Re: [git pull] IDE updates #4

From: Sergei Shtylyov
Date: Wed Oct 22 2008 - 18:57:20 EST


Bartlomiej Zolnierkiewicz wrote:

and number of new submitted patches is < 10 (I'll take
care of fixing them up, ditto for all other new stuff that will be using old
naming scheme).
Thanks for clarifying this.
This rename only added more uncertainty for my pending patchset (which had been already dependant on at least TX4939 driver which keeps being recast by Atsushi and being stale in pata-2.6 series) as I can't predict when you and Linus will merge the changes and this is getting on my nerves, as I don't have time on any extra rework and I'm running out of time with the submission. I know I should have done this earlier and

Maybe some parts could be submitted separately?
(so keeping them up-to-date in pata-2.6 would be my task)

2 (maybe even 3) out of 4 can be but that doesn't make much sense already (and would incur the patch reordering for me) -- the best thing you can do is to merge ASAP the last verison of TX4939 which has my ACK. I'm not sure about TX4938 driver yet -- will look at it after some sleep...

Also I didn't know anything about your patchset and its
dependency on TX4939, otherwise I'll be pushing things in

The patchset consists of a large patch moving read_sff_dma_status() to its porper place, one small preparatory patch, and 2 followup patches, so unfortunately it's dependent on TX4939 in its main patch (worse, the relevant part of this driver has changed after your last merged driver version)...

different order or even skip this pull request if needed
(TX493x drivers are new stuff and were still under review,
such things can be also submitted after the merge window
closes so they were given the lowest priority).

Unfortunately, that driver has been submitted first back 9/09, long before my patchset was even created, so the dependence was just natural.

Please ping me about such issues, I'm certainly willing to
adjust my workflow if this is going to help other people with
their work but I can't do much if I don't know anything...

I probably should've followed up to your first mail expressing the intent to remove the subdirs in 2.6.28 but haven't... anyway, it's hard to predict how the things would go (unless you always expect them to go the worst possible way -- that expectation seems to never fail :-).

I'd have done this if its not for my unfortunate vacations... I aslso kept finding new issues looking on the IOC4 driver. :-/

Though I only was looking at it because of the TX4939 driver being outdated in pata-2.6 series... otherwise I'd have spent my time on the patchset in question. BTW, the reviews I've been doing have also been time consuming... and that's not speaking of the MUSB issues I've been trying to deal with at the same time at work. :-)

I remember about your SGIIOC4 fixes and they are going to
make the next git pull request so sooner this one gets in the
better. Also maybe you could get some help with fixing this
driver from SGI folks if there are still some issues left?

Frankly speaking, I don't see much SGI's interest in this driver (e. g., I bet they could've implemented DMA tuning having the docs but never done that). I've just fixed what my sight fell at (there are 3 pending cleanup patches but those will have to wait now).

To continue my whining, the bad news of the past day were that I discovered that my pet Geode board stopped to come up, and that hard drive plugged into HPT371N chip on my other pet Alchemy board didn't get detected by the recent kernel -- and that's not due to my patches (which I was sanity-testing on this board). When I plugged the drive into Promise IDE card (PDC20269), it got detected alright. I don't know what that means yet -- need to try Rocket100 (HPT370) card yet...


MBR, Sergei

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at