1. 27 Jan, 2016 1 commit
    • Michael Schanz's avatar
      CGT000029 QMX6: remove revision check of EXT CSD · 8477d370
      Michael Schanz authored
      Revision checking of EXT CSD area avoids that new eMMC devices are recognized
      at all. In the past, it has been observed that new types are still backward
      compatible and handled correctly, although not supported with complete feature set.
      This patch removes the revision checking of the EXT CSD area in order to get
      the driver working for devices with EXT_CSD rev. 6 and above.
      Signed-off-by: Michael Schanz's avatarMichael Schanz <michael.schanz@congatec.com>
      8477d370
  2. 30 Nov, 2015 1 commit
  3. 29 Nov, 2015 21 commits
  4. 12 Jun, 2015 1 commit
    • Shawn Xiao's avatar
      MGS-532 [#1630] P13 driver meet "AXI BUS ERROR" on GC2000 · 6eefa0c2
      Shawn Xiao authored
      When run Antutu benchmark several times continuously, it will
      show "AXI BUS ERROR".
      
      The root cause is When unlock a surface, driver must flush
      its GPU cache. Otherwise, GPU may write out the cache after
      GPU address is freed, which will cause MMU Exception or
      AXI bus error. Even if GPU address isn't freed, data in cache
      may pollute memory after memory is used by another surface.
      
      The solution is Before an event is written to command buffer,
      driver checks the group of records linked to this event to
      find out which kinds of cache need to flush. This method
      doesn't care whether the record is from user side or kernel.
      And driver generates only one flush for a group of records.
      
      Date: Jun 11, 2015
      Signed-off-by: 's avatarMeng Mingming <b51843@freescale.com>
      6eefa0c2
  5. 15 May, 2015 10 commits
    • Shawn Xiao's avatar
      MA-4701 ENGR00278179-3 [#ccc] query openvg video memory database · 6d5f951e
      Shawn Xiao authored
      query openvg video memory database with gpu memory tool
      
      Date: Aug 11, 2014
      Signed-off-by: 's avatarXianzhong <b07117@freescale.com>
      Acked-by: Jason Liu
      6d5f951e
    • Shawn Xiao's avatar
      MGS-715 [#ccc] Uninitialized pointer cause compiler warning · de4988db
      Shawn Xiao authored
       Some variables is not initialized, which causes compiler prompt there
       may be potential harm.
      
       This patch from Conti gives initial value to such variables.
      
       Date May 7, 2015
      Signed-off-by: 's avatarShawn Xiao <b49994@freescale.com>
      de4988db
    • Shawn Xiao's avatar
      MGS-722 [#ccc] Dump kernel command, event queue and DB status when GPU hang · a49e3c68
      Shawn Xiao authored
       To make it easy to debug GPU hang issue in 4.x driver, add the function
       to dump kernel command buffer, event queue and DB status when GPU hang.
       This logic only work when GPU hang, and have no side effect when GPU
       run
       well.
      
       Date May 6, 2015
      Signed-off-by: 's avatarShawn Xiao <b49994@freescale.com>
      a49e3c68
    • Shawn Xiao's avatar
      MGS-332 [#ccc] Support N VSYNC feature in weston client. · 6edd31c9
      Shawn Xiao authored
      Support eglSwapInterval >1 to throttle rendering performance of an
      application.
      Such as setting it to 3 can throttle the application to 20fps on a 60Hz
      display.
      
      Date: Dec 17, 2014
      Signed-off-by Yong Gan <yong.gan@freescale.com>
      6edd31c9
    • Shawn Xiao's avatar
      MGS-293-1 [#ccc] Wayland EGL is not throttling to VSYNC by default · 80da3e4a
      Shawn Xiao authored
       Introduced a new event queue in which the swap worker can dispatch its
       frame callback events so it does not interfere with the main EGL thread
       waiting for the buffer release event.
       Otherwise there would be a race between checking for received events
       and
       entering dispatch function.
      
