Re: sata_sx4 vs. Promise SX4060

[Jeff Garzik <jeff@garzik.org>]

Lasse Makholm wrote:
> 
> Hi,
> 
> I know it's been brought up here before, at least a few years ago
> (http://search.gmane.org/?query=sx4060&group=gmane.linux.ide) but here
> goes anyway...
> 
> I'm one of those semi-happy owners of a Promise SX4060 PATA RAID
> controller which is apparently only supported for some 2.4-something
> kernels...
> 
> Obviously I would a lot more happy with a 2.6 driver... :-)
> 
> So I've been googling high and low without much result, except finding
> out that the SX4060 is based on the same chip (the PDC20621) as the SATA
> 150 SX4, leading me to believe that the sata_sx4 driver is probably the
> best bet for getting the SX4060 working under 2.6 kernels...
> 
> Also, this comment in sata_sx4.c seems to support this suspicion:
> 
>     The SX4 behaves like a PATA chip, with no SATA controls or
>     knowledge whatsoever, leading to the presumption that
>     PATA<->SATA bridges exist on SX4 boards, external to the
>     PDC20621 chip itself.
> 
> All documentation I've been able to find on the devices indicate that 
> they are very similar...
> 
> I tried hacking the driver to recognize the SX4060 board number,
> essentially the same thing as described here:
> http://article.gmane.org/gmane.linux.ide/5435
> 
> ...with similar results. The controller is recognized and it also
> recognizes the attached drives, but it seems to have trouble 
> initializing the DIMM and with the ATA_SET_FEATURES command...
> 
> With ATA_DEBUG and ATA_DEBUG_VERBOSE, I get something like this (I have 
> the full log if someone wants to see it):
> 
> [  271.122187] pdc20621_dimm_init: Time Counter Register (0x44): 0xf0a6fb86
> [  271.122250] pdc20621_dimm_init: Num counters 0xf590479 (257492089)
> [  271.122309] pdc20621_dimm_init: 10 * Internal clk = 0x35a (858)
> [  271.122369] pdc20621_dimm_init: 10 * Internal clk * 33 = 0x6e9a (28314)
> [  271.122428] pdc20621_dimm_init: PLL F Param: 0x2f (47)
> [  271.122486] pdc20621_dimm_init: pci_status: 0x8a2f1824
> [  271.122721] pdc20621_dimm_init: Local DIMM Speed = 100
> [  271.124883] pdc20621_dimm_init: Local DIMM Size = 64MB
> [  271.125118] Local DIMM ECC Enabled
> [  271.226153] 0, 0,
> [  271.226220] 55, aa, Promise Not Yet Defined 1.1098
> [  271.226291] 55, aa, Promise Not Yet Defined 1.1098
> [  271.226524] pdc20621_dimm_init: Start ECC initialization
> [  281.972847] BUG: soft lockup - CPU#0 stuck for 11s! [modprobe:6153]
> Register and stack dump here...
> [  293.775043] BUG: soft lockup - CPU#0 stuck for 11s! [modprobe:6153]
> Register and stack dump here...
> [  305.581239] BUG: soft lockup - CPU#0 stuck for 11s! [modprobe:6153]
> Register and stack dump here...
> [  312.567607] pdc20621_dimm_init: Finish ECC initialization
> Lots of noise follows...
> [  312.768567] ata5: PATA max UDMA/44 mmio m1048576@0xfe900000 dimm 
> m32768@0xfe800000 port 0xfe900200 irq 17
> [  312.768571] ata6: PATA max UDMA/44 mmio m1048576@0xfe900000 dimm 
> m32768@0xfe800000 port 0xfe900280 irq 17
> [  312.768574] ata7: PATA max UDMA/44 mmio m1048576@0xfe900000 dimm 
> m32768@0xfe800000 port 0xfe900300 irq 17
> [  312.768577] ata8: PATA max UDMA/44 mmio m1048576@0xfe900000 dimm 
> m32768@0xfe800000 port 0xfe900380 irq 17
> More noise follows...
> [  342.918886] ata_port_flush_task: ENTER
> [  342.918947] ata5: ata_port_flush_task: EXIT
> [  342.918959] ata5.00: qc timeout (cmd 0xef)
> [  342.918962] ata_dev_set_xfermode: EXIT, err_mask=4
> [  342.919021] ata5.00: failed to set xfermode (err_mask=0x4)
> 
> Even more noise follows...
> 
> [  373.068199] ata5.00: failed to set xfermode (err_mask=0x4)
> [  373.068260] ata5.00: limiting speed to UDMA/44:PIO3
> [  373.068263] pdc_20621_phy_reset: ENTER
> [  373.068320] ata_bus_reset: ENTER, host 5, port 0
> [  373.068388] ata_bus_softreset: ata5: bus reset via SRST
> [  373.224003] ata_dev_classify: found ATA device by sig
> [  373.224068] ata_bus_reset: EXIT
> [  373.224125] ata5.00: ata_dev_read_id: ENTER
> [  373.224127] ata5: ata_dev_select: ENTER, device 0, wait 1
> [  373.224157] ata5: ata_dev_select: ENTER, device 0, wait 1
