Date: Wed, 09 Nov 2005 00:07:01 -0800 From: Nate Lawson <nate@root.org> To: Poul-Henning Kamp <phk@phk.freebsd.dk> Cc: Craig Rodrigues <rodrigc@FreeBSD.org>, cvs-src@FreeBSD.org, src-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/kern vfs_mount.c Message-ID: <4371AE25.3090008@root.org> In-Reply-To: <26185.1131519693@critter.freebsd.dk> References: <26185.1131519693@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
Poul-Henning Kamp wrote: > In message <43716BD6.5050604@root.org>, Nate Lawson writes: > > >>While I don't have ideas for a better general mechanism for this, I >>think it sets a bad precedent. We can't have every complex syscall >>transporting its own error message strings back to the user program. >>And we can't expand errno to be the union of every single API-specific >>error either. > > > That was one the main points of my ioctl talk at BSDcan. > > One way to solve it: > > Hang a sbuf pointer in each thread and APIs cound stuff their error > message there. > > Add a new syscall to return the string to userland. > > Modify perror(3), err(3) and similar to pull out the "extended" > error, if there is one. This sounds better. Not sure if it's fully baked, but definitely a better direction. -- Nate
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4371AE25.3090008>