Permalink
Commits on Sep 19, 2015
  1. mt2: Build TWRP

    mdmower committed Sep 19, 2015
    Change-Id: I874f096b9b748fe6b0b1aed75ea91b91f4e4c447
Commits on Sep 18, 2015
  1. mt2: Use BFQ I/O scheduler

    u-ra committed Sep 18, 2015
    Change-Id: I47368b1ed98948ed0e096b73cbcc3904b49b855a
Commits on Sep 17, 2015
  1. mt2: Add vibrator intensity control

    Kali- authored and u-ra committed Oct 14, 2013
    Change-Id: I45b986ca3c798b777e5059323754ef295322b9de
Commits on Sep 16, 2015
  1. mt2: Update fstab

    u-ra committed Sep 16, 2015
    * Add support for F2FS
    * Don't track atime
    * Barriers are enabled by default
    
    Change-Id: I8497ef83869b917bc4721ffab39cf00a6ba39148
  2. Revert "mt2: Disable HDR for the time being"

    u-ra committed Sep 14, 2015
    This reverts commit 36e3b74.
    
    Change-Id: I22b5599bfaac717e286fe0204bca1ad4d6c8a9fa
Commits on Sep 14, 2015
  1. mt2: Add missing GPS blob

    u-ra committed Sep 14, 2015
    Change-Id: Id0d0d6f6b6ef018321d6d922e507af6b35015043
  2. mt2: No need for reserved vector symbols anymore

    u-ra committed Sep 11, 2015
    Change-Id: I5461052bf1b61b057a6e73b737b947caceed968c
  3. mt2: Delete unused touchscreen firmware

    u-ra committed Sep 11, 2015
    Change-Id: Ia978f0ca7cb7f8bb278e1093f58b39d60399139f
  4. mt2: Update camera stack from B322 release

    u-ra committed Sep 11, 2015
    Change-Id: I98be4d7da22dbb89f4b2a1f32e43162697854e97
Commits on Sep 11, 2015
  1. Revert "mt2: Build with Huawei symbols"

    mdmower committed Sep 10, 2015
    This reverts commit 06ae312.
    
    Change-Id: I7d1d2d84bbc7c91e8d7ed61dcd9a8f6c3a34bf6b
  2. mt2: Provide global Huawei symbols via libc

    mdmower committed Sep 10, 2015
    Huawei included some non-standard functions in libcutils and liblog.
    Make these functions globally available by including them in libc.
    
    Parts of this commit originated from:
      libcutils/liblog: Add functions required by Huawei g620_a2 binaries
      Thomas Wendt
      I58604686008340ef4e10f3f67dc329fe7bd864cc
    
    Change-Id: Ie3a84ce5b6b969ef4cc17b2028cc03298a43a79b
Commits on Sep 9, 2015
  1. mt2: Allow breakfast sync to succeed if no vendor

    mdmower committed Sep 9, 2015
    Maintainers are of different opinions on this, but I'm in the camp of
    letting breakfast finish error-free when initially syncing a device.
    Vendor can be populated subsequent to device sync.
    
    Change-Id: I1e203fb5c6ff3dcc24156b1681ebcee28345e591
  2. mt2: Fix kernel repo dependency

    u-ra committed Sep 9, 2015
    Change-Id: I823d38df019c675c068f1ac8dd037a8dcd99def3
  3. mt2: Update repo dependencies

    u-ra committed Sep 8, 2015
    Change-Id: If5a5fe014602e8ce9816168427ab02cd29e7de3e
Commits on Sep 8, 2015
  1. mt2: Restrict installs to B322 bootloader

    mdmower committed Sep 8, 2015
    Rule out old firmware issues in releases going forward.
  2. mt2: libril: ril.cpp: fix misuse of strncat

    nickkral authored and u-ra committed Feb 8, 2015
    strncat(dest,src,size) appends size+1 bytes to the end of
    dest, so sizeof(dest) must be greater than
    strlen(dest) + size + 1. Passing the buffer size to strncat
    instead of sizeof(dest) - strlen(dest) - 1 is a common
    strncat bug. Use strlcat instead, as it has more intuitive
    behavior and ensures the buffer is properly null terminated.
    
    Addresses the following compiler warning:
    
      In file included from system/core/include/cutils/sockets.h:22:0,
                       from hardware/ril/libril/ril.cpp:24:
      In function 'char* strncat(char*, const char*, size_t)',
          inlined from 'void android::RIL_register(const RIL_RadioFunctions*)' at hardware/ril/libril/ril.cpp:4258:62:
      bionic/libc/include/string.h:199:61: warning: call to char* __builtin___strncat_chk(char*, const char*, unsigned int, unsigned int) might overflow destination buffer
           return __builtin___strncat_chk(dest, src, n, __bos(dest));
                                                                 ^
    
    (line numbers are from internal master and do not match AOSP)
    
    Even with this change, this code feels weird.
    MAX_DEBUG_SOCKET_NAME_LENGTH is 12, and rildebug is initialized to be
    SOCKET_NAME_RIL_DEBUG ("rild-debug"), which is 11 bytes including null
    terminator. The strlcat call here can append a maximum of 1 byte before
    the buffer is full. I don't know if this is intended or not.
    
    Change-Id: I49801ad1ea3aa6173bbc9fd7cf00f3d308693253
  3. mt2: libril: Fix for rild crash becasue of buffer overflow

    Prasath Balakrishnan authored and u-ra committed Aug 5, 2015
    There is a buffer overflow which causing rild crash. For
    appendPrintBuf, "sprintf" is replaced with "snprintf" for
    the safe copy operation.
    
    Change-Id: I25430a5c6f2a8b19a3271235323499b6ee802746
    CRs-Fixed: 881515
  4. mt2: libril: Work around list corruption, take 2

    u-ra committed Sep 8, 2015
    Given that we simply can't trust RequestInfo* lifetime (thanks Huawei,
    thanks a bunch), don't even try to keep a list of them. Instead, keep a
    list of token ids and invalidate them as needed.
    
    Previous attempt crashed too often, delaying the network init at boot.
    This still blows, though.
Commits on Sep 4, 2015
Commits on Sep 3, 2015
  1. mt2: Ship QCOM sound effects

    u-ra committed Sep 3, 2015
  2. mt2: Ship OMX decoders

    u-ra committed Sep 2, 2015
Commits on Sep 2, 2015
  1. mt2: Use dlmalloc by default

    u-ra committed Sep 2, 2015
  2. mt2: Set EGL cache sizes

    u-ra committed Sep 2, 2015
Commits on Aug 31, 2015
Commits on Aug 28, 2015
  1. Revert "mt2: Stop using COMMON_GLOBAL_C(PP)FLAGS"

    u-ra committed Aug 28, 2015
    hardware/ril-caf thinks it's too good to pay attention to
    TARGET_GLOBAL_CFAGS.
    
    This reverts commit f6872f6.
  2. mt2: Build with Huawei symbols

    u-ra committed Aug 28, 2015
  3. mt2: Set ACDB files path

    u-ra committed Aug 28, 2015
  4. Make the implementations of RIL_onUnsolicitedResponse match the proto…

    Bernhard Rosenkränzer authored and u-ra committed Nov 17, 2014
    …types
    
    Prototypes expect the data as a const void*, whereas the implementation
    takes a void*.
    
    This causes the build to break with clang and may also have other
    unexpected side effects. (I'd expect to see unresolvable references in
    some gcc builds.)
    
    Change-Id: Ic34787da20abd2821730a18e49c0a435a82c2613
    Signed-off-by: Bernhard Rosenkränzer <Bernhard.Rosenkranzer@linaro.org>