Re: + fix-ide-dma-resource-managment.patch added to -mm tree

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

Hello.

Sergei Shtylylov wrote:

> Bartlomiej Zolnierkiewicz wrote:

>>> [skipped]

>>>>>>> - release the channel's secondary DMA I/O ports for real

>>>>>> NAK, looking and the code - secondary DMA base is only used
>>>>>> by siimage.c and sgiioc4.c (both use ->mmio == 2, for siimage
>>>>>> ->dma_base2
>>>>>> is unused, for sgiioc4 it is used for some weird DMA transfer ending
>>>>>> stuff)

>>>    Hm, concerning the other field of ide_hwif_t, dma_master: I 
>>> currently see
>>> no users to it but pdc202xx_old.c. The way it's computed makes it 
>>> dependent on
>>> what channels are enabled (though all channels are forced present in 
>>> that
>>> driver by the most recent patch) which makes it invalid if the 
>>> primary channel
>>> happens to be disabled. I'd like to replace it with extra_base fields 
>>> which
>>> would point to the extra DMA ports regardless of the channels 
>>> present/enabled,
>>> is this OK?

>> Isn't it better to just use pci_resource_start(dev, 4) in pdc202xx_old.c?

>    Not only better but is actually correct -- extra_base would belong 
> only to a sole channel in a pair... :-< I'm still not sure this field is 
> worth introducing... I probably won't.

    OTOH, it's handy in ide-dma.c itself. And hwif->extra_base is surely more 
efficient than pci_resource_start(hwif->dev, 4). So, I'm going to introduce 
it, just changing the initial code somewhat...

WBR, Sergei