        0 15268 1 c71585a0 22 4 16 4 -1
   9: 0000000000000000FFFF00000100007F:8A88 0000000000000000FFFF00000100007F:8A31 01 00000000:00000000 00:00000000 00000000     0        0 15503 1 c71265a0 22 4 30 2 -1
  10: 0000000000000000FFFF00000100007F:D3A4 0000000000000000FFFF00000100007F:C0E9 01 00000000:00000000 00:00000000 00000000     0        0 15273 1 c7126080 22 4 12 4 -1
  11: 0000000000000000FFFF00000100007F:8A31 0000000000000000FFFF00000100007F:8A88 01 00000000:00000000 00:00000000 00000000     0        0 15502 1 c730a060 22 0 0 3 -1

cat /proc/net/udp6:
  sl  local_address                         remote_address                        st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode ref pointer drops
   86: 00000000000000000000000000000000:A256 00000000000000000000000000000000:0000 07 00000000:00000000 00:00000000 00000000  1000        0 14302 2 c6d13820 0
  102: 00000000000000000000000000000000:A466 00000000000000000000000000000000:0000 07 00000000:00000000 00:00000000 00000000  1000        0 14301 2 c6d13ac0 0
  125: 00000000000000000000000000000000:8B7D 00000000000000000000000000000000:0000 07 00000000:00000000 00:00000000 00000000  1000        0 14295 2 c6d13d60 0

netstat:
Proto Recv-Q Send-Q Local Address          Foreign Address        State
tcp        0      0 127.0.0.1:5037         0.0.0.0:*              LISTEN
tcp        0      0 127.0.0.1:42034        127.0.0.1:8787         ESTABLISHED
tcp6       0      0 :::35464               :::*                   LISTEN
tcp6       0      0 :::8787                :::*                   LISTEN
tcp6       0      0 :::51483               :::*                   LISTEN
tcp6       0      0 ::ffff:127.0.0.1:51483 ::ffff:127.0.0.1:37432 ESTABLISHED
tcp6       0      0 ::ffff:127.0.0.1:49385 ::ffff:127.0.0.1:54180 ESTABLISHED
tcp6       0      0 ::ffff:127.0.0.1:50202 ::ffff:127.0.0.1:45105 ESTABLISHED
tcp6       0      0 ::ffff:127.0.0.1:8787  ::ffff:127.0.0.1:42034 ESTABLISHED
tcp6       0      0 ::ffff:127.0.0.1:37432 ::ffff:127.0.0.1:51483 ESTABLISHED
tcp6       0      0 ::ffff:127.0.0.1:45105 ::ffff:127.0.0.1:50202 ESTABLISHED
tcp6       0      0 ::ffff:127.0.0.1:35464 ::ffff:127.0.0.1:35377 ESTABLISHED
tcp6       0      0 ::ffff:127.0.0.1:54180 ::ffff:127.0.0.1:49385 ESTABLISHED
tcp6       0      0 ::ffff:127.0.0.1:35377 ::ffff:127.0.0.1:35464 ESTABLISHED
udp6       0      0 :::41558               :::*                   CLOSE
udp6       0      0 :::42086               :::*                   CLOSE
udp6       0      0 :::35709               :::*                   CLOSE

Bug: 2511871

Change-Id: I25e0470ba65e8a88e57506ae26232a7c43a9c16b

7 years agoQualify the source argument of atomic loads as a const pointer.
Carl Shapiro [Tue, 28 Sep 2010 20:47:03 +0000 (13:47 -0700)]
Qualify the source argument of atomic loads as a const pointer.

Also normalizes the opening brace placment in a few locations.

Change-Id: I8f518e933094337d5d3371321326ffc03b3a5f5a

7 years agoAdd new USB vendor IDs.
Xavier Ducrohet [Tue, 28 Sep 2010 16:54:07 +0000 (09:54 -0700)]
Add new USB vendor IDs.

Change-Id: Ie3b1ee09c67c4c8fdb294cfc99fa28e4ebe8dfae

7 years agoAdd AID_GPS user ID for GPS daemon
Mike Lockwood [Mon, 27 Sep 2010 12:01:58 +0000 (08:01 -0400)]
Add AID_GPS user ID for GPS daemon

Change-Id: I2f8ffba40e3f5145b37296f19c698e08d768ce59
Signed-off-by: Mike Lockwood <lockwood@android.com>
7 years agoMake netd restart when system server hangs on it.
Brad Fitzpatrick [Tue, 21 Sep 2010 22:49:06 +0000 (15:49 -0700)]
Make netd restart when system server hangs on it.

netd has been seen hanging (cause suspected but unconfirmed)
which then brings down the system server which blocks forever
on its main thread trying to connect to netd with its hung
event loop.

