From: "Ævar Arnfjörð Bjarmason" <avarab@gmail.com>
To: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Cc: Jeff King <peff@peff.net>,
git@vger.kernel.org, Junio C Hamano <gitster@pobox.com>,
Erik Faye-Lund <kusmabite@gmail.com>,
Jonathan Nieder <jrnieder@gmail.com>
Subject: Re: [RFC PATCH 02/10] range-diff.c: don't use st_mult() for signed "int"
Date: Fri, 10 Dec 2021 15:58:43 +0100 [thread overview]
Message-ID: <211210.868rwscxcw.gmgdl@evledraar.gmail.com> (raw)
In-Reply-To: <nycvar.QRO.7.76.6.2112101526540.90@tvgsbejvaqbjf.bet>
On Fri, Dec 10 2021, Johannes Schindelin wrote:
> Hi Peff,
>
> On Fri, 10 Dec 2021, Jeff King wrote:
>
>> On Fri, Dec 10, 2021 at 11:22:59AM +0100, Ævar Arnfjörð Bjarmason wrote:
>>
>> > > Dropping the st_mult() does nothing to fix the actual problem (which is
>> > > that this function should use a more appropriate type), but introduces
>> > > new failure modes.
>> >
>> > Yes you're entirely right. I had some stupid blinders on while writing
>> > this. FWIW I think I was experimenting with some local macros and
>> > conflated a testing of the overflow of n*n in gdb with the caste'd
>> > version, which you rightly point out here won't have the overflow issue
>> > at all. Sorry.
>>
>> I'm not sure if this is helpful or not, but this is the minimal fix I
>> came up with that runs the testcase I showed earlier. It's basically
>> just swapping out "int" for "ssize_t" for any variables we use to index
>> the arrays (though note a few are themselves held in arrays, and we have
>> to cross some function boundaries).
>>
>> I won't be surprised if it doesn't hit all cases, or if it even hits a
>> few it doesn't need to (e.g., should "phase" be dragged along with "i"
>> and "j" in the first hunk?). I mostly did guess-and-check on the
>> test-case, fixing whatever segfaulted and then running again until it
>> worked. I didn't even really read the code very carefully.
>>
>> I think you _did_ do more of that careful reading, and broke down the
>> refactorings into separate patches in your series. Which is good. So I
>> think what we'd want is to pick out those parts of your series that end
>> up switching the variable type. My goal in sharing this here is just to
>> show that the end result of the fix can (and IMHO should) be around this
>> same order of magnitude.
>
> I am in favor of this patch. Will you have time to submit this with a
> commit message?
I'd also be happy to pick it up as a massaging of my s/int/intmax_t/
change. I think per[1] that intmax_t is more portable here than ssize_t,
but I'm very likely to be missing something. Corrections most welcome.
Per [1] I ejected that out of my v2 because I think the "cost" being
larger than 1<<16 might not be all that useful. I.e. the limiting that's
in get_correspondences().
But I'll happily admit ignorance on how the actual guts of range-diff
work, I just wanted to fix a segfault I kept running into locally at
some point, and figured I'd submit this RFC.
Doesn't an enlargement of the "int" from an assumed 32 bit unsigned to
say a 64bit unsigned require that 16bit unsigned COST_MAX to be
correspondingly bumped to 32bit unsigned? I.e. we'd define it as 1/2 of
whatever "intmax_t" (or "ssize_t" or "long long int" or whatever) is
defined as?
That may be a question under the umbrella of "Ævar doesn't actually
understand range-diff", but think I recall playing with bumping one and
not the other (or bumping COST_MAX too close to the size of the
container type) and running into errors...
