https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=233377
Mark Linimon changed:
What|Removed |Added
Assignee|b...@freebsd.org|p...@freebsd.org
--
You are receiv
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=233377
--- Comment #7 from Mark Johnston ---
(In reply to Sean Bruno from comment #6)
Indeed, this points again at the UMA bucket zones. The stack is somewhat
bogus; I believe the only way that we can call uma_zalloc() while freeing
something is
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=233377
--- Comment #6 from Sean Bruno ---
This looks a bit more useful.
panic: Memory modified after free 0xc0036e8621a0(32) val=0 @
0xc0036e8621a0
cpuid = 3
time = 1543128633
KDB: stack backtrace:
0xe0008ac7cd80: at .kdb_backtrac
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=233377
--- Comment #5 from Sean Bruno ---
(In reply to Mark Johnston from comment #4)
Its *almost* all deadc0de, except for one or two bits:
db> x/gx 0xc0037d3c,200
0xc0037d3c: deadc0dedeadc0dedeadc0dedeadc0de
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=233377
--- Comment #4 from Mark Johnston ---
We will want to repro this a number of times to see if the zone is consistent.
I will provide a patch to embed the name of the zone in the panic message
instead.
Could you provide the output of:
db>
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=233377
--- Comment #3 from Sean Bruno ---
I'm not sure this is helpful, but with the debug printf in the kernel the panic
looks like this on the console. Is there any debugging I can do from the db>
prompt to give us a clue here?
draining zone 2
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=233377
--- Comment #2 from Sean Bruno ---
(In reply to Mark Johnston from comment #1)
Oof. That's dumping a lot of text to the serial console at the moment. Every
drain is emitting a line of text and its happening a lot. It probably won't
hit t
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=233377
--- Comment #1 from Mark Johnston ---
The first step will be to figure out which zone this is. I think it'll be
tricky to track this down without vmcores to look at, but this patch will give
us a starting point.
diff --git a/sys/vm/uma_co
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=233377
Bug ID: 233377
Summary: [PowerPC64] Panic during high disk I/O activity
Product: Base System
Version: CURRENT
Hardware: powerpc
OS: Any
Status: New