This then results in an infinite watchdog runtime restart loop
because the situation never improves:  netd is still borked.

Instead, we should kill netd and give it another chance when
the system server dies.

Bug: 3018996
Bug: 3017876
Change-Id: Ibf9ecf0ffcb78e48c38e5167f21cbeefe0e2a47f

7 years agoadd /vendor support
Brian Swetland [Sun, 19 Sep 2010 10:36:39 +0000 (03:36 -0700)]
add /vendor support

- look in /vendor/firmware/... in addition to /system/etc/firmware/... for firmware loading
- add /vendor/bin to path before /system/bin
- add /vendor/lib to ldpath before /system/lib
- configure appropriate permissions for /system/vendor/bin
- symlink /vendor -> /system/vendor

Change-Id: I0c06ca1e38a44f0c7024cee6cea8907aa93a4532

7 years agoadd graphics events to the eventlog definitions
Mathias Agopian [Tue, 14 Sep 2010 06:05:56 +0000 (23:05 -0700)]
add graphics events to the eventlog definitions

by default this logging is turned off

Change-Id: Ic7e5fe2b377c4760e1673e7d19a69d34cc9333c9

7 years agoMerge "Fetch peer credentials for local sockets" into gingerbread
Kenny Root [Tue, 14 Sep 2010 21:37:08 +0000 (14:37 -0700)]
Merge "Fetch peer credentials for local sockets" into gingerbread

7 years agoFetch peer credentials for local sockets
Kenny Root [Tue, 14 Sep 2010 21:26:12 +0000 (14:26 -0700)]
Fetch peer credentials for local sockets

Fetch the PID, UID, and GID of the remote side of a local socket
connection in case any users of this library class want to check it.

Change-Id: Ia3230e6bc68ab6f93160df9f5996d2bf744b872c

7 years agoAdd dvm_lock_sample's new fields.
Brad Fitzpatrick [Tue, 14 Sep 2010 17:22:48 +0000 (10:22 -0700)]
Add dvm_lock_sample's new fields.

Change-Id: I2ec442834c38d3716695ab90bf0db220bd8dae27

7 years agoMerge "Add missing lock in SocketListener." into gingerbread
Brad Fitzpatrick [Tue, 14 Sep 2010 17:44:15 +0000 (10:44 -0700)]
Merge "Add missing lock in SocketListener." into gingerbread

7 years agoAdd missing lock in SocketListener.
Brad Fitzpatrick [Tue, 14 Sep 2010 17:22:14 +0000 (10:22 -0700)]
Add missing lock in SocketListener.

Change-Id: I3d97a06381fce67ef13b1ccdeaa4e8e2830ab2aa

7 years agoReturn false on socket read error
Kenny Root [Tue, 14 Sep 2010 16:55:22 +0000 (09:55 -0700)]
Return false on socket read error

FrameworkListener was returning the errno from a function marked as
returning bool which caused an implicit conversion to true since we were
in an error block where errno was set to something non-zero.

This caused the clients that had errors to stick around forever and not
get removed from the set of file descriptors that SocketListener was
listening to.

Change-Id: Ia27a4cac47459f3a3c2bb6a7f66803a3165c894a

7 years agomake df more readable
Brian Swetland [Sun, 12 Sep 2010 01:19:35 +0000 (18:19 -0700)]
make df more readable

Filesystem             Size   Used   Free   Blksize
/dev                   164M    32K   164M   4096
/system                442M   117M   325M   4096
/cache                 492M     8M   483M   4096
/data                  984M    59M   925M   4096
/mnt/sdcard             12G    27M    12G   32768

Change-Id: I9a84b7b84ae528ee5cf0b1e92a8bee032b87703b

7 years agoMerge "remount / as read-only only on post-fs to allow per-target config of /" into...
Brian Swetland [Wed, 8 Sep 2010 22:05:44 +0000 (15:05 -0700)]
Merge "remount / as read-only only on post-fs to allow per-target config of /" into gingerbread

7 years agoremount / as read-only only on post-fs to allow per-target config of /
Brian Swetland [Wed, 8 Sep 2010 22:06:45 +0000 (15:06 -0700)]
remount / as read-only only on post-fs to allow per-target config of /

Change-Id: Ia89dd2021e0f960201b4cee573227f0addd48431

7 years agoUse makedev helper for device major and minor numbers
Colin Cross [Fri, 3 Sep 2010 19:25:34 +0000 (12:25 -0700)]
Use makedev helper for device major and minor numbers

Major and minor numbers can be greater than 255.  Major numbers
are 12 bits, and minor numbers are 20 bits.  Supporting major
and minor numbers > 255 allows the use of the extended block
device for extra mmc partitions.