       Date: Nov 28, 2014
       Signed-off-by Yong Gan <yong.gan@freescale.com>
      80da3e4a
    • Shawn Xiao's avatar
      MGS-732 [#ccc] fix gpu hang when disable power management · fc6e7105
      Shawn Xiao authored
      gpu hang when disable power management, because GPU PU is turned off
      before gpu probe,
      gpu power management need handle this case and turn the power on first.
      
      Date: May 11, 2015
      Signed-off-by: 's avatarXianzhong <b07117@freescale.com>
      fc6e7105
    • Shawn Xiao's avatar
      MA-6054 [#ccc] fix gckKERNEL_QueryProcessDB crash on i.MX6SL · 0bb86910
      Shawn Xiao authored
       this crash issue is caused by kernel NULL pointer when access GPU
       database,
       GPU database is shared by all kernels, it can be queried with any valid
       kernel.
       this patch will find the valid kernel pointer to avoid GPU kernel
       crash.
      
       the backtrace when crash is triggered with lowermem killer on i.mx6sl:
      
       [<c03713b4>] (gckKERNEL_QueryProcessDB+0x24/0x1c4) from [<c0367d00>]
       (viv_gpu_resmem_query+0x28/0x4c)
       [<c0367d00>] (viv_gpu_resmem_query+0x28/0x4c) from [<c039b228>]
       (lowmem_shrink+0x278/0x448)
       [<c039b228>] (lowmem_shrink+0x278/0x448) from [<c00c59d8>]
       (shrink_slab+0x160/0x1e4)
       [<c00c59d8>] (shrink_slab+0x160/0x1e4) from [<c00c82c0>]
       (kswapd+0x664/0x954)
       [<c00c82c0>] (kswapd+0x664/0x954) from [<c008683c>] (kthread+0x80/0x88)
       [<c008683c>] (kthread+0x80/0x88) from [<c003c874>]
       (kernel_thread_exit+0x0/0x8)
      
       Date: Dec 26, 2014
      Signed-off-by: 's avatarXianzhong <b07117@freescale.com>
       Acked-by: Jason Liu
      0bb86910
    • Shawn Xiao's avatar
      MGS-732 ENGR00325693-2 [#1318] Add eglSwapInterval support in Wayland client · b0e9a4af
      Shawn Xiao authored
       Add gcoOS_SetSwapIntervalEx implement for Wayland.
       Wait frame callback if swapInverval enabled.
      
       Date: Aug 04, 2014
       Signed-off-by Yong Gan <yong.gan@freescale.com>
      b0e9a4af
    • Shawn Xiao's avatar
      MGS-327 [#ccc] Unexpected exit like Ctrl-C cause Kernel crash · 487aca18
      Shawn Xiao authored
       Issue in MGS-327 happens when Apps exit unexpectedly such as receiving
       Ctrl-C terminated signal. This is caused by some operations like memory
       free are done twice. If the exit happens when ioctl back from kernel
       space, the event queue will have no chance to be freed but the events
       in
       the queue have been put to GPU. To prevent memory leak, when Apps exit
       GPU driver will clean the resource it occupied. So when flush the
       command buffer, the original event queue will be put to kernel again.
       This will lead the events in the queue will be redo.
      
       To avoid this problem, after the AddList operation in Kernel, clean the
       iface and copy it to user space. Once the same iface is transfered to
       kernel again, its key information will be null and it will not pass
       Addlist parameter check. So it will not be done twice.
      
       Date Dec 15, 2014
      Signed-off-by: 's avatarShawn Xiao <b49994@freescale.com>
      487aca18
    • Shawn Xiao's avatar
      MGS-213 [#1469] detect VG state in power management · 4b1f045a
      Shawn Xiao authored
      this patch is provided by Vivante to enhance VG power up.
      
      comparing to 2D/3D power management, there is no state detection in VG
      power up,
      that will cause that VG core hang due to the abnormal GPU power up.
      
      Date: Dec 01, 2014
      Signed-off-by: 's avatarXianzhong <b07117@freescale.com>
      4b1f045a
  6. 03 Apr, 2015 1 commit
  7. 18 Sep, 2014 3 commits
  8. 20 Aug, 2014 2 commits