public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Benny Halevy <bhalevy@panasas.com>
To: Jens Axboe <jens.axboe@oracle.com>
Cc: torvalds@linux-foundation.org, linux-kernel@vger.kernel.org,
	mingo@elte.hu
Subject: Re: [PATCH 08/10] [SG] Update arch/ to use sg helpers
Date: Mon, 22 Oct 2007 23:10:52 +0200	[thread overview]
Message-ID: <471D11DC.20007@panasas.com> (raw)
In-Reply-To: <1193076664-13652-9-git-send-email-jens.axboe@oracle.com>

It looks like it could be nice to define and use a helper for
page_address(sg_page(sg)) (although 11 call sites could use it
after this patch)

#define sg_pgaddr(sg) page_address(sg_page(sg))

Note that mips sg_{un,}map_sg checked for page_address(sg->page) != 0
before calling __dma_sync(addr + sg->offset, sg->length, direction)
and you changed it to addr = (unsigned long) sg_virt(sg) which
takes sg->offset into account.  That said I'm not sure if the original
code was correct for the (page_address(sg->page) == 0 && sg->offset != 0)
case...


On Oct. 22, 2007, 20:11 +0200, Jens Axboe <jens.axboe@oracle.com> wrote:
<snip>

> diff --git a/arch/mips/mm/dma-default.c b/arch/mips/mm/dma-default.c
> index 98b5e5b..b0b034c 100644
> --- a/arch/mips/mm/dma-default.c
> +++ b/arch/mips/mm/dma-default.c
> @@ -165,12 +165,11 @@ int dma_map_sg(struct device *dev, struct scatterlist *sg, int nents,
>  	for (i = 0; i < nents; i++, sg++) {
>  		unsigned long addr;
>  
> -		addr = (unsigned long) page_address(sg->page);
> +		addr = (unsigned long) sg_virt(sg);
>  		if (!plat_device_is_coherent(dev) && addr)
> -			__dma_sync(addr + sg->offset, sg->length, direction);
> +			__dma_sync(addr, sg->length, direction);
>  		sg->dma_address = plat_map_dma_mem(dev,
> -				                   (void *)(addr + sg->offset),
> -						   sg->length);
> +				                   (void *)addr, sg->length);
>  	}
>  
>  	return nents;
> @@ -223,10 +222,9 @@ void dma_unmap_sg(struct device *dev, struct scatterlist *sg, int nhwentries,
>  	for (i = 0; i < nhwentries; i++, sg++) {
>  		if (!plat_device_is_coherent(dev) &&
>  		    direction != DMA_TO_DEVICE) {
> -			addr = (unsigned long) page_address(sg->page);
> +			addr = (unsigned long) sg_virt(sg);
>  			if (addr)
> -				__dma_sync(addr + sg->offset, sg->length,
> -				           direction);
> +				__dma_sync(addr, sg->length, direction);
>  		}
>  		plat_unmap_dma_mem(sg->dma_address);
>  	}

<snip>

> diff --git a/arch/x86/kernel/pci-calgary_64.c b/arch/x86/kernel/pci-calgary_64.c
> index 5098f58..1a20fe3 100644
> --- a/arch/x86/kernel/pci-calgary_64.c
> +++ b/arch/x86/kernel/pci-calgary_64.c
> @@ -411,8 +411,10 @@ static int calgary_nontranslate_map_sg(struct device* dev,
>  	int i;
>  
>  	for_each_sg(sg, s, nelems, i) {
> -		BUG_ON(!s->page);
> -		s->dma_address = virt_to_bus(page_address(s->page) +s->offset);
> +		struct page *p = sg_page(s);
> +
> +		BUG_ON(!p);

why not just BUG_ON(!sg_page(s))?

> +		s->dma_address = virt_to_bus(sg_virt(s));
>  		s->dma_length = s->length;
>  	}
>  	return nelems;


  reply	other threads:[~2007-10-22 21:11 UTC|newest]

