`./build_qemu_support.sh` broken on debian buster

[dbrumley]

When I compile on debian buster (I used dockerhub), qemu fails to build due a conflict with memfd_create (perhaps due to newer libc's including this name). I've not been following qemu dev, so it could just be the version used by AFL needs to be updated.

Here is the build error:

 CC      util/memfd.o
util/memfd.c:40:12: error: static declaration of 'memfd_create' follows non-static declaration
 static int memfd_create(const char *name, unsigned int flags)
            ^~~~~~~~~~~~
In file included from /usr/include/x86_64-linux-gnu/bits/mman-linux.h:117,
                 from /usr/include/x86_64-linux-gnu/bits/mman.h:49,
                 from /usr/include/x86_64-linux-gnu/sys/mman.h:41,
                 from /workdir/AFL-2.53b/qemu_mode/qemu-2.10.0/include/sysemu/os-posix.h:29,
                 from /workdir/AFL-2.53b/qemu_mode/qemu-2.10.0/include/qemu/osdep.h:104,
                 from util/memfd.c:28:
/usr/include/x86_64-linux-gnu/bits/mman-shared.h:46:5: note: previous declaration of 'memfd_create' was here
 int memfd_create (const char *__name, unsigned int __flags) __THROW;
     ^~~~~~~~~~~~
make: *** [/workdir/AFL-2.53b/qemu_mode/qemu-2.10.0/rules.mak:66: util/memfd.o] Error 1

(Sorry for no patch; limit time :( )

[Dor1s]

Thanks for the report! @pietroferretti I'm wondering if your patch could've caused this? If yes, should we revert or could you think of a follow up fix please?

[pietroferretti]

I'm a little confused, since the patch was supposed to fix this exact issue. I'm going to try on Debian Buster and see if I can reproduce.

[Dor1s]

Oh, I see. @dbrumley were you using the master branch OR the latest release of AFL? In case of the latter, it doesn't have the aforementioned fix, but it is merged in on master.

[Dor1s]

(GitHub's autocomplete almost ruined my previous comment, should be fixed now, sorry for the spam)

[pietroferretti]

I have no issues when compiling from master on Debian buster, FYI.

[Dor1s]

Thanks a lot for checking! Now waiting for David to confirm whether he still runs into the issue on the latest master or not. If not, I think we should release a new version including the fix.

[dbrumley]

I made a mistake. I pulled the current release, not master, and it gave me the issue. Can confirm master is honkey-dorey and works.

[Dor1s]

Thank you, @dbrumley ! I'm releasing a new version then, given the importance of the fix landed by @pietroferretti (thanks again!).

[jedav]

Would it be worth adding some tests in travis that we build properly in various distros? Possibly via docker?

[Dor1s]

Would it be worth adding some tests in travis that we build properly in various distros? Possibly via docker?

I'm not sure. That doesn't sound too simple, and having more supported configurations means more maintenance burden. Is it worth it?