Date: Thu, 1 Nov 2001 09:19:46 +0100 (CET) From: Søren Schmidt <sos@freebsd.dk> To: "Andrey A. Chernov" <ache@nagual.pp.ru> Cc: cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/dev/ata ata-pci.c Message-ID: <200111010819.fA18Jku26675@freebsd.dk> In-Reply-To: <20011101111236.A66386@nagual.pp.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
It seems Andrey A. Chernov wrote: [ Charset KOI8-R unsupported, converting... ] > On Thu, Nov 01, 2001 at 08:58:55 +0100, S?ren Schmidt wrote: > > It seems Andrey A. Chernov wrote: > > > ache 2001/10/31 21:21:50 PST > > > > > > Modified files: > > > sys/dev/ata ata-pci.c > > > Log: > > > Use devclass_find_free_unit() in ata_pci_add_child() because no more > > > implicit auto-scanning up present in device_add_child(). Functionally it is > > > fully equivalent to old code. > > > > What happend with having things reviewed by the maintainer ? > > There is absolutely no change at this place comparing to original variant > of ATA code, so this change is outside one and needed because of > duplicating devices handling code changed, to finish it, because there is > no reports of this place abuse in other drivers. > > > I havn't looked at this, but are you sure it handles the case where > > the onboard controllers come _after_ the PCI bus is scanned, the > > controller that is a MASTER_DEV _must_ end up as ata0/1. > > I just functionally restore what was at this place before, > based on N. Dudorov posted patch showing the bug place. > > I not touch ATA_MASTERDEV part of this function and don't know what > happens there. Uh, oh, how the MASTERDEV thing works is *vital* to this. > The idea behind of interface change is: if you want automatic find of free > slot, specify this action in a visible way, not using some bad > internal magic of devclass_alloc_unit(). That was exactly what I was afraid off, the reason the code used the hard value '2', was that I dont care what number I get, but its *must* be bigger than or equal to '2'. If this doesn't happen now, booting on aforementioned boards are totally broken :( *That* is why you should check with the maintainer before comitting :) -Søren To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200111010819.fA18Jku26675>