Re: inlining failed

[Nix <nix at esperi dot org dot uk>]

On 29 Nov 2008, Carlos O'Donell verbalised:
>> I posted a patch a while back that allows glibc to build and pass all
>> but one of its tests with stack protection turned on in CFLAGS, which
>> might help. Let's see, ah, here:
>>
>> <http://sourceware.org/ml/libc-help/2008-07/msg00062.html>
>>
>> (2.8, sorry: I'm about to update it to 2.9.)
>
> This looks like a nice patch, have you submitted it as an enhancement
> in the bugzilla?

Nope. I didn't know that was where enhancements were tracked (as opposed
to bugfixes), and I didn't have any idea if the patch was any good at
all, or even useful... I was sort of hoping for comments. 'Nice patch'
counts :)

(I also assumed that every distro out there had probably done something
like this, until I checked and found that a lot of them just force off
stack protection for all of libc. Since only ld.so is really allergic to
it, this is excessive.)


I also thought most of the work on new glibc features was done on
libc-hacker, not bugzilla, and hence that not much work *was* being
done, since libc-hacker seems to be pretty much dead. If bugzilla is the
place to be these days, that puts a different light on things.

> Could you also submit this to eglibc? Embedded targets could make good
> use of this.
> (patches@eglibc.org http://www.eglibc.org/mailing_lists)

I'll attach/submit once I've updated the patch to 2.9, probably
tomorrow.

>> Nobody commented, not even with regard to the apprarent stack corruption
>> in strtol() spotted by running the testsuite with the stack protector on.
>
> If you think there is a bug in strtol() please open a bugzilla issue
> along with your analysis. After all we are a volunteer project.

I got lost inside the code tangle. All I really know is where the
testcase crashes (in glibc 2.8, anyway) with -fstack-protector on, and
with the patch applied anyone will be able to reproduce that, one hopes :)

I'll open a bug once I've retested and verified that it still occurs,
again, probably tomorrow.