----- Original Message ----- > Dave Anderson <[email protected]> writes: > > > ----- Original Message ----- > >> OGAWA Hirofumi <[email protected]> writes: > >> > >> OK. More simpler proof, the following is enough to convince you? > > [...] > > > But going back to the error report, the "slab: 0" is kind of confusing: > > > > crash> kmem -s kmalloc-32 > > CACHE NAME OBJSIZE ALLOCATED TOTAL SLABS > > SSIZE > > kmem: kmalloc-32: slab: 0 invalid freepointer: ffff001090e33f80 > > ffff880333001c00 kmalloc-32 32 122658 125440 980 > > 4k > > crash> > > [...] > > > I'm thinking we should clarify that error message, perhaps by storing the > > cpu > > number in si->cpu, and displaying it when si->slab is NULL? > > Just a idea for now though (means not tested error path at all), how > about the following? > > We know already slab pointer at that point. So, this sets si->slab > temporarily. (and slub_page_objects() just uses slab in argument, not > si->slab).
OK great, that works nicely -- let's go with that. Both patches are now queued for crash-7.1.5: https://github.com/crash-utility/crash/commit/f6918567d72bd1ca98dd4119092e25157961e3c2 https://github.com/crash-utility/crash/commit/7f324a41449728fb0a1c36b2264a69aeb65a87b1 Thanks, Dave > > Thanks. > > --- > > memory.c | 23 +++++++++++++++++------ > 1 file changed, 17 insertions(+), 6 deletions(-) > > diff -puN memory.c~crash-slub-freelist-error-fix memory.c > --- crash-64/memory.c~crash-slub-freelist-error-fix 2016-04-21 > 05:04:11.660627253 +0900 > +++ crash-64-hirofumi/memory.c 2016-04-21 05:20:01.673010107 +0900 > @@ -17947,12 +17947,12 @@ static ushort slub_page_objects(struct m > if (CRASHDEBUG(1) && (objects != si->objects)) > error(NOTE, "%s: slab: %lx oo objects: %ld " > "slab objects: %d\n", > - si->curname, si->slab, > + si->curname, page, > si->objects, objects); > > if (objects == (ushort)(-1)) { > error(INFO, "%s: slab: %lx invalid page.objects: -1\n", > - si->curname, si->slab); > + si->curname, page); > return 0; > } > } else > @@ -18039,14 +18039,22 @@ get_kmem_cache_slub_data(long cmd, struc > > switch (cmd) > { > - case GET_SLUB_OBJECTS: > + case GET_SLUB_OBJECTS: { > + /* For better error report, set cur slab to si->slab. */ > + ulong orig_slab = si->slab; > + si->slab = cpu_slab_ptr; > + > if (!readmem(cpu_slab_ptr + OFFSET(page_inuse), > - KVADDR, &inuse, sizeof(short), > - "page inuse", RETURN_ON_ERROR)) > + KVADDR, &inuse, sizeof(short), > + "page inuse", RETURN_ON_ERROR)) { > + si->slab = orig_slab; > return FALSE; > + } > objects = slub_page_objects(si, cpu_slab_ptr); > - if (!objects) > + if (!objects) { > + si->slab = orig_slab; > return FALSE; > + } > > free_objects += objects - inuse; > free_objects += count_free_objects(si, cpu_freelist); > @@ -18055,6 +18063,9 @@ get_kmem_cache_slub_data(long cmd, struc > if (!node_total_avail) > total_objects += inuse; > total_slabs++; > + > + si->slab = orig_slab; > + } > break; > > case GET_SLUB_SLABS: > _ > -- > OGAWA Hirofumi <[email protected]> > -- Crash-utility mailing list [email protected] https://www.redhat.com/mailman/listinfo/crash-utility
