svn commit: r183981 - head/sys/dev/ata/chipsets

[John Baldwin]

On Wednesday 22 October 2008 03:47:21 am Søren Schmidt wrote:
> Well, it has been busy around here..
> 
> Anyhow, I think this is an overly pessimistic change, I've not seen  
> nor heard about problems with the 63K transfer size on anything but  
> the HT1000 based machines. It seems that all serverworks chips are  
> bugridden in some way or another, just avoid :)

Unfortunately I have to support existing machines, so I'm stuck with them.  
Given Scott's comments, it seems that the 1k runt transfers are probably 
non-optimal anyway.

> On the MIO change, the original docs I got from serverworks states  
> that PCI id to not support MIO mode and be for compat/legacy setups,  
> so this might produce some hickups if thats true.
> However their docs are often as flawed as their silicon so, guess this  
> is up to the board/BIOS producer to decide.

Note that the Linux driver does use MIO, and w/o MIO it can't see 4 SATA 
ports.  This has been in use on ${BIGNUM} Dell 1435 machines for over a year 
btw.

> -Søren
> 
> On 17Oct, 2008, at 18:03 , John Baldwin wrote:
> 
> > Author: jhb
> > Date: Fri Oct 17 16:03:37 2008
> > New Revision: 183981
> > URL: http://svn.freebsd.org/changeset/base/183981
> >
> > Log:
> >  - For chipsets that can't do 64k transfers, fall back to 32k  
> > transfers
> >    (still a power of 2) rather than 63k transfers.  Even with 63k  
> > transfers
> >    some machines (such as Dell SC1435's) were experiencing chronic  
> > data
> >    corruption.
> >  - Use the MIO method to talk to the Serverworks HT1000_S1 SATA  
> > controller
> >    like all the other SATA controllers rather than the compat PATA
> >    method.  This lets the controller see all 4 SATA ports and also
> >    matches the behavior of the Linux driver.
> >
> >  Silence from:	sos
> >  MFC after:	3 days
> >
> > Modified:
> >  head/sys/dev/ata/chipsets/ata-cyrix.c
> >  head/sys/dev/ata/chipsets/ata-marvell.c
> >  head/sys/dev/ata/chipsets/ata-national.c
> >  head/sys/dev/ata/chipsets/ata-serverworks.c
> >
> > Modified: head/sys/dev/ata/chipsets/ata-cyrix.c
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > ======================================================================
> > --- head/sys/dev/ata/chipsets/ata-cyrix.c	Fri Oct 17 15:11:12 2008	 
> > (r183980)
> > +++ head/sys/dev/ata/chipsets/ata-cyrix.c	Fri Oct 17 16:03:37 2008	 
> > (r183981)
> > @@ -109,7 +109,7 @@ ata_cyrix_setmode(device_t dev, int mode
> > 	/* dont try to set the mode if we dont have the resource */
> > 	if (ctlr->r_res1) {
> > 	    ch->dma.alignment = 16;
> > -	    ch->dma.max_iosize = 126 * DEV_BSIZE;
> > +	    ch->dma.max_iosize = 64 * DEV_BSIZE;
> >
> > 	    if (mode >= ATA_UDMA0) {
> > 		ATA_OUTL(ch->r_io[ATA_BMCMD_PORT].res,
> >
> > Modified: head/sys/dev/ata/chipsets/ata-marvell.c
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > ======================================================================
> > --- head/sys/dev/ata/chipsets/ata-marvell.c	Fri Oct 17 15:11:12 2008	 
> > (r183980)
> > +++ head/sys/dev/ata/chipsets/ata-marvell.c	Fri Oct 17 16:03:37 2008	 
> > (r183981)
> > @@ -536,7 +536,7 @@ ata_marvell_edma_dmainit(device_t dev)
> > 	ch->dma.max_address = BUS_SPACE_MAXADDR;
> >
> >     /* chip does not reliably do 64K DMA transfers */
> > -    ch->dma.max_iosize = 126 * DEV_BSIZE;
> > +    ch->dma.max_iosize = 64 * DEV_BSIZE;
> > }
> >
> > ATA_DECLARE_DRIVER(ata_marvell);
> >
> > Modified: head/sys/dev/ata/chipsets/ata-national.c
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > ======================================================================
> > --- head/sys/dev/ata/chipsets/ata-national.c	Fri Oct 17 15:11:12  
> > 2008	(r183980)
> > +++ head/sys/dev/ata/chipsets/ata-national.c	Fri Oct 17 16:03:37  
> > 2008	(r183981)
> > @@ -101,7 +101,7 @@ ata_national_setmode(device_t dev, int m
> >     int error;
> >
> >     ch->dma.alignment = 16;
> > -    ch->dma.max_iosize = 126 * DEV_BSIZE;
> > +    ch->dma.max_iosize = 64 * DEV_BSIZE;
> >
> >     mode = ata_limit_mode(dev, mode, ATA_UDMA2);
> >
> >
> > Modified: head/sys/dev/ata/chipsets/ata-serverworks.c
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > = 
> > ======================================================================
> > --- head/sys/dev/ata/chipsets/ata-serverworks.c	Fri Oct 17 15:11:12  
> > 2008	(r183980)
> > +++ head/sys/dev/ata/chipsets/ata-serverworks.c	Fri Oct 17 16:03:37  
> > 2008	(r183981)
> > @@ -79,7 +79,7 @@ ata_serverworks_probe(device_t dev)
> >      { ATA_CSB6,      0x00, SWKS_100, 0, ATA_UDMA5, "CSB6" },
> >      { ATA_CSB6_1,    0x00, SWKS_66,  0, ATA_UDMA4, "CSB6" },
> >      { ATA_HT1000,    0x00, SWKS_100, 0, ATA_UDMA5, "HT1000" },
> > -     { ATA_HT1000_S1, 0x00, SWKS_100, 4, ATA_SA150, "HT1000" },
> > +     { ATA_HT1000_S1, 0x00, SWKS_MIO, 4, ATA_SA150, "HT1000" },
> >      { ATA_HT1000_S2, 0x00, SWKS_MIO, 4, ATA_SA150, "HT1000" },
> >      { ATA_K2,        0x00, SWKS_MIO, 4, ATA_SA150, "K2" },
> >      { ATA_FRODO4,    0x00, SWKS_MIO, 4, ATA_SA150, "Frodo4" },
> > @@ -184,7 +184,7 @@ ata_serverworks_allocate(device_t dev)
> >     ch->hw.tf_write = ata_serverworks_tf_write;
> >
> >     /* chip does not reliably do 64K DMA transfers */
> > -    ch->dma.max_iosize = 126 * DEV_BSIZE;
> > +    ch->dma.max_iosize = 64 * DEV_BSIZE;
> >
> >     return 0;
> > }
> >
> 
> -Søren
> 
> 
> 
> 
> 
> 
> 



-- 
John Baldwin