Change-Id: I4c42bfe0781069e8e4cb13ebe26e860cdc03f866

7 years agoIt seems that Windows actually does have a header for <stdbool.h>.
Dan Bornstein [Wed, 1 Sep 2010 23:21:09 +0000 (16:21 -0700)]
It seems that Windows actually does have a header for <stdbool.h>.

Change-Id: I443929d07972389a8c9837926a4b7df37bd6a16a

7 years agoMerge "get rid of compile time warning: "warning: ‘android_ids’ defined but not used...
Nick Kralevich [Wed, 1 Sep 2010 00:45:04 +0000 (17:45 -0700)]
Merge "get rid of compile time warning: "warning: ‘android_ids’ defined but not used"" into gingerbread

7 years agoget rid of compile time warning: "warning: ‘android_ids’ defined but not used"
Nick Kralevich [Tue, 31 Aug 2010 16:44:43 +0000 (09:44 -0700)]
get rid of compile time warning: "warning: ‘android_ids’ defined but not used"

Change-Id: If365bd88f735d425f37d61bcd5a6903f99f6a454

7 years agoam 2ad6067c: (-s ours) do not merge: cherry-picked 44db990d3a4ce0edbdd16fa7ac20693ef6...
Nick Kralevich [Mon, 30 Aug 2010 21:29:13 +0000 (14:29 -0700)]
am 2ad6067c: (-s ours) do not merge: cherry-picked 44db990d3a4ce0edbdd16fa7ac20693ef601b723 from master branch

Merge commit '2ad6067ce491446ab22f59a363d36575a942f5c7' into gingerbread

* commit '2ad6067ce491446ab22f59a363d36575a942f5c7':
  do not merge: cherry-picked 44db990d3a4ce0edbdd16fa7ac20693ef601b723 from master branch

7 years agodo not merge: cherry-picked 44db990d3a4ce0edbdd16fa7ac20693ef601b723 from master...
Nick Kralevich [Fri, 27 Aug 2010 21:35:07 +0000 (14:35 -0700)]
do not merge: cherry-picked 44db990d3a4ce0edbdd16fa7ac20693ef601b723 from master branch

Change-Id: I03e99bfa70ab86eb8eaa2984b6da26a92bd2718a

7 years agoFix bug 2950316. Check return values.
Nick Kralevich [Fri, 27 Aug 2010 21:35:07 +0000 (14:35 -0700)]
Fix bug 2950316.  Check return values.

Change-Id: I687bb5fb8195d4c1fc863e32a5e233a8b9e74196

7 years agoFix the pagesize argument parsing
Maarten Hooft [Tue, 24 Aug 2010 22:24:50 +0000 (15:24 -0700)]
Fix the pagesize argument parsing

Change-Id: I8a3d1a2a70c5ad972d3e150b320ee2184c9d5b4a

7 years agoFix usage string.
Ying Wang [Mon, 23 Aug 2010 19:59:23 +0000 (12:59 -0700)]
Fix usage string.

Change-Id: Iaa99a1bb8bb6998082f3553689b5b812a6471514

7 years agoIncluding dl heap stubs for the dalvikvm host build
Jesse Wilson [Thu, 19 Aug 2010 01:17:55 +0000 (18:17 -0700)]
Including dl heap stubs for the dalvikvm host build

Change-Id: I9d0ce98d433f71db25c58bf464ec9317552cfae6

7 years agoFix a long standing bug in the output buffer size computation.
Carl Shapiro [Sat, 10 Apr 2010 01:23:25 +0000 (18:23 -0700)]
Fix a long standing  bug in the output buffer size computation.

git cherry-pick -e 2a7f2ae7d4b25f89e36be04e47b9e7a3d76e0cfd

7 years agosdcard: Implement statfs, stat time values and change mount point to /mnt/sdcard
Mike Lockwood [Mon, 16 Aug 2010 18:14:44 +0000 (14:14 -0400)]
sdcard: Implement statfs, stat time values and change mount point to /mnt/sdcard

