> > > I think the issue is much more fundamental than that; you appear to have > page-table corruption (for example, "pte:262d2626292a2627" and > "pte:1110111111111111" are definitely corrupted) and so anything dealing > with 'struct page' derived from the physical address in the pte is going to > go wonky. > > From the logs here, the pmds look ok but these are the pte values I > spotted: > > 0x1110111111111111 > 0x262d2626292a2627 > 0xa098a09aa29ea8a4 > 0x212725231f242323 > 0x2626262023222323 > > which don't seem to correspond to any sort of poison, but are possibly > artifacts of repeated patterns with random bits cleared? > > Will Very valuable thought. I am streaming camera data, and generally the image captured from my desk is very dark (so these might be 8-bit brightness values from a mono ov9281 sensor). I then set the camera to generate a test pattern, i.e. it will only produce output like this 00000090: ffff ffff ffff ffff ffff 00ff 00ff 00ff 000000a0: 00ff 00ff 00ff 00ff 00ff 00ff 00ff 00ff ( also ff00 and 0000 ). And those were the crashes that followed: [ 54.838582] BUG: Bad page map in process pool-upowerd pte:ff00ffffffffff pmd:800000001801003 [ 54.839187] BUG: Bad page map in process pool-upowerd pte:ff00ff00ff00ff pmd:800000001801003 [ 54.839615] BUG: Bad page map in process pool-upowerd pte:ff00ff00ff00ff pmd:800000001801003 [ 54.845439] BUG: Bad page map in process pool-upowerd pte:ff00ff00ff00ff pmd:800000001801003 [ 49.664066] BUG: Bad page map in process gmain pte:ff00ffffffffff pmd:800000001801003 [ 49.664515] BUG: Bad page map in process gmain pte:ff00ff00ff00ff pmd:800000001801003 [ 49.664879] BUG: Bad page map in process gmain pte:ff00ff00ff00ff pmd:800000001801003 [ 49.665245] BUG: Bad page map in process gmain pte:ff00ff00ff00ff pmd:800000001801003 [ 49.665610] BUG: Bad page map in process gmain pte:ff00ff00ff00ff pmd:800000001801003 [ 49.665973] BUG: Bad page map in process gmain pte:ff00ff00ff00ff pmd:800000001801003 [ 49.666336] BUG: Bad page map in process gmain pte:ff00ff00ff00ff pmd:800000001801003 [ 49.666698] BUG: Bad page map in process gmain pte:ff00ff00ff00ff pmd:800000001801003 [ 49.667061] BUG: Bad page map in process gmain pte:ff00ff00ff00ff pmd:800000001801003 That confirms my suspicion of a dma write-out-of-bounds from a very separate angle (thanks again for the valuable idea). I have since taken this to https://github.com/raspberrypi/linux/issues/5138 but will update the list once investigation has found a solution. Thanks, Max