1. https://lore.kernel.org/git/211210.86czm4d3zo.gmgdl@evledraar.gmail.com/
next prev parent reply other threads:[~2021-12-10 15:07 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-12-09 19:19 [RFC PATCH 00/10] range-diff: fix segfault due to integer overflow Ævar Arnfjörð Bjarmason
2021-12-09 19:19 ` [RFC PATCH 01/10] string-list API: change "nr" and "alloc" to "size_t" Ævar Arnfjörð Bjarmason
2021-12-09 19:19 ` [RFC PATCH 02/10] range-diff.c: don't use st_mult() for signed "int" Ævar Arnfjörð Bjarmason
2021-12-10 3:39 ` Jeff King
2021-12-10 10:22 ` Ævar Arnfjörð Bjarmason
2021-12-10 11:41 ` Jeff King
2021-12-10 12:31 ` Ævar Arnfjörð Bjarmason
2021-12-10 19:24 ` Phillip Wood
2021-12-14 14:34 ` Jeff King
2021-12-10 14:27 ` Johannes Schindelin
2021-12-10 14:58 ` Ævar Arnfjörð Bjarmason [this message]
2021-12-11 14:01 ` Johannes Schindelin
2021-12-12 17:44 ` Ævar Arnfjörð Bjarmason
2021-12-14 14:42 ` Jeff King
2021-12-09 19:19 ` [RFC PATCH 03/10] range-diff.c: use "size_t" to refer to "struct string_list"'s "nr" Ævar Arnfjörð Bjarmason
2021-12-09 19:19 ` [RFC PATCH 04/10] range-diff: zero out elements in "cost" first Ævar Arnfjörð Bjarmason
2021-12-09 19:19 ` [RFC PATCH 05/10] linear-assignment.c: split up compute_assignment() function Ævar Arnfjörð Bjarmason
2021-12-09 19:19 ` [RFC PATCH 06/10] linear-assignment.c: take "size_t", not "int" for *_count Ævar Arnfjörð Bjarmason
2021-12-09 19:19 ` [RFC PATCH 07/10] linear-assignment.c: convert a macro to a "static inline" function Ævar Arnfjörð Bjarmason
2021-12-09 19:19 ` [RFC PATCH 08/10] linear-assignment.c: detect signed add/mul on GCC and Clang Ævar Arnfjörð Bjarmason
2021-12-10 3:56 ` Jeff King
2021-12-09 19:19 ` [RFC PATCH 09/10] linear-assignment.c: add and use intprops.h from Gnulib Ævar Arnfjörð Bjarmason
2021-12-09 19:19 ` [RFC PATCH 10/10] linear-assignment.c: use "intmax_t" instead of "int" Ævar Arnfjörð Bjarmason
2021-12-10 4:00 ` Jeff King
2021-12-10 12:30 ` [RFC PATCH v2 0/5] range-diff: fix segfault due to integer overflow Ævar Arnfjörð Bjarmason
2021-12-10 12:30 ` [RFC PATCH v2 1/5] range-diff: zero out elements in "cost" first Ævar Arnfjörð Bjarmason
2021-12-14 13:36 ` Jeff King
2021-12-10 12:30 ` [RFC PATCH v2 2/5] linear-assignment.c: split up compute_assignment() function Ævar Arnfjörð Bjarmason
2021-12-14 13:39 ` Jeff King
2021-12-10 12:30 ` [RFC PATCH v2 3/5] linear-assignment.c: take "size_t", not "int" for *_count Ævar Arnfjörð Bjarmason
2021-12-14 13:40 ` Jeff King
2021-12-10 12:30 ` [RFC PATCH v2 4/5] range-diff.c: rename "n" to "column_count" in get_correspondences() Ævar Arnfjörð Bjarmason
2021-12-14 13:42 ` Jeff King
2021-12-10 12:30 ` [RFC PATCH v2 5/5] range-diff: fix integer overflow & segfault on cost[i + n * j] Ævar Arnfjörð Bjarmason
2021-12-14 14:04 ` Jeff King
2021-12-10 14:31 ` [RFC PATCH 00/10] range-diff: fix segfault due to integer overflow Johannes Schindelin
2021-12-10 15:07 ` Ævar Arnfjörð Bjarmason
2021-12-21 23:22 ` Philip Oakley
2021-12-21 23:36 ` Ævar Arnfjörð Bjarmason
2021-12-22 20:50 ` Johannes Schindelin
2021-12-22 21:11 ` Jeff King
2021-12-24 11:15 ` Philip Oakley
2021-12-24 16:46 ` Ævar Arnfjörð Bjarmason
2021-12-24 18:31 ` Philip Oakley
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=211210.868rwscxcw.gmgdl@evledraar.gmail.com \
--to=avarab@gmail.com \
--cc=Johannes.Schindelin@gmx.de \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=jrnieder@gmail.com \
--cc=kusmabite@gmail.com \
--cc=peff@peff.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).