Re: [PATCH 2/2] hpt366: UltraDMA filtering for SATA cards

[Sergei Shtylyov <sshtylyov@ru.mvista.com>]

Bartlomiej Zolnierkiewicz wrote:

>>>> 101 files changed, 1880 insertions(+), 2828 deletions(-)

>>>>please look at -mm or IDE quilt tree instead. :)

>>>   Looking...

>>    When are you planning to push out to Linus the 
>>ide-mode-limiting-fixes-for-user-requested-speed-changes.patch? I'd like my 
>>HPT37x SATA mode filtering stuff to be atop of this one, after looking at it.

> Preferably 2.6.24 material and ide_rate_filter() FIXME (respecting device PIO
> limits) still needs to be addressed before pushing all mode limiting patches
> upstream.

    Maybe I'll look into this...

>>>>ide_rate_filter() happily uses ide_find_dma_mode() nowadays (however this
>>>>hpt366 patch is for vanilla kernel which doesn't have the needed 
>>>>changes).

>>    Yeah, it keeps being in the same vein (same bug rather :-) as the old 
>>code, i.e. not looking at hwif->mwdma_mask when falling back in 
>>ide_rate_filter()...

> Worth fixing but deserves a separate patch.

    It does. Unfortunately, after you said that this issue has been already
dealt width, I've dropped (and lost) the code fixing this -- will have to redo
it now. :-/
    I'm now envisioning the HPT37[24] SATA filtering work as series of n patches:

[1/4] introduce drive_is_sata() helper + minor fix to eighty_ninty_three()
[2/4] hpt366: UltraDMA mode filter for SATA cards
[3/4] fix ide_rate_filter() to respect hwif->mwdma_mask
[4/4] introduce mwdma_filter() method and use it for HPT37x-based SATA cards

> Bart

MBR, Sergei