|
From: Julian S. <js...@ac...> - 2015-04-13 11:16:38
|
> A possible cause could be the page size: as I understand, > mips have different page size setup. > If your valgrind has been compiled with a pagesize > not matching your kernel/setup, then maybe Valgrind might ask > wrongly aligned mmap requests, giving then this EINVAL Yes, that would be my guess too. At the time the failure happened there was still a block of 1759MB available: | --1956:0:aspacem <<< SHOW_SEGMENTS: out_of_memory (13 segments, 1 [..] | --1956:0:aspacem 6: 0011948000-007f880fff 1759m so it is clearly not the case that it has really run out of memory. J |