This bug only occurs in TT-mode so it is either a memory-manager problem
or the tt-engine in it self.
When the jffs walks the pages in the mtd device, a random page (at my
machine somewhere at page 90 through 110), the kernel panics.
When ubd driver reads it data from the host with the standard read()
command, it fills 512 bytes into 0xa09b2000, but the same call then also
trash a bhs list located at 0xa280421c. The first 120 entries get
trashed with random(?) data. All pages red before at this point (and
using same buffers doesn't fail)
It does work as expected in SKAS mode, so I give up at this point. I
don't now anything about the TT and SKAS engine, and don't have much
time currenly to investigate any further then I now can simulate what I
want for now.
Get latest updates about Open Source Projects, Conferences and News.