On Tue, 2017-02-14 at 16:25 -0800, Bart Van Assche wrote: > Hello Nic, > So I decided to take a step back from this series last week, because at this point any non bug-fix patches will not be considered for v4.11-rc1. > Several additional patches have been added to this series after a stress > test with the iSCSI, ib_srpt and tcm_qla2xxx drivers. The iSCSI target driver > should work at least as reliable as the current upstream code. The ib_srpt > driver now passes my fio --verify concurrently with sg_reset -d stress test > which is not the case for the upstream code. Abort handling in the tcm_qla2xxx > driver is not yet perfect but has been improved significantly over the behavior > of the upstream code. Please review the patches in this patch series. > The layout and ordering of this series is still a mess. I can't emphasize enough how important it is to put the bug-fixes first, preceding all of the other changes. Otherwise, I don't know if the bug-fix is related to your other changes, or a stand-alone bug-fix to upstream needs to be applied to stable, et al. That said, I've gone through your previous series multiple times to extract out various bug-fixes and improvements, and I've asked to have separate individual series for the different areas. So I'm going to go through this *one* more time to pick up bug-fixes for v4.11-rc1. But be warned, but if you don't split up the other changes into separate series like I asked multiple times now, I'll not be continuing to waste time on non bug-fix patches until you do so. Also, one other thing that annoys me with our interactions is that each time I've taken the time to go through and comment, you usually end up snipping large parts of my comments, which means I end up having to repeat myself over and over. I want this practice to stop now. If you ignore my feedback, I'll ignore your non bug-fix patches. It's that simple. I don't even need a long response to each comment. Just either acknowledge it, or say 'this doesn't make sense', or 'I think this is wrong because..' or whatever. But the days of ignoring and snipping my comments, and expecting non bug-fix patches to get merged are going to come to a stop today. So that all said, what I want to see moving forward beyond the bug fixes here is the following split of this series, into smaller more digestible series that are posted individually. 1) Any new bug-fixes to upstream need to always precede the other changes, and not be intermixed with everything else. 2) qla2xxx TMR improvements + target specific changes for qla2xxx 3) Various improvements to TMR handling 4) Various improvements to session shutdown handling 5) Various improvements to target-core To repeat, I don't want to be flooded with all of this stuff in different areas at once. #1 can be sent at any time. #2 is pretty much independent of everything else, but #3 needs to happen before #4 and #5. That way we can focus on the specific areas one at a time, without me having to comment and you having to respond in many different areas at the same time. -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