Change-Id: Iac2c4ec47af7d47d76a82916866ad36782caf25c
Signed-off-by: Mike Lockwood <lockwood@android.com>
7 years agofix up permissions
Brian Swetland [Fri, 13 Aug 2010 01:21:12 +0000 (18:21 -0700)]
fix up permissions

Change-Id: I93c828ebc755ad2a2055066c8af65dfde7dc7b2e

7 years agosdcard: a program to create a "virtual" /sdcard pointed at a path
Brian Swetland [Fri, 13 Aug 2010 01:01:08 +0000 (18:01 -0700)]
sdcard: a program to create a "virtual" /sdcard pointed at a path

sdcard is a program that uses FUSE to emulate FAT-on-sdcard style
directory permissions (all files are given fixed owner, group, and
permissions at creation, owner, group, and permissions are not
changeable, symlinks and hardlinks are not createable, etc.

usage: sdcard <path> <uid> <gid>

It must be run as root, but will change to uid/gid as soon as it
mounts a filesystem on /sdcard.  It will refuse to run if uid or
gid are zero.

Change-Id: I9a5d2e5daaebeee632f8470172cbb77b7fa689f8
Signed-off-by: Brian Swetland <swetland@google.com>
7 years agohandle 4096 byte pages if requested via --pagesize
Brian Swetland [Wed, 11 Aug 2010 23:56:49 +0000 (16:56 -0700)]
handle 4096 byte pages if requested via --pagesize

Change-Id: I9299ebeced85dcf5e6e4686b99b28b9786be4f25
Signed-off-by: Brian Swetland <swetland@google.com>
7 years agoTracking merge of dalvik-dev to gingerbread
Brian Carlstrom [Wed, 4 Aug 2010 22:45:52 +0000 (15:45 -0700)]
Tracking merge of dalvik-dev to gingerbread

git cherry-pick --no-commit b99a099c
git cherry-pick --no-commit 9811671e023ceb82029040f7acf599013c773e6b
git cherry-pick --no-commit d0645bf0cc76ef7be11cf77a8e1995233a1972e4

Change-Id: Ie70fecaeaba97a25f114cd2717e381b5107533b1

7 years agoMerge "Add basic lsof command" into gingerbread
Kenny Root [Tue, 27 Jul 2010 19:58:23 +0000 (12:58 -0700)]
Merge "Add basic lsof command" into gingerbread

7 years agoCorrect copyright header for uptime and NOTICE
Kenny Root [Tue, 27 Jul 2010 17:57:00 +0000 (10:57 -0700)]
Correct copyright header for uptime and NOTICE

Change-Id: I89f91c21db87955b20b1ef9b4590f7009f106b26

7 years agoAdd basic lsof command
Kenny Root [Tue, 27 Jul 2010 16:20:02 +0000 (09:20 -0700)]
Add basic lsof command

Basic "list open files" command.

Supports reading all processes for their "exe", "cwd", "root", and "fd"
entries.

Change-Id: I2e2f0911dd65a1993aa1ce9f51222bf264a9d68d

7 years agoMerge "Added Toshiba's USB VID (0x0930) upon their request." into gingerbread
Xavier Ducrohet [Wed, 21 Jul 2010 01:01:02 +0000 (18:01 -0700)]
Merge "Added Toshiba's USB VID (0x0930) upon their request." into gingerbread

7 years agoAdded Toshiba's USB VID (0x0930) upon their request.
Takeshi Kishimoto [Tue, 20 Jul 2010 09:08:09 +0000 (17:08 +0800)]
Added Toshiba's USB VID (0x0930) upon their request.

Change-Id: I9f7fca6b98b0664ef9af4f7b86f0f5863e91e4c1

7 years agoresolved conflicts for merge of 5f5d5c8c to gingerbread
Nick Kralevich [Mon, 19 Jul 2010 22:53:05 +0000 (15:53 -0700)]
resolved conflicts for merge of 5f5d5c8c to gingerbread

Change-Id: Ieee2cd67d33f3e774fcfd96f8a5f67a0374aa873

7 years agovalidate the source of uevent messages
Nick Kralevich [Mon, 19 Jul 2010 21:31:20 +0000 (14:31 -0700)]
validate the source of uevent messages

Bug: 2844206
Change-Id: If2eee54181abfc6c7fda0232f98fa6bb5d12c60c

7 years agoMerge "Fix bug in debuggerd so it can successfully find the exidx section in librarie...
Mike Dodd [Fri, 16 Jul 2010 23:45:38 +0000 (16:45 -0700)]
Merge "Fix bug in debuggerd so it can successfully find the exidx section in libraries. This should fix the bug where the exception stack wasn't being printed past the PC." into gingerbread

