[PATCH] fsl-rio: Correct IECSR register clear value

[Liu Gang-B34182]

Hi, Andrew,

Thanks for your comments.

I think you are right and a more detailed description will be better.

This bug causes the IECSR register clear failure.
In this case, the RETE (retry error threshold exceeded) interrupt will be generated and cannot be cleared.
So the related ISR may be called persistently.

Thanks again!

Regards,

Liu Gang

-----Original Message-----
From: Andrew Morton [mailto:akpm at linux-foundation.org] 
Sent: Thursday, August 25, 2011 2:54 AM
To: Liu Gang-B34182
Cc: 'linuxppc-dev at lists.ozlabs.org'; Li Yang-R58472; Gala Kumar-B11780; Zang Roy-R61911
Subject: Re: [PATCH] fsl-rio: Correct IECSR register clear value

On Wed, 24 Aug 2011 09:31:21 +0000
Liu Gang-B34182 <B34182 at freescale.com> wrote:

> Hi, Andrew,
> 
> Thank you for applying the patch "[PATCH] rio: Use discovered bit to test if enumeration is complete "!
> 
> So far the following patch "[PATCH] fsl-rio: Correct IECSR register clear value " has no comment or response.

Perhaps because you didn't tell anyone what the patch does.

> To: linuxppc-dev at ozlabs.org
> Cc: akpm at linux-foundation.org; Li Yang-R58472; Gala Kumar-B11780; Zang 
> Roy-R61911; Liu Gang-B34182; Liu Gang-B34182
> Subject: [PATCH] fsl-rio: Correct IECSR register clear value
> 
> The RETE bit in IECSR is cleared by writing a 1 to it.
> 
> Signed-off-by: Liu Gang <Gang.Liu at freescale.com>
> ---
>  arch/powerpc/sysdev/fsl_rio.c |    5 +++--
>  1 files changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/powerpc/sysdev/fsl_rio.c 
> b/arch/powerpc/sysdev/fsl_rio.c index b3fd081..cdd765b 100644
> --- a/arch/powerpc/sysdev/fsl_rio.c
> +++ b/arch/powerpc/sysdev/fsl_rio.c
> @@ -54,6 +54,7 @@
>  #define ODSR_CLEAR		0x1c00
>  #define LTLEECSR_ENABLE_ALL	0xFFC000FC
>  #define ESCSR_CLEAR		0x07120204
> +#define IECSR_CLEAR		0x80000000
>  
>  #define RIO_PORT1_EDCSR		0x0640
>  #define RIO_PORT2_EDCSR		0x0680
> @@ -1089,11 +1090,11 @@ static void port_error_handler(struct 
> rio_mport *port, int offset)
>  
>  	if (offset == 0) {
>  		out_be32((u32 *)(rio_regs_win + RIO_PORT1_EDCSR), 0);
> -		out_be32((u32 *)(rio_regs_win + RIO_PORT1_IECSR), 0);
> +		out_be32((u32 *)(rio_regs_win + RIO_PORT1_IECSR), IECSR_CLEAR);
>  		out_be32((u32 *)(rio_regs_win + RIO_ESCSR), ESCSR_CLEAR);
>  	} else {
>  		out_be32((u32 *)(rio_regs_win + RIO_PORT2_EDCSR), 0);
> -		out_be32((u32 *)(rio_regs_win + RIO_PORT2_IECSR), 0);
> +		out_be32((u32 *)(rio_regs_win + RIO_PORT2_IECSR), IECSR_CLEAR);
>  		out_be32((u32 *)(rio_regs_win + RIO_PORT2_ESCSR), ESCSR_CLEAR);
>  	}
>  }

Apparently it fixes some bug.  But because you didn't tell us what the user-visible effects of that bug are, I am unable to determine what kernel versions (if any) the patch should be merged into.