Date: Sat, 10 Feb 2018 11:42:08 -0600 From: Justin Hibbits <jrh29@alumni.cwru.edu> To: Pedro Giffuni <pfg@freebsd.org> Cc: Ed Maste <emaste@freebsd.org>, Conrad Meyer <cem@freebsd.org>, src-committers <src-committers@freebsd.org>, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: Re: svn commit: r329077 - head/usr.bin/tftp Message-ID: <CAHSQbTB73rde=ggryo4cX-Xy_QVfjPq9FWx_Ln-YDxGwun03mg@mail.gmail.com> In-Reply-To: <e0ded532-bd6b-648a-82c4-7734c56ba73b@FreeBSD.org> References: <201802091946.w19Jkqb3038864@repo.freebsd.org> <CAPyFy2A5zF5D7vDsnoYweh2DUNsO3i1mMJgCCj7uf4vZp=_qeA@mail.gmail.com> <CAHSQbTAfDo6dwPGrNfS9r_f07Z=SCWU9fJJc4PV9S9GEi_PMPg@mail.gmail.com> <e0ded532-bd6b-648a-82c4-7734c56ba73b@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Feb 10, 2018 at 11:10 AM, Pedro Giffuni <pfg@freebsd.org> wrote: > > > On 02/10/18 11:13, Justin Hibbits wrote: >> >> On Fri, Feb 9, 2018 at 7:22 PM, Ed Maste <emaste@freebsd.org> wrote: >>> >>> On 9 February 2018 at 14:46, Conrad Meyer <cem@freebsd.org> wrote: >>>> >>>> Author: cem >>>> Date: Fri Feb 9 19:46:51 2018 >>>> New Revision: 329077 >>>> URL: https://svnweb.freebsd.org/changeset/base/329077 >>>> >>>> Log: >>>> tftp(1): Fix libedit state corruption involving signals >>> >>> From https://ci.freebsd.org/job/FreeBSD-head-mips-build/391/console: >>> >>> 01:15:24 --- all_subdir_usr.bin/tftp --- >>> 01:15:24 --- main.o --- >>> 01:15:24 cc1: warnings being treated as errors >>> 01:15:24 /usr/src/usr.bin/tftp/main.c: In function 'main': >>> 01:15:24 /usr/src/usr.bin/tftp/main.c:182: warning: 'hist' may be used >>> uninitialized in this function >>> 01:15:24 /usr/src/usr.bin/tftp/main.c:181: warning: 'el' may be used >>> uninitialized in this function >>> >> After initializing these, I see the following interesting error: >> >> /home/chmeee/freebsd/head/usr.bin/tftp/main.c:181: warning: variable >> 'el' might be clobbered by 'longjmp' or 'vfork' >> /home/chmeee/freebsd/head/usr.bin/tftp/main.c:182: warning: variable >> 'hist' might be clobbered by 'longjmp' or 'vfork' >> >> This appears to be a known bug in gcc, at >> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=24239 so I don't know >> what the best way to fix it is. > > > Independent of the compiler bug, it does seem like 'el' and 'hist' can be > used uninitialized for the non-interactive case. > > Untested, but perhaps an only-when-needed initialization like this may > workaround the compiler bug. > > Cheers, > > Pedro. Nope, I tried that as well, and hit the compiler bug. r329106 works around both. - Justin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAHSQbTB73rde=ggryo4cX-Xy_QVfjPq9FWx_Ln-YDxGwun03mg>