Date: Tue, 24 Nov 2015 00:07:16 +0200 From: Konstantin Belousov <kostikbel@gmail.com> To: Bryan Drewery <bdrewery@FreeBSD.org> Cc: Ed Maste <emaste@freebsd.org>, David Wolfskill <david@catwhisker.org>, "current@freebsd.org" <current@freebsd.org> Subject: Re: Panic: GPF in kernel mode in fork_exit() (prior to FS mouont) Message-ID: <20151123220716.GD58629@kib.kiev.ua> In-Reply-To: <5653678F.3040800@FreeBSD.org> References: <20151123132320.GJ1119@albert.catwhisker.org> <20151123133317.GA58629@kib.kiev.ua> <CAPyFy2BFqN4asLVOTbViCOO6wXvuOXAJJA74PPhXPbLz5fJU_g@mail.gmail.com> <5653678F.3040800@FreeBSD.org>
index | next in thread | previous in thread | raw e-mail
On Mon, Nov 23, 2015 at 11:22:55AM -0800, Bryan Drewery wrote: > On 11/23/2015 9:58 AM, Ed Maste wrote: > > On 23 November 2015 at 08:33, Konstantin Belousov <kostikbel@gmail.com> wrote: > >> > >> The revision 291171 changed layout of the dereferenced structure > >> sysentvec. Was your kernel build clean, or did you used -DNO_CLEAN or > >> similar option ? If yes, remove the kernel build directory and start > >> from scratch. > > > > Every time removing the build directory and starting from scratch > > fixes an issue like this it's demonstrated a bug in our build's > > dependency tracking. Starting over is a fine workaround to make > > progress, but I'd also like to try to extract sufficient detail when > > this happens so that we can fix these cases. > > My guess is that some files are missing proper dependencies on sysent.h. Probably yes. > > For example these 2 would not rebuild. I think any consumers of these > headers which dereferenced the struct pointers would need to include > sys/sysent.h, but that this still may lead to a broken situation somewhere. > > sys/sys/imgact.h:struct sysentvec; > sys/sys/ktrace.h:struct sysentvec; > Why would headers need a rebuild, what does it mean for header anyway ? Indeed, any code which dereferences the structure or otherwise depends on its layout, have to include sys/sysent.h to get the full definition, otherwise compiler cannot generate code. In other words, the problem cannot be in the source, it is somewhere in the build system.home | help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20151123220716.GD58629>