vm_page_fault: unhandled page at 0xdeadbeef, ip 0x8006176d

[humdinger]

This is hrev42901.

I recently had a KDL while only having idly open 2 Tracker windows, a query result window, a Mail window. Plus a Web+ with ca. 6 tabs, 4 of those loading pages.

First line:

PANIC: vm_page_fault: unhandled page fault in kernel space at 0xdeadbeef, ip 0x8006176d

The rest attached as KDL.jpg.

[humdinger]

KDL

[anevilyak]

Looks like an uninitialized pointer in the ports code somewhere...

[mmlr]

Replying to anevilyak:

Looks like an uninitialized pointer in the ports code somewhere...

Actually it looks like use of an already freed pointer. It's similar to the other port heap issues reported. Interestingly it actually ends up in 0xdeadbeef, which could now come from either heap implementation, the old heap still used by the port code, and the slab we currently use as kernel heap. Until now I thought there weren't any 0xdeadbeef's in those KDLs, so it may point at the slab, seeing that we only introduced 0xdeadbeef there last week.

[mmlr]

Replying to mmlr:

Until now I thought there weren't any 0xdeadbeef's in those KDLs, so it may point at the slab, seeing that we only introduced 0xdeadbeef there last week.

Ah never mind, I see now that this was against hrev42901, which didn't yet include those changes.

[waddlesplash]

Anyone seen this since?

[waddlesplash]

It's highly probable that this was fixed by hrev46290 or a related commit, as nobody seems to have seen it since.

[nielx]

Assign tickets with status=closed and resolution=fixed within the R1/beta2 development window to the R1/beta2 Milestone