Re: RE:cciss patches for 2.4.21-rc1, 4 of 4

[Arjan van de Ven <arjanv@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1232 bytes --]

On Fri, 2003-04-25 at 00:12, mikem@beardog.cca.cpqcorp.net wrote:
> 20030424
> 
> Changes:
> 	1. Sets the DMA mask to 64 bits. Removes RH's code for the DMA mask.
> 
> diff -urN lx2421rc1-p3/drivers/block/cciss.c lx2421rc1/drivers/block/cciss.c
> --- lx2421rc1-p3/drivers/block/cciss.c	Wed Apr 23 14:40:48 2003
> +++ lx2421rc1/drivers/block/cciss.c	Wed Apr 23 14:51:55 2003
> @@ -106,7 +106,7 @@
>  #define NR_CMDS		 128 /* #commands that can be outstanding */
>  #define MAX_CTLR 8
>  
> -#define CCISS_DMA_MASK	0xFFFFFFFF	/* 32 bit DMA */
> +#define CCISS_DMA_MASK 0xFFFFFFFFFFFFFFFF /* 64 bit DMA */
>  
>  static ctlr_info_t *hba[MAX_CTLR];
>  
> @@ -2861,17 +2861,6 @@
>  	hba[i]->ctlr = i;
>  	hba[i]->pdev = pdev;
>  
> -	/* configure PCI DMA stuff */
> -	if (!pci_set_dma_mask(pdev, (u64) 0xffffffffffffffff))
> -		printk("cciss: using DAC cycles\n");
> -	else if (!pci_set_dma_mask(pdev, (u64) 0xffffffff))
> -		printk("cciss: not using DAC cycles\n");
> -	else {
> -		printk("cciss: no suitable DMA available\n");
> -		free_hba(i);
> -		return -ENODEV;
> -	}
> -		


this is wrong. The code there is EXACTLY what is needed as per
Documentation/DMA-mapping.txt, removing it is a bug.


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]