linux-kernel - Re: [PATCH v2 77/71] ncr5380: Fix wait for 53C80 registers registers after PDMA

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Mon, 7 Dec 2015 14:16:14 +1100 (AEDT)
From:	Finn Thain <fthain@...egraphics.com.au>
To:	Ondrej Zary <linux@...nbow-software.org>
cc:	Michael Schmitz <schmitzmic@...il.com>, linux-m68k@...r.kernel.org,
	linux-scsi@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 77/71] ncr5380: Fix wait for 53C80 registers registers
 after PDMA


On Mon, 7 Dec 2015, Ondrej Zary wrote:

> The check for 53C80 registers accessibility was commented out because
> it was broken (inverted). Fix and enable it.
> 
> Signed-off-by: Ondrej Zary <linux@...nbow-software.org>
> ---
>  drivers/scsi/g_NCR5380.c |   37 ++++++-------------------------------
>  1 file changed, 6 insertions(+), 31 deletions(-)
> 
> diff --git a/drivers/scsi/g_NCR5380.c b/drivers/scsi/g_NCR5380.c
> index 038dddf..a7479c6 100644
> --- a/drivers/scsi/g_NCR5380.c
> +++ b/drivers/scsi/g_NCR5380.c
> @@ -603,14 +603,10 @@ static inline int NCR5380_pread(struct Scsi_Host *instance, unsigned char *dst,
>  	if (!(NCR5380_read(hostdata->c400_ctl_status) & CSR_GATED_53C80_IRQ))
>  		printk("53C400r: no 53C80 gated irq after transfer");
>  
> -#if 0
> -	/*
> -	 *	DON'T DO THIS - THEY NEVER ARRIVE!
> -	 */
> -	printk("53C400r: Waiting for 53C80 registers\n");
> -	while (NCR5380_read(hostdata->c400_ctl_status) & CSR_53C80_REG)
> +	/* wait for 53C80 registers to be available */
> +	while (!(NCR5380_read(hostdata->c400_ctl_status) & CSR_53C80_REG))
>  		;

In the previous patch, udelay(4) was added to a CSR_GATED_53C80_IRQ 
polling loop. It is interesting that you don't need it here when polling 
CSR_53C80_REG.

> -#endif
> +
>  	if (!(NCR5380_read(BUS_AND_STATUS_REG) & BASR_END_DMA_TRANSFER))
>  		printk(KERN_ERR "53C400r: no end dma signal\n");
>  		
> @@ -632,7 +628,6 @@ static inline int NCR5380_pwrite(struct Scsi_Host *instance, unsigned char *src,
>  	struct NCR5380_hostdata *hostdata = shost_priv(instance);
>  	int blocks = len / 128;
>  	int start = 0;
> -	int i;
>  
>  	NCR5380_write(hostdata->c400_ctl_status, CSR_BASE);
>  	NCR5380_write(hostdata->c400_blk_cnt, blocks);
> @@ -681,36 +676,16 @@ static inline int NCR5380_pwrite(struct Scsi_Host *instance, unsigned char *src,
>  		blocks--;
>  	}
>  
> -#if 0
> -	printk("53C400w: waiting for registers to be available\n");
> -	THEY NEVER DO ! while (NCR5380_read(hostdata->c400_ctl_status) & CSR_53C80_REG);
> -	printk("53C400w: Got em\n");
> -#endif
> -
> -	/* Let's wait for this instead - could be ugly */
> -	/* All documentation says to check for this. Maybe my hardware is too
> -	 * fast. Waiting for it seems to work fine! KLL
> -	 */
> -	while (!(i = NCR5380_read(hostdata->c400_ctl_status) & CSR_GATED_53C80_IRQ)) {
> +	/* wait for 53C80 registers to be available */
> +	while (!(NCR5380_read(hostdata->c400_ctl_status) & CSR_53C80_REG)) {
>  		udelay(4); /* DTC436 chip hangs without this */

... based on the above, this udelay is probably not needed.

(Or perhaps it is only needed once, in between the final host_buf register 
access and the first ctl_status access??)

Is there any reference in the docs to timing sensitivity?

>  		/* FIXME - no timeout */
>  	}
>  
> -	/*
> -	 * I know. i is certainly != 0 here but the loop is new. See previous
> -	 * comment.
> -	 */

Thanks for cleaning up this mess!

-- 

> -	if (i) {
> -		if (!((i = NCR5380_read(BUS_AND_STATUS_REG)) & BASR_END_DMA_TRANSFER))
> -			printk(KERN_ERR "53C400w: No END OF DMA bit - WHOOPS! BASR=%0x\n", i);
> -	} else
> -		printk(KERN_ERR "53C400w: no 53C80 gated irq after transfer (last block)\n");
> -
> -#if 0
>  	if (!(NCR5380_read(BUS_AND_STATUS_REG) & BASR_END_DMA_TRANSFER)) {
>  		printk(KERN_ERR "53C400w: no end dma signal\n");
>  	}
> -#endif
> +
>  	while (!(NCR5380_read(TARGET_COMMAND_REG) & TCR_LAST_BYTE_SENT))
>  		; 	// TIMEOUT
>  	return 0;
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/