From mboxrd@z Thu Jan 1 00:00:00 1970 From: Luben Tuikov Subject: Re: [PATCH] 2.5.x add back missing scsi_queue_next_request calls Date: Fri, 21 Mar 2003 14:20:48 -0500 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <3E7B6610.90707@splentec.com> References: <20030319184444.A9694@beaverton.ibm.com> <3E7A317A.702@splentec.com> <3E7A529C.5050607@torque.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Return-path: List-Id: linux-scsi@vger.kernel.org To: dougg@torque.net Cc: Patrick Mansfield , linux-scsi@vger.kernel.org Douglas Gilbert wrote: > > If you don't like what Patrick is proposing, please explain > why. I think I _did_ explain why: contamination of scsi_put_command(). I also think that I explain it in quite a dry and formal way with examples, trying really hard to convey what I meant. Please feel free to give me pointers on where I ``missed the mark''. > [Rather than re-explain, you could supply a url to the > prior post in the marc.theaimsgroup.com archive.] On slab allocation: http://MARC.10East.com/?l=linux-scsi&m=104438145823604&w=2 The whole message is about that. The second ``Point 2'', on ``change'', ``new'' and ``before'': http://MARC.10East.com/?l=linux-scsi&m=104387923212007&w=2 Excerpt: 2. This is a pickle and might I mention that Doug's queue depth changes were *for a reason*, so there's no point in saying ``the way it was before''. Furthermore ``the number of outstanding commands'' is ambiguous and any which way it is it, is NOT enough. I.e. outstaning in LLDD, oustanding free, etc. -- it just doesn't compute. This was about the ``way it was before'', i.e. we should not improve SCSI Core by means of its old workings, but by means of new infrastructure... (Here's I'd be repeating myself, as I've said those things before on linux-scsi... sorry, no time to search the archives now.) > I don't > think boldface exclamations are required. I was just trying to make myself clear. > Naturally you are at liberty to present another patch which > merges what you are trying to do with Patrick's work. No, I don't want to. This has to do with managing ppl and the reasons are more or less as follows: Patrick is already doing the work -- let him continue to do it, he's doing a great job; I don't believe in taking someone's patch (idea) and rewriting it and resubmitting it as my own. If they step out of line, I'd let them know, they can correct it, resubmit the patch, and thusly continue to do the great work they've been doing. This way they learn, get more enthusiastic and feel really good about themselves and next time they'll do an even better job. If you guys were closer, I'd all buy you a beer! :-) -- Luben