Date: Thu, 2 Oct 2003 02:23:03 +0200 From: Erik Trulsson <ertr1013@student.uu.se> To: Murray Stokely <murray@freebsd.org>, stable@freebsd.org, qa@freebsd.org Subject: Re: FreeBSD 4.9 RC1 (i386) now available Message-ID: <20031002002303.GA33156@falcon.midgard.homeip.net> In-Reply-To: <20031001235809.GY91404@gsmx07.alcatel.com.au> References: <20030929151905.GD3743@freebsdmall.com> <20030929160905.GB42388@falcon.midgard.homeip.net> <20031001235809.GY91404@gsmx07.alcatel.com.au>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Oct 02, 2003 at 09:58:10AM +1000, Peter Jeremy wrote: > On 2003-Sep-29 18:09:05 +0200, Erik Trulsson <ertr1013@student.uu.se> wrote: > >On Mon, Sep 29, 2003 at 08:19:05AM -0700, Murray Stokely wrote: > >> Not all FTP sites have the first release candidate, but it is at least > >> available from ftp.freebsd.org. Please download and install this > >> candidate and help us find bugs BEFORE we call it 4.9-RELEASE. > > > ><sarcasm> > >You mean apart from the minor bug that non-pci kernels using ata won't even > >compile, and hasn't worked for the last three weeks or so? > ></sarcasm> > > > >See > > > >http://docs.freebsd.org/cgi/getmsg.cgi?fetch=44649+0+archive/2003/freebsd-stable/20030928.freebsd-stable > > > >or > > > >http://docs.freebsd.org/cgi/getmsg.cgi?fetch=816008+0+archive/2003/cvs-all/20030921.cvs-all > > > >for details and a patch. > > Whilst both Erik and I independently came up with the same patch, upon > reflection, I'm not sure that this is the correct patch. None of the > callers to ata_dmastart() check for a return value and therefore this > probably should be a void function - so the code in ata-isa.c is > correct and the remaining declaration and definitions are incorrect. The "real" version of ata_dmastart() found in ata-dma.c does return different values for different situations, so for compatibility that function would also need to be modified, which I am not sure is a good idea. You are correct in that the return value is currently not checked by the callers to ata_dmastart, but perhaps it should be? I believe the return value probably should be kept, in case some future caller wants to check how the call succeeded, but for an authoritative answer you would have to ask one of the ATA-experts. Anyway, it doesn't really matter. The code compiles with the patch, and a return value that is ignored is quite harmless, so any changes should wait until after 4.9-RELEASE at least. > > In either case, I would request that this be fixed before 4.9-RELEASE. It has already been fixed. The commit to fix it was made by luoqi about 20 hours ago. -- <Insert your favourite quote here.> Erik Trulsson ertr1013@student.uu.se
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20031002002303.GA33156>