[lkml]   [2002]   [Aug]   [24]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: IDE janitoring comments
>>  - Do we really want to keep all those _P versions around ?
>> A quick grep showed _only_ by the non-portable x86 specific
>> recovery timer stuff that taps ISA timers (well, I think ports
>> 0x40 and 0x43 and an ISA timer). I would strongly suggest to
>I'd like to keep them around for the moment. They should be using
>udelay() but thats a general issue with _p inb/outb etc.

I don't think we need them at all in hwif (see below)

>> After much thinking about the above, I came to the conslusion
>> we probably want to just kill all the IN_BYTE, OUT_BYTE, etc.
>Agreed entirely
>> Also, getting rid of the _P version would make things a lot
>> easier as well here too.
>What currently uses the _P versions ?

Ok, I looked and found out that those are used by

- some weird stuff in ide.c tapping IO ports at 0x40 and 0x43 that
I assume is a timer. (Can someone make sure that code don't get used
on anything but legacy x86 ?).

- a couple of interface drivers like cmd640 tapping the PCI config
IO stuffs for probing.

In all cases, I firmly beleive those are way outside of the domain
of application of the hwif-> abstraction. Those are code that knows
it's tapping legacy stuff on a IO port, and can/should directly use
the in/out_p function. I don't think those have anything to do in
hwif. All that should be in hwif is what is needed by the generic
code to tap the IDE registers themselves.

Do you agree ? (I'd love to get rid of them :)

If you agree, I'll send you a patch tomorrow along with the fixing
of ide-pmac that will do

- Remove the _P versions from hwif->iops
- slightly change ide-iops to define both sets of iops, one set
providing PIO ops using directly in/outx & int/outsx, and one
set providing MMIO ops using directly read/writex

Anybody that need different routines for the generic IDE code to
tap the taskfile (or eventually DMA) registers (typically cris
arch) will provide it's own set of routines to hwif. I can probably
fix cris (though I don't know anything about that arch, but the
code seem obvious enough), and i'll rely on you to fix m86k :)


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

 \ /
  Last update: 2005-03-22 13:28    [W:0.040 / U:1.352 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site