> [  373.224195] ata_tf_load: feat 0x0 nsect 0x0 lba 0x0 0x0 0x0
> [  373.224253] ata_tf_load: device 0xA0
> [  373.224321] ata_exec_command: ata5: cmd 0xEC
> [  373.231994] ata_hsm_move: ata5: protocol 2 task_state 2 (dev_stat 0x58)
> [  373.232057] ata_pio_sector: data read
> [  373.232419] ata_hsm_move: ata5: protocol 2 task_state 3 (dev_stat 0x50)
> [  373.232479] ata_hsm_move: ata5: dev 0 command complete, drv_stat 0x50
> [  373.232562] ata_port_flush_task: ENTER
> [  373.232618] ata5: ata_port_flush_task: EXIT
> [  373.232622] ata5.00: ata_dev_configure: ENTER
> [  373.232634] pdc20621_nodata_prep: ata5: ENTER
> [  373.232692] pdc20621_ata_pkt: ENTER, dimm_sg == 0x200180, 2097536
> [  373.232760] pdc20621_nodata_prep: ata pkt buf ofs 282, mmio copied
> [  373.232819] pdc20621_packet_start: ata5: ENTER
> [  373.232879] pdc20621_packet_start: submitted ofs 0x200100 (2097408), 
> seq 1
> [  373.232987] pdc20621_interrupt: ENTER
> [  373.233045] pdc20621_interrupt: mask == 0x2
> [  373.233103] pdc20621_interrupt: seq 1, port_no 0, ap de81c000, tmp 2
> [  373.233162] pdc20621_host_intr: ENTER
> [  373.233230] pdc20621_host_intr: BUS_NODATA (drv_stat 0x50)
> [  373.233297] pdc20621_interrupt: seq 2, port_no 1, ap de820000, tmp 0
> [  373.233357] pdc20621_interrupt: seq 3, port_no 2, ap de824000, tmp 0
> [  373.233416] pdc20621_interrupt: seq 4, port_no 3, ap de828000, tmp 0
> [  373.233475] pdc20621_interrupt: seq 5, port_no 0, ap de81c000, tmp 0
> [  373.233535] pdc20621_interrupt: seq 6, port_no 1, ap de820000, tmp 0
> [  373.233594] pdc20621_interrupt: seq 7, port_no 2, ap de824000, tmp 0
> [  373.233654] pdc20621_interrupt: seq 8, port_no 3, ap de828000, tmp 0
> [  373.233713] pdc20621_interrupt: mask == 0x2
> [  373.233769] pdc20621_interrupt: EXIT
> [  373.233828] ata_port_flush_task: ENTER
> [  373.233886] ata5: ata_port_flush_task: EXIT
> [  373.233890] ata5.00: ata_dev_configure: cfg 49:2f00 82:7c6b 83:7b09 
> 84:4003 85:7c69 86:3a01 87:4003 88:087f
> [  373.233894] ata_dump_id: 49==0x2f00  53==0x0007  63==0x0007 
> 64==0x0003  75==0x0000
> [  373.233964] ata_dump_id: 80==0x00fe  81==0x001e  82==0x7c6b 
> 83==0x7b09  84==0x4003
> [  373.234033] ata_dump_id: 88==0x087f  93==0x600b
> [  373.234092] ata5.00: ATA-7: Maxtor 6Y080P0, YAR41BW0, max UDMA/133
> [  373.234095] ata5.00: 160086528 sectors, multi 0: LBA
> [  373.234105] ata5.00: ata_dev_configure: EXIT, drv_stat = 0x50
> [  373.234110] ata_dev_set_xfermode: set features - xfer mode
> [  373.234168] pdc20621_nodata_prep: ata5: ENTER
> [  373.234226] pdc20621_ata_pkt: ENTER, dimm_sg == 0x200180, 2097536
> [  373.234292] pdc20621_nodata_prep: ata pkt buf ofs 282, mmio copied
> [  373.234351] pdc20621_packet_start: ata5: ENTER
> [  373.234411] pdc20621_packet_start: submitted ofs 0x200100 (2097408), 
> seq 1
> [  373.234520] pdc20621_interrupt: ENTER
> [  373.234577] pdc20621_interrupt: mask == 0x2
> [  373.234634] pdc20621_interrupt: seq 1, port_no 0, ap de81c000, tmp 2
> [  373.234694] pdc20621_interrupt: seq 2, port_no 1, ap de820000, tmp 0
> [  373.234754] pdc20621_interrupt: seq 3, port_no 2, ap de824000, tmp 0
> [  373.234813] pdc20621_interrupt: seq 4, port_no 3, ap de828000, tmp 0
> [  373.234872] pdc20621_interrupt: seq 5, port_no 0, ap de81c000, tmp 0
> [  373.234932] pdc20621_interrupt: seq 6, port_no 1, ap de820000, tmp 0
> [  373.234991] pdc20621_interrupt: seq 7, port_no 2, ap de824000, tmp 0
> [  373.235051] pdc20621_interrupt: seq 8, port_no 3, ap de828000, tmp 0
> [  373.235110] pdc20621_interrupt: mask == 0x2
> [  373.235166] pdc20621_interrupt: EXIT
> Above stuff seems to happen a couple of times...
> [  403.217240] ata_port_flush_task: ENTER
> [  403.217301] ata5: ata_port_flush_task: EXIT
> [  403.217313] ata5.00: qc timeout (cmd 0xef)
> [  403.217316] ata_dev_set_xfermode: EXIT, err_mask=4
> [  403.217375] ata5.00: failed to set xfermode (err_mask=0x4)
> [  403.217434] ata5.00: disabled
> It finally bombs... Same thing happens for all 4 ports...
> 
> While I do know my way around Linux, C and a debugger, I'm still pretty 
> blank when it comes ATA stuff... So any tips on where to go from here 
> would highly appreciated...

