Re: [PATCH libata#upstream] sata_promise: decode and report error reasons

[Jeff Garzik <jgarzik@pobox.com>]

Mikael Pettersson wrote:
> This patch adds much needed error reason decoding and
> reporting to sata_promise. It's simplistic but should
> log all relevant error info the controller provides.
> 
> Signed-off-by: Mikael Pettersson <mikpe@it.uu.se>
> ---
>  drivers/ata/sata_promise.c |   64 ++++++++++++++++++++++++++++++++++++++-------
>  1 files changed, 55 insertions(+), 9 deletions(-)
> 
> Changes since the preliminary version:
> * dropped SError decoding, as all bits there are standard
> * PDC2_HTO_ERR is mapped to AC_ERR_HOST_BUS not AC_ERR_TIMEOUT
> * PDC_UNDERRUN_ERR and PDC_OVERRUN_ERR are mapped to AC_ERR_HSM
>   not AC_ERR_ATA_BUS
> * added brief explanations to error constant declarations
> * some cleanups and simplifications
> 
> --- linux-2.6.21-rc3+upstream/drivers/ata/sata_promise.c.~1~	2007-03-13 23:41:43.000000000 +0100
> +++ linux-2.6.21-rc3+upstream/drivers/ata/sata_promise.c	2007-03-14 00:48:49.000000000 +0100
> @@ -45,7 +45,7 @@
>  #include "sata_promise.h"
>  
>  #define DRV_NAME	"sata_promise"
> -#define DRV_VERSION	"2.03"
> +#define DRV_VERSION	"2.04"
>  
>  
>  enum {
> @@ -70,8 +70,23 @@ enum {
>  	PDC_TBG_MODE		= 0x41C, /* TBG mode (not SATAII) */
>  	PDC_SLEW_CTL		= 0x470, /* slew rate control reg (not SATAII) */
>  
> -	PDC_ERR_MASK		= (1<<19) | (1<<20) | (1<<21) | (1<<22) |
> -				  (1<<8) | (1<<9) | (1<<10),
> +	/* PDC_GLOBAL_CTL bit definitions */
> +	PDC_PH_ERR		= (1 <<  8), /* PCI error while loading packet */
> +	PDC_SH_ERR		= (1 <<  9), /* PCI error while loading S/G table */
> +	PDC_DH_ERR		= (1 << 10), /* PCI error while loading data */
> +	PDC2_HTO_ERR		= (1 << 12), /* host bus timeout */
> +	PDC2_ATA_HBA_ERR	= (1 << 13), /* error during SATA DATA FIS transmission */
> +	PDC2_ATA_DMA_CNT_ERR	= (1 << 14), /* DMA DATA FIS size differs from S/G count */
> +	PDC_OVERRUN_ERR		= (1 << 19), /* S/G byte count larger than HD requires */
> +	PDC_UNDERRUN_ERR	= (1 << 20), /* S/G byte count less than HD requires */
> +	PDC_DRIVE_ERR		= (1 << 21), /* drive error */
> +	PDC_PCI_SYS_ERR		= (1 << 22), /* PCI system error */
> +	PDC1_PCI_PARITY_ERR	= (1 << 23), /* PCI parity error (from SATA150 driver) */
> +	PDC1_ERR_MASK		= PDC1_PCI_PARITY_ERR,
> +	PDC2_ERR_MASK		= PDC2_HTO_ERR | PDC2_ATA_HBA_ERR | PDC2_ATA_DMA_CNT_ERR,
> +	PDC_ERR_MASK		= (PDC_PH_ERR | PDC_SH_ERR | PDC_DH_ERR | PDC_OVERRUN_ERR
> +				   | PDC_UNDERRUN_ERR | PDC_DRIVE_ERR | PDC_PCI_SYS_ERR
> +				   | PDC1_ERR_MASK | PDC2_ERR_MASK),
>  
>  	board_2037x		= 0,	/* FastTrak S150 TX2plus */
>  	board_20319		= 1,	/* FastTrak S150 TX4 */
> @@ -615,17 +630,48 @@ static void pdc_post_internal_cmd(struct
>  		pdc_reset_port(ap);
>  }
>  
> +static void pdc_error_intr(struct ata_port *ap, struct ata_queued_cmd *qc,
> +			   u32 port_status, u32 err_mask)
> +{
> +	struct ata_eh_info *ehi = &ap->eh_info;
> +	unsigned int ac_err_mask = 0;
> +
> +	ata_ehi_clear_desc(ehi);
> +	ata_ehi_push_desc(ehi, "port_status 0x%08x", port_status);
> +	port_status &= err_mask;
> +
> +	if (port_status & PDC_DRIVE_ERR)
> +		ac_err_mask |= AC_ERR_DEV;
> +	if (port_status & (PDC_OVERRUN_ERR | PDC_UNDERRUN_ERR))
> +		ac_err_mask |= AC_ERR_HSM;
> +	if (port_status & (PDC2_ATA_HBA_ERR | PDC2_ATA_DMA_CNT_ERR))
> +		ac_err_mask |= AC_ERR_ATA_BUS;
> +	if (port_status & (PDC_PH_ERR | PDC_SH_ERR | PDC_DH_ERR | PDC2_HTO_ERR
> +			   | PDC_PCI_SYS_ERR | PDC1_PCI_PARITY_ERR))
> +		ac_err_mask |= AC_ERR_HOST_BUS;
> +
> +	ehi->action |= ATA_EH_SOFTRESET;
> +	qc->err_mask |= ac_err_mask;
> +	ata_port_freeze(ap);
> +}
> +
>  static inline unsigned int pdc_host_intr( struct ata_port *ap,
>                                            struct ata_queued_cmd *qc)
>  {
>  	unsigned int handled = 0;
> -	u32 tmp;
> -	void __iomem *mmio = ap->ioaddr.cmd_addr + PDC_GLOBAL_CTL;
> +	void __iomem *port_mmio = ap->ioaddr.cmd_addr;
> +	struct pdc_host_priv *hp = ap->host->private_data;
> +	u32 port_status, err_mask;
>  
> -	tmp = readl(mmio);
> -	if (tmp & PDC_ERR_MASK) {
> -		qc->err_mask |= AC_ERR_DEV;
> -		pdc_reset_port(ap);

AFAICS pdc_reset_port() will no longer be called, after this change.

Since you are freezing the port with ata_port_freeze() before 
pdc_error_handler() is called, the one remaining call to 
pdc_reset_port() never occurs:

         if (!(ap->pflags & ATA_PFLAG_FROZEN))
                 pdc_reset_port(ap);

And Promise hardware definitely needs to be "kicked" by pdc_reset_port() 
after most errors.

Also, PCI error are controller-wide, so ideally you should presume that 
all transactions on all ports are bad, and reset everything.  See 
mv_pci_error() in sata_mv.c in libata-dev.git#mv-eh for mostly-generic 
code that initiates error handling on all ports, rather than just one.

	Jeff