Skip to contentSkip to conversation

BestComm ATA task on MPC52xx

DISCUSSION Marian Balakowicz 2 participants 3 messages Jun 1, 2007 – Jun 2, 2007

lore
Share & tools
lore
#0 lore raw

Hi,

I am working on bringing up a DMA support for ATA controller on MPC52xx and I am facing problems for which I am not pretty sure what the actual root cause is.

There was a John Rigby's patch last year, which was based on arch/ppc and drivers/ide (mpc52xx_ide.c). I use arch/powerpc and adopted to libata version of mpc52xx_ide.c which is now merged to vanilla and is called pata_mpc52xx.c. On top of this I have John Rigby's modifications converted from ide to libata. Problems appear with the first DMA transfer attempt, which happens during a driver probe when partition table is read. There shall be a interrupt signaling a transfer end and the handler (not present in John Rigby's code) should do the transfer closing. But the interrupt is not received and after the cmd timeouts it looks like ATA controller registers, BestComm registers, buffer descriptors and BestComm ATA task variables are unchanged. So, it seems that for some reason ATA BestComm task did not run/do it's job.

Did anyone had any success with ATA DMA on 52xx? Or maybe hear about newer version of the ATA BestComm task, or some useful ways of debugging BestComm stuff? I am planning to post my working code but need a day or two to clean it up a little.

Thanks,
Marian

#1 lore raw

Which chip revision are you using?

Signature · 2 lines
Matt Sealey <[email protected]>
Genesi, Manager, Developer Relations
Quoted 30 lines · Marian Balakowicz

Hi,

I am working on bringing up a DMA support for ATA controller on MPC52xx and I am facing problems for which I am not pretty sure what the actual root cause is.

There was a John Rigby's patch last year, which was based on arch/ppc and drivers/ide (mpc52xx_ide.c). I use arch/powerpc and adopted to libata version of mpc52xx_ide.c which is now merged to vanilla and is called pata_mpc52xx.c. On top of this I have John Rigby's modifications converted from ide to libata. Problems appear with the first DMA transfer attempt, which happens during a driver probe when partition table is read. There shall be a interrupt signaling a transfer end and the handler (not present in John Rigby's code) should do the transfer closing. But the interrupt is not received and after the cmd timeouts it looks like ATA controller registers, BestComm registers, buffer descriptors and BestComm ATA task variables are unchanged. So, it seems that for some reason ATA BestComm task did not run/do it's job.

Did anyone had any success with ATA DMA on 52xx? Or maybe hear about newer version of the ATA BestComm task, or some useful ways of debugging BestComm stuff? I am planning to post my working code but need a day or two to clean it up a little.

Thanks,
Marian
_______________________________________________
Linuxppc-dev mailing list
[email protected]
https://ozlabs.org/mailman/listinfo/linuxppc-dev

#2 lore raw
Quoted 1 line · Matt Sealey

Which chip revision are you using?

Here's the configuration:

CPU: MPC5200B v2.2, Core v1.4 at 350 MHz

       Bus 100 MHz, IPB 100 MHz, PCI 25 MHz

The board uses not quite typical 100 MHz bus clock, for which I don't have a optimized ATA DMA timing parameters yet, instead, temporarily I use too wasteful 132MHz settings.

Marian