Well, the hardware programming guide for SX4 is available at
http://gkernel.sourceforge.net/specs/promise/

Can you post "lspci -vvv" entry for this device?  I'm betting that the 
SX4060 BIOS information is merely echoing what is in the PCI BARs for 
the PCI device.


> Some questions:
> 
> * I guess the soft lockups during DIMM init are suspicious... Is it
>   possible that the driver got the memory map wrong? The detected
>   64 MB is correct... The SX4060 BIOS lists this information:
> 
>     Installed Memory Address: FE800000 Size: 64M
>     Controller IRQ: 11
>     ATA IO Address: E000
>     XOR and HDMA IO Address: ED00
>     Sequence Control IO Address: EC00
>     Flash and SRAM Memory Address: FE900000
> 
>    But I don't know how to relate this to what's in the driver
>    source...

Anything is possible when two pieces of h/w are different :)

You could try commenting out the DIMM init stuff in the driver 
completely, on the hope that BIOS configured it properly.


> *  Is it a reasonable assumption that SX4060 support could be achieved
>    with fairly small changes to the sata_sx4 driver? (I realize that
>    knowing exactly /what/ changes are needed is an entirely different
>    matter...)

Dunno.  Your progress so far is a really good sign, and Promise 
certainly re-uses designs, so it is a reasonable theory, at least...


>  * Is there any point in contacting Promise about information on this
>    device? Or would I just be wasting my time?

I've been the one getting Promise to release docs, but it's a slow 
process that (literally) has taken years.

However, that said, they _do_ respond and _are_ interested.  You might 
try promise_linux@promise.com and point out that the existing SX4 docs 
are publicly released at http://gkernel.sourceforge.net/specs/promise/ 
with the blessing of Chi and Sam @ Promise.


>  * Is it a correct assumption that the sata_sx4 driver knows nothing the
>    various array configurations and that it will always just see 4
>    independent drives, regardless of what arrays have been configured in
>    the BIOS?

Correct.


>  * Also, is there any (easy) way to choke debug output from some ATA
>    devices, while keeping it for others? Apart from grep... :-)

Alas, not at present.

There is a half-finished infrastructure in libata, but messages need to 
be moved over to it (ap->msg_enable).


>  * Would there be any useful information to gather from running the card
>    with the Promise drivers under Windows or a 2.4 kernel?

<shrug>  Those are all useful tools for the toolbox, certainly.  Nothing 
pops into my head immediately, but those are definitely tools to keep 
around for this type of endeavor.


> I'd be happy to put in some time and effort in moving this forward but I 
> probably won't make much progress need some pointers from some ATA-gurus...

If you are looking for generic ATA programming information, 
http://ata.wiki.kernel.org/index.php/Developer_Resources tries to give 
the basics.  Even if there is a broken link, you can usually google for 
a filename to find a bazillion copies of the spec in question.

Regards,

	Jeff