linux-kernel - Re: [PATCH] mapletree-vs-khugepaged

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <yt9dwnel4d4x.fsf@linux.ibm.com>
Date:   Mon, 16 May 2022 19:10:22 +0200
From:   Sven Schnelle <svens@...ux.ibm.com>
To:     Liam Howlett <liam.howlett@...cle.com>
Cc:     Heiko Carstens <hca@...ux.ibm.com>,
        Guenter Roeck <linux@...ck-us.net>,
        Andrew Morton <akpm@...ux-foundation.org>,
        "linux-mm@...ck.org" <linux-mm@...ck.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] mapletree-vs-khugepaged

Liam Howlett <liam.howlett@...cle.com> writes:

> * Sven Schnelle <svens@...ux.ibm.com> [220516 11:37]:
>> Hi Liam,
>> 
>> Liam Howlett <liam.howlett@...cle.com> writes:
>> 
>> > * Sven Schnelle <svens@...ux.ibm.com> [220515 16:02]:
>> >
>> > I tried the above on my qemu s390 with kernel 5.18.0-rc6-next-20220513,
>> > but it runs without issue, return code is 0.  Is there something the VM
>> > needs to have for this to trigger?
>> 
>> A coworker said the same. Reason for this seems to be that i've run the
>> code in a unittest environment which seems to make a difference. When
>> compiling the code above with gcc on my system it also doesn't crash.
>> So i have to figure out what makes this unittest binary special.
>> 
>> >> I've added a few debug statements to the maple tree code:
>> >> 
>> >> [   27.769641] mas_next_entry: offset=14
>> >> [   27.769642] mas_next_nentry: entry = 0e00000000000000, slots=0000000090249f80, mas->offset=15 count=14
>> >
>> > Where exactly are you printing this?
>> 
>> I added a lot of debug statements to the code trying to understand
>> it. I'll attach it to this mail.
>
> Thanks.  Can you check to see if that diff you sent was the correct
> file?  It appears to be the git stats and not the changes themselves.

No, it wasn't. I did git stash show with -p, which doesn't make sense.
I've attached the correct diff.

>> > If this is the case, then I think any task that has more than 14 VMAs
>> > would have issues.  I also use mas_next_entry() in mas_find() which is
>> > used for the mas_for_each() macro/iterator.  Can you please enable
>> > CONFIG_DEBUG_VM_MAPLE_TREE ?  mmap.c tests the tree after pretty much
>> > any change and will dump useful information if there is an issue -
>> > including the entire tree. See validate_mm_mt() for details.
>> >
>> > You can find CONFIG_DEBUG_VM_MAPLE_TREE in the config:
>> > kernel hacking -> Memory debugging -> Debug VM -> Debug VM maple trees
>> 
>> I have both DEBUG_MAPPLE_TREE and DEBUG_VM_MAPLE_TREE enabled, but don't
>> see anything printed.
>
> I suspect that this means the issue is actually in the mmap code and not
> the tree.  It may be sensitive to the task-specific layout.  Do you have
> randomization on as well?  I'm thinking maybe I return NULL because I'm
> asking for the next element after the last VMA and not checking that
> correctly or similar.


> Does ./scripts/faddr2line work for you?  What does it say about
> mmap_region+0x19e/0x848 ?

2629        next = mas_next(&mas, ULONG_MAX);
2630        prev = mas_prev(&mas, 0);
2631        if (vm_flags & VM_SPECIAL)
2632               goto cannot_expand;
2633
2634        /* Attempt to expand an old mapping */
2635        /* Check next */
2636        if (next && next->vm_start == end && !vma_policy(next) &&

next above is 0x0e00000000000000 and next->vm_start will trigger the crash.

2637            can_vma_merge_before(next, vm_flags, NULL, file, pgoff+pglen,
2638                              NULL_VM_UFFD_CTX, NULL)) {
2639               merge_end = next->vm_end;
2640               vma = next;
2641               vm_pgoff = next->vm_pgoff - pglen;
2642        }

View attachment "mapple-debug.diff" of type "text/x-diff" (6470 bytes)