7 years agoadd meta-files about 3rd party projects
The Android Open Source Project [Fri, 16 Jul 2010 18:41:39 +0000 (11:41 -0700)]
add meta-files about 3rd party projects

Change-Id: I7a6205fa12f2aa08eff591fb0ce808a48b1e9e74

7 years agoMerge "Add directories for OBB mounting" into gingerbread
Kenny Root [Thu, 15 Jul 2010 20:38:51 +0000 (13:38 -0700)]
Merge "Add directories for OBB mounting" into gingerbread

7 years agoAdd directories for OBB mounting
Kenny Root [Thu, 15 Jul 2010 19:14:44 +0000 (12:14 -0700)]
Add directories for OBB mounting

Change-Id: Ib73e9bca50fb168ab5d147cc260666a770092961

7 years agoFix bug in debuggerd so it can successfully find the exidx section in
Mike Dodd [Wed, 14 Jul 2010 18:28:29 +0000 (11:28 -0700)]
Fix bug in debuggerd so it can successfully find the exidx section in
libraries. This should fix the bug where the exception stack wasn't
being printed past the PC.

Change-Id: Ia69ebd49606ee34e329ba43c07b325660248e8fa

7 years agoinit: use tmpfs/ftruncate for properties backing store instead of ashmem
Brian Swetland [Tue, 13 Jul 2010 23:43:56 +0000 (16:43 -0700)]
init: use tmpfs/ftruncate for properties backing store instead of ashmem

This removes the need for ashmem for early bringup and avoids an issue
with permissions enforcement.

Change-Id: I405b080660934d73048c79d614b6b2ebc43ab182
Signed-off-by: Brian Swetland <swetland@google.com>
7 years agofastboot: make it so I can actually see the bootloader chatter
Brian Swetland [Mon, 28 Jun 2010 18:14:26 +0000 (11:14 -0700)]
fastboot: make it so I can actually see the bootloader chatter

- prefix bootloader messages with (bootloader)
- remove the fancy right justified no newline business

Change-Id: Ica16ebe6a25971db423c5db06300a280d93faf3e
Signed-off-by: Brian Swetland <swetland@google.com>
7 years agoMerge "Add a HAVE define for pthread_setname_np()." into gingerbread
Andy McFadden [Thu, 1 Jul 2010 16:05:51 +0000 (09:05 -0700)]
Merge "Add a HAVE define for pthread_setname_np()." into gingerbread

7 years agoMerge "Fix sim-eng build." into gingerbread
Andy McFadden [Thu, 1 Jul 2010 14:30:51 +0000 (07:30 -0700)]
Merge "Fix sim-eng build." into gingerbread

7 years agoAdd a HAVE define for pthread_setname_np().
Andy McFadden [Wed, 30 Jun 2010 21:49:25 +0000 (14:49 -0700)]
Add a HAVE define for pthread_setname_np().

Due to conflicts with recent WebKit, we need to call it
HAVE_ANDROID_PTHREAD_SETNAME_NP.

https://review.source.android.com/#change,15227 has the original
change, which was reverted after the conflict was found.

Change-Id: I7111960a16505558338756dba91eb50cadb582ff

7 years agoFix sim-eng build.
Andy McFadden [Wed, 30 Jun 2010 22:26:03 +0000 (15:26 -0700)]
Fix sim-eng build.

Include <stddef.h> to get definition of size_t.  Remove inclusion
of <stdint.h>, which doesn't seem to be needed.

Change-Id: Ib7ff231b5ec4b8c672084801fbb294a50cb7c3d7

7 years agoSet explicit timeout on USB control transfer
Werner Johansson [Thu, 1 Jul 2010 01:42:21 +0000 (18:42 -0700)]
Set explicit timeout on USB control transfer

This attempts to solve the intermittent issues with serial number
showing up as all question marks when issuing "fastboot devices"

7 years agoDisplay file sizes larger than 2GB correctly
Kenny Root [Fri, 25 Jun 2010 16:08:05 +0000 (09:08 -0700)]
Display file sizes larger than 2GB correctly

File size was cast to a signed integer which displayed files over 2GB as
negative sizes on devices. Use a long long for printing instead which
matches the stat struct.

Change-Id: I0ec07f33716bb4c2e5d120633ed435f637238f46

7 years agoDefine inline atomic operations for x86 and ARM.
Carl Shapiro [Fri, 4 Jun 2010 00:05:15 +0000 (17:05 -0700)]
Define inline atomic operations for x86 and ARM.

