public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: Karsten Keil <kkeil@suse.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	linux-kernel@vger.kernel.org, isdn4linux@listserv.isdn4linux.de
Subject: Re: [PATCH 3/4] Fix remaining big endian issue of hfcmulti
Date: Tue, 05 Aug 2008 14:29:48 +1000	[thread overview]
Message-ID: <1217910588.24157.151.camel@pasglop> (raw)
In-Reply-To: <20080802151532.DE017A3C09@pingi.kke.suse.de>

On Sat, 2008-08-02 at 16:35 +0200, Karsten Keil wrote:
> The driver was not so bad at big endian at all, only the optimised fifo
> read/write functions need a fix, with this fix the driver works on
> a pegasus PPC machine.

This is however very broken... IE, you should instead use iomap
and thus get ioreadXX_rep() and writeXX_rep() (XX = 16 or 32) that
will do the right thing for you. IE, they will do the right amount
of memory barriers and will avoid the unnecessary double-swapping
you are doing there.

Your code will happen to "work" but it's just way too ugly especially
since you are basically re-implementing what iomap already gives you.

Ben.

> Signed-off-by: Karsten Keil <kkeil@suse.de>
> ---
>  drivers/isdn/hardware/mISDN/hfcmulti.c |   27 +++++++++++++--------------
>  1 files changed, 13 insertions(+), 14 deletions(-)
> 
> diff --git a/drivers/isdn/hardware/mISDN/hfcmulti.c b/drivers/isdn/hardware/mISDN/hfcmulti.c
> index 10144e8..e36360a 100644
> --- a/drivers/isdn/hardware/mISDN/hfcmulti.c
> +++ b/drivers/isdn/hardware/mISDN/hfcmulti.c
> @@ -140,7 +140,7 @@
>   * #define HFC_REGISTER_DEBUG
>   */
>  
> -static const char *hfcmulti_revision = "2.00";
> +static const char *hfcmulti_revision = "2.01";
>  
>  #include <linux/module.h>
>  #include <linux/pci.h>
> @@ -427,12 +427,12 @@ write_fifo_regio(struct hfc_multi *hc, u_char *data, int len)
>  {
>  	outb(A_FIFO_DATA0, (hc->pci_iobase)+4);
>  	while (len>>2) {
> -		outl(*(u32 *)data, hc->pci_iobase);
> +		outl(cpu_to_le32(*(u32 *)data), hc->pci_iobase);
>  		data += 4;
>  		len -= 4;
>  	}
>  	while (len>>1) {
> -		outw(*(u16 *)data, hc->pci_iobase);
> +		outw(cpu_to_le16(*(u16 *)data), hc->pci_iobase);
>  		data += 2;
>  		len -= 2;
>  	}
> @@ -447,17 +447,19 @@ void
>  write_fifo_pcimem(struct hfc_multi *hc, u_char *data, int len)
>  {
>  	while (len>>2) {
> -		writel(*(u32 *)data, (hc->pci_membase)+A_FIFO_DATA0);
> +		writel(cpu_to_le32(*(u32 *)data),
> +			hc->pci_membase + A_FIFO_DATA0);
>  		data += 4;
>  		len -= 4;
>  	}
>  	while (len>>1) {
> -		writew(*(u16 *)data, (hc->pci_membase)+A_FIFO_DATA0);
> +		writew(cpu_to_le16(*(u16 *)data),
> +			hc->pci_membase + A_FIFO_DATA0);
>  		data += 2;
>  		len -= 2;
>  	}
>  	while (len) {
> -		writeb(*data, (hc->pci_membase)+A_FIFO_DATA0);
> +		writeb(*data, hc->pci_membase + A_FIFO_DATA0);
>  		data++;
>  		len--;
>  	}
> @@ -468,12 +470,12 @@ read_fifo_regio(struct hfc_multi *hc, u_char *data, int len)
>  {
>  	outb(A_FIFO_DATA0, (hc->pci_iobase)+4);
>  	while (len>>2) {
> -		*(u32 *)data = inl(hc->pci_iobase);
> +		*(u32 *)data = le32_to_cpu(inl(hc->pci_iobase));
>  		data += 4;
>  		len -= 4;
>  	}
>  	while (len>>1) {
> -		*(u16 *)data = inw(hc->pci_iobase);
> +		*(u16 *)data = le16_to_cpu(inw(hc->pci_iobase));
>  		data += 2;
>  		len -= 2;
>  	}
> @@ -490,18 +492,18 @@ read_fifo_pcimem(struct hfc_multi *hc, u_char *data, int len)
>  {
>  	while (len>>2) {
>  		*(u32 *)data =
> -			readl((hc->pci_membase)+A_FIFO_DATA0);
> +			le32_to_cpu(readl(hc->pci_membase + A_FIFO_DATA0));
>  		data += 4;
>  		len -= 4;
>  	}
>  	while (len>>1) {
>  		*(u16 *)data =
> -			readw((hc->pci_membase)+A_FIFO_DATA0);
> +			le16_to_cpu(readw(hc->pci_membase + A_FIFO_DATA0));
>  		data += 2;
>  		len -= 2;
>  	}
>  	while (len) {
> -		*data = readb((hc->pci_membase)+A_FIFO_DATA0);
> +		*data = readb(hc->pci_membase + A_FIFO_DATA0);
>  		data++;
>  		len--;
>  	}
> @@ -5251,9 +5253,6 @@ HFCmulti_init(void)
>  	if (debug & DEBUG_HFCMULTI_INIT)
>  		printk(KERN_DEBUG "%s: init entered\n", __func__);
>  
> -#ifdef __BIG_ENDIAN
> -#error "not running on big endian machines now"
> -#endif
>  	hfc_interrupt = symbol_get(ztdummy_extern_interrupt);
>  	register_interrupt = symbol_get(ztdummy_register_interrupt);
>  	unregister_interrupt = symbol_get(ztdummy_unregister_interrupt);


  parent reply	other threads:[~2008-08-05  4:30 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-08-02 14:35 [PATCH 3/4] Fix remaining big endian issue of hfcmulti Karsten Keil
2008-08-04 12:03 ` David Woodhouse
2008-08-04 14:29   ` Karsten Keil
2008-08-05  4:29 ` Benjamin Herrenschmidt [this message]
2008-08-05 11:31   ` Karsten Keil
2008-08-05 13:04     ` Benjamin Herrenschmidt
2008-08-05 17:25       ` Karsten Keil
2008-08-05 18:42         ` Linus Torvalds
2008-08-05 21:02           ` Karsten Keil
2008-08-05 21:23             ` Sean MacLennan
2008-08-05 21:37               ` Linus Torvalds
2008-08-05 21:59                 ` Sean MacLennan
2008-08-05 23:04                   ` Benjamin Herrenschmidt
2008-08-05 23:38                     ` Sean MacLennan
2008-08-06  0:18                   ` Karsten Keil
2008-08-06  0:33                     ` Benjamin Herrenschmidt
2008-08-05 21:46               ` Benjamin Herrenschmidt
2008-08-05 21:45             ` Benjamin Herrenschmidt
2008-08-05 21:43         ` Benjamin Herrenschmidt
  -- strict thread matches above, loose matches on Subject: below --
2008-08-07 11:01 AW: " Andreas.Eversberg
2008-08-07 13:46 ` Karsten Keil

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=1217910588.24157.151.camel@pasglop \
    --to=benh@kernel.crashing.org \
    --cc=isdn4linux@listserv.isdn4linux.de \
    --cc=kkeil@suse.de \
    --cc=linux-kernel@vger.kernel.org \
    --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