海思3536平台经常运行时出现MMB LEAK错误

本帖最后由 Mirol 于 2016-12-22 14:37 编辑

经常在运行时出现MMB LEAK错误，然后重启。
具体LOG如下：
[12-22 09:19:04](!) [  NO NAME        30678.521,689] ( 2809) Direct/Signals:                --> Caught signal 11 (at 0x64, invalid address) <--
[12-22 09:19:04] (!!!)  *** WARNING [still objects in 'Window Pool'] *** [object.c:239 in fusion_object_pool_destroy()]
[12-22 09:19:04] (!!!)  *** WARNING [still objects in 'Layer Region Pool'] *** [object.c:239 in fusion_object_pool_destroy()]
[12-22 09:19:04] (!!!)  *** WARNING [still objects in 'Layer Context Pool'] *** [object.c:239 in fusion_object_pool_destroy()]
[12-22 09:19:04] (!!!)  *** WARNING [still objects in 'GraphicsState Pool'] *** [object.c:239 in fusion_object_pool_destroy()]
[12-22 09:19:04] (!!!)  *** WARNING [still objects in 'Surface Pool'] *** [object.c:239 in fusion_object_pool_destroy()]
[12-22 09:19:05](!) DirectFB/core/vt: Unable to disallocate VT!
[12-22 09:19:05]    --> Device or resource busy
[12-22 09:19:05](-) [  NO NAME        30678.681,368] ( 2809) Direct/Trap:               Raising signal 11 from SigHandler...
[12-22 09:19:05]MMB LEAK(pid=2630): 0x694F4000, 65536 bytes, 'AO(0,1) CirBuf'
[12-22 09:19:05]mmz_userdev_release: mmb<0x694F4000> mapped to userspace 0x9b0b0000 will be force unmaped!
[12-22 09:19:05]MMB LEAK(pid=2630): 0x69531000, 65536 bytes, 'AO(1,1) CirBuf'
[12-22 09:19:05]mmz_userdev_release: mmb<0x69531000> mapped to userspace 0x9b0a0000 will be force unmaped!
[12-22 09:19:05]Segmentation fault


请问该怎么解决，有人遇见相同的情况吗？