This change moves the ARM definitions into GCC extended inline
assembler.  In addition, the same set of x86 definitions are now
shared among all x86 targets.

Change-Id: I6e5aa3a413d0af2acbe5d32994983d35a01fdcb3

7 years agoAdded NEC and PMC's USB VIDs upon their request.
Takeshi Kishimoto [Fri, 18 Jun 2010 01:37:35 +0000 (09:37 +0800)]
Added NEC and PMC's USB VIDs upon their request.

Change-Id: I332ee1e68ff960d41a365aeb96d012253056edd2

7 years agoUpdate OOM adj levels.
Dianne Hackborn [Fri, 11 Jun 2010 00:52:37 +0000 (17:52 -0700)]
Update OOM adj levels.

This introduces some new buckets that will be used by the activity
manager, and adjusts the default levels to match what we had been
setting for Droid.

It turns out Passion has been using the old levels, which makes no
sense for how much memory it has.  Also it is a problem with these
definitions spreading to other projects for different devices, so
I am going to update the defaults to be appropriate for our newer
devices, and put here a file to tune them down for something like
the sapphire.

Change-Id: Ia4f33c4c3b94aeb3656f60b9222072a1d11e9e68

7 years agoam c8c8c050: merge from open-source master
The Android Open Source Project [Fri, 4 Jun 2010 15:24:02 +0000 (08:24 -0700)]
am c8c8c050: merge from open-source master

Merge commit 'c8c8c050259a0b1222b827b53775e0496c280b7f' into kraken

* commit 'c8c8c050259a0b1222b827b53775e0496c280b7f':
  Improve linenoise.c compatability.

7 years agomerge from open-source master
The Android Open Source Project [Fri, 4 Jun 2010 15:19:30 +0000 (08:19 -0700)]
merge from open-source master

Change-Id: I70c032cb8d1b3fbc4ae59c1e408b6fd0d23d4f2a

7 years agoam 6be163b1: merge from open-source master
The Android Open Source Project [Fri, 4 Jun 2010 15:16:25 +0000 (08:16 -0700)]
am 6be163b1: merge from open-source master

Merge commit '6be163b167e986b4fafda9290e5d74fc3945b273' into kraken

* commit '6be163b167e986b4fafda9290e5d74fc3945b273':
  Emit shell prompt even when in non-interactive mode.

7 years agomerge from open-source master
The Android Open Source Project [Fri, 4 Jun 2010 15:00:52 +0000 (08:00 -0700)]
merge from open-source master

Change-Id: I8b1e7e238d9d0e828bea72530b5356b8386e48bc

7 years agoImprove linenoise.c compatability.
Jack Palevich [Fri, 4 Jun 2010 06:32:01 +0000 (14:32 +0800)]
Improve linenoise.c compatability.

+ Don't flush characters when switching into and out of raw mode. This
avoids eating characters that follow '\n'. (Such characters can occur
when pasting multiple line input, or when scripts are driving input.)

+ Try to be slightly cleverer about calculating the length of the prompt,
so that prompts with embedded '\n' characters are handled OK. This is
an area that really needs to be replaced with a query of the cursor
position from the terminal.

+ As a hack, just assume the screen is very wide if we don't know how
wide it is. This allows dexpropt to work correctly. (It was getting
confused by the editing commands emitted when the end-of-line was reached.)

Change-Id: I988dd0f0bceb22b298e915be0dde085c9358ef66

7 years agoEmit shell prompt even when in non-interactive mode.
Jack Palevich [Fri, 4 Jun 2010 04:03:14 +0000 (12:03 +0800)]
Emit shell prompt even when in non-interactive mode.

Change-Id: If85c509efe13e4bdb40216c79d560ea7f83ba813

7 years agoReap firmware child processes to prevent zombies
Colin Cross [Thu, 3 Jun 2010 19:21:01 +0000 (12:21 -0700)]
Reap firmware child processes to prevent zombies

Change-Id: Ic6c50b929a025f05450cd309238e60a857690d9b

7 years agoam 324a98f1: merge from open-source master
The Android Open Source Project [Thu, 3 Jun 2010 16:16:51 +0000 (09:16 -0700)]
am 324a98f1: merge from open-source master

Merge commit '324a98f1901fb84ec4c88196e01c3e14c2b5a4a4' into kraken

* commit '324a98f1901fb84ec4c88196e01c3e14c2b5a4a4':
  sh: Fix problem with stray '\r' character that was introduced in previous change