Thread overview: 85+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-10-22 18:10 [PATCH 00/10] SG updates Jens Axboe
2007-10-22 18:10 ` [PATCH 01/10] [SG] Add helpers for manipulating SG entries Jens Axboe
2007-10-22 18:10 ` [PATCH 02/10] [SG] Update block layer to use sg helpers Jens Axboe
2007-10-23  5:13   ` Heiko Carstens
2007-10-23  5:16     ` Jens Axboe
2007-10-23  5:42       ` [PATCH] fix ll_rw_blk.c build on s390 Heiko Carstens
2007-10-23  5:44       ` [PATCH] net: fix xfrm build - missing scatterlist.h include Heiko Carstens
2007-10-23  7:28         ` Jens Axboe
2007-10-23 14:32   ` [PATCH 02/10] [SG] Update block layer to use sg helpers John Stoffel
2007-10-22 18:10 ` [PATCH 03/10] [SG] Update crypto/ to " Jens Axboe
2007-10-22 18:10 ` [PATCH 04/10] [SG] Update drivers to use " Jens Axboe
2007-10-23  6:28   ` Heiko Carstens
2007-10-23  7:14     ` Jens Axboe
2007-10-23  7:16       ` Heiko Carstens
2007-10-22 18:10 ` [PATCH 05/10] [SG] Update fs/ " Jens Axboe
2007-10-22 18:11 ` [PATCH 06/10] [SG] Update net/ " Jens Axboe
2007-10-23 10:44   ` Christian Borntraeger
2007-10-23 10:45     ` Jens Axboe
2007-10-22 18:11 ` [PATCH 07/10] [SG] Update swiotlb " Jens Axboe
2007-10-22 18:11 ` [PATCH 08/10] [SG] Update arch/ " Jens Axboe
2007-10-22 21:10   ` Benny Halevy [this message]
2007-10-23  7:26     ` Jens Axboe
2007-10-22 18:11 ` [PATCH 09/10] Change table chaining layout Jens Axboe
2007-10-22 19:39   ` Geert Uytterhoeven
2007-10-22 19:49     ` Linus Torvalds
2007-10-22 19:52       ` Jens Axboe
2007-10-22 20:16       ` Alan Cox
2007-10-22 20:38         ` Matt Mackall
2007-10-22 20:44         ` Linus Torvalds
2007-10-22 21:43           ` Alan Cox
2007-10-22 21:47             ` Linus Torvalds
2007-10-23  0:07               ` David Miller
2007-10-23  7:18               ` Geert Uytterhoeven
2007-10-23  9:29               ` Boaz Harrosh
2007-10-23  9:41                 ` Jens Axboe
2007-10-23  9:50                   ` Boaz Harrosh
2007-10-23  9:55                     ` Jens Axboe
2007-10-23 10:23                       ` Boaz Harrosh
2007-10-23 10:29                         ` Jens Axboe
2007-10-23 15:22                         ` Linus Torvalds
2007-10-24  8:05                           ` Jens Axboe
2007-10-24  9:03                             ` Geert Uytterhoeven
2007-10-24  9:12                               ` Jens Axboe
2007-10-24 13:35                                 ` Olivier Galibert
2007-10-24 13:38                                   ` Jens Axboe
2007-10-24 13:45                                     ` Olivier Galibert
2007-10-24 15:16                                 ` Linus Torvalds
2007-10-25  8:40                           ` Rusty Russell
2007-10-25  9:11                             ` Jens Axboe
2007-10-25 11:54                               ` Rusty Russell
2007-10-26  0:03                                 ` Rusty Russell
2007-10-25 15:40                             ` Linus Torvalds
2007-10-25 16:03                               ` Benny Halevy
2007-10-26  5:01                               ` Paul Mackerras
2007-10-26 14:52                                 ` Linus Torvalds
2007-10-26 17:28                                   ` Jens Axboe
2007-11-05  6:11                                   ` [RFC PATCH 1/2] sg_ring instead of scatterlist chaining Rusty Russell
2007-11-05  6:15                                     ` [RFC PATCH 2/2] sg_ring instead of scatterlist chaining in virtio Rusty Russell
2007-11-05 16:40                                     ` [RFC PATCH 1/2] sg_ring instead of scatterlist chaining Randy Dunlap
2007-10-23 10:33                   ` [PATCH 09/10] Change table chaining layout Ingo Molnar
2007-10-23 10:56                     ` Jens Axboe
2007-10-23 11:27                       ` Ingo Molnar
2007-10-23 19:23                         ` Geert Uytterhoeven
2007-10-23 21:46                           ` Jens Axboe
2007-10-24  6:56                           ` Jens Axboe
2007-10-22 21:16         ` Benny Halevy
2007-10-22 21:21         ` Jeff Garzik
2007-10-22 21:47           ` Matt Mackall
2007-10-22 22:52             ` Alan Cox
2007-10-22 23:46               ` Matt Mackall
2007-10-23  0:11                 ` Jeff Garzik
2007-10-23  4:09   ` powerpc: Fix fallout from sg_page() changes Olof Johansson
2007-10-23  4:31     ` IB/ehca: Fix sg_page() fallout Olof Johansson
2007-10-23  5:05       ` Jens Axboe
2007-10-23  5:54         ` Olof Johansson
2007-10-23  7:12           ` Jens Axboe
2007-10-23  7:13     ` powerpc: Fix fallout from sg_page() changes Jens Axboe
2007-10-23 17:08   ` [PATCH 09/10] Change table chaining layout Boaz Harrosh
2007-10-23 18:33     ` Jens Axboe
2007-10-23 19:56       ` Andi Kleen
2007-10-23 20:20         ` Jens Axboe
2007-10-23 20:57           ` Andi Kleen
2007-10-23 21:44             ` Jens Axboe
2007-10-22 18:11 ` [PATCH 10/10] Add CONFIG_DEBUG_SG sg validation Jens Axboe
2007-10-23 14:48 ` [PATCH][SG] fix typo in ps3rom.c Arnd Bergmann

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=471D11DC.20007@panasas.com \
    --to=bhalevy@panasas.com \
    --cc=jens.axboe@oracle.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=torvalds@linux-foundation.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox