kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by jalyv » Thu, 22 Jan 2004 04:16:03


CR 53c8xx fails with model 810 chip (motherboard an Alpha Avanti
machine,
an Alphaserver 400 4/233). I downloaded and built the 2.4.24 kernel
from kernel.org, the config file is a cut-down version from the
original hp config file (to remove all the useless modules). Of
course I used make newversion and make xconfig to do that.

Can anyone suggest a fix or change?


aboot: starting kernel vmlinuz-2.4.24jal with arguments root=/dev/sdf3
console=t
tyS0 ncr53c8xx=verb:2
Linux version 2.4.24jal (jal@alp) (gcc version 2.96 20000731 (Red Hat
Linux 7.2
2.96-112.7.2)) #6 Sun Jan 18 15:27:05 PST 2004
Booting on Avanti using machine vector Avanti from MILO
Major Options: LEGACY_START
Command line: root=/dev/sdf3 console=ttyS0 ncr53c8xx=verb:2
<snip>
RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
SCSI subsystem driver Revision: 1.00
ncr53c8xx: setup=disc:y,specf:3,tags:8,sync:10,burst:7,wide:y,diff:0,revprob:n,b
uschk:0x1
ncr53c8xx: setup=mpar:y,spar:y,fsn=n,verb:2,debug:0x0,led:n,settle:2,irqm:0x0,nv
ram:0x1,pcifix:0x0
ncr53c8xx: at PCI bus 0, device 6, function 0
ncr53c8xx: 53c810 detected
ncr53c810-0: rev 0x1 on pci bus 0 device 6 function 0 irq 11
ncr53c810-0: ID 7, Fast-10, Parity Checking
ncr53c810-0: initial SCNTL3/DMODE/DCNTL/CTEST3/4/5 = (hex)
00/00/00/00/00/00
ncr53c810-0: final SCNTL3/DMODE/DCNTL/CTEST3/4/5 = (hex)
03/c8/00/00/08/00
ncr53c810-0: resetting, command processing suspended for 2 seconds
ncr53c810-0: restart (scsi reset).
scsi0 : ncr53c8xx-3.4.3b-20010512
ncr53c810-0: command processing resumed
scsi : aborting command due to timeout : pid 0, scsi0, channel 0, id
0, lun 0 In
quiry 00 00 00 ff 00
ncr53c8xx_abort: pid=0 serial_number=1 serial_number_at_timeout=1
ncr53c810-0: abort ccb=fffffc0007f02000 (cancel)
SCSI host 0 abort (pid 0) timed out - resetting
SCSI bus is being reset for host 0 channel 0.
ncr53c8xx_reset: pid=0 reset_flags=2 serial_number=1
serial_number_at_timeout=1
ncr53c810-0: resetting, command processing suspended for 2 seconds
ncr53c810-0: command processing resumed
SCSI host 0 abort (pid 1) timed out - resetting
SCSI bus is being reset for host 0 channel 0.
ncr53c8xx_reset: pid=1 reset_flags=2 serial_number=2
serial_number_at_timeout=2
ncr53c810-0: resetting, command processing suspended for 2 seconds
ncr53c810-0: command processing resumed
SCSI host 0 abort (pid 2) timed out - resetting
SCSI bus is being reset for host 0 channel 0.
ncr53c8xx_reset: pid=2 reset_flags=2 serial_number=3
serial_number_at_timeout=3
ncr53c810-0: resetting, command processing suspended for 2 seconds
ncr53c810-0: command processing resumed


here is boot from HP standard Redhat 7.2 for Alpha (2.4.18-27.7.x.hp)
where
the scsi driver works:

<snip>
RAMDISK: Compressed image found at block 0
Freeing initrd memory: 670k freed
VFS: Mounted root (ext2 filesystem).
Red Hat nash version 3.2.6 starting
Loading sd_mod module
Loading ncr53c8xx module
ncr53c8xx: at PCI bus 0, device 6, function 0
ncr53c8xx: 53c810 detected
ncr53c810-0: rev 0x1 on pci bus 0 device 6 function 0 irq 11
ncr53c810-0: ID 7, Fast-10, Parity Checking
scsi0 : ncr53c8xx-3.4.3b-20010512
Vendor: DEC Model: RZ29B (C) DEC Rev: 0014
Type: Direct-Access ANSI SCSI revision: 02
Vendor: DEC Model: RZ29B (C) DEC Rev: 0016
Type: Direct-Access ANSI SCSI revision: 02
Vendor: DEC Mod
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by Michal Jae » Tue, 27 Jan 2004 05:01:09


Try instead sym53c8xx_2 driver. I have no way to test that but
this version, as opposed to an older sym53c8xx, is supposed to
work with 810 controllers.

It would be interesting to know what results you achieved.

Michal

 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by Graham Pay » Sun, 01 Feb 2004 09:25:48

might be wrong (I'm somewhat new to the Alpha), but I'm running the same
chip in my LX164 533, it's on a Intraserver card with eth0. I've not had any
problems with it at all. Running 2.4.24 with card in 32 bit PCI slot.

Here's my dmesg (exerpt):

Uniform CD-ROM driver Revision: 3.12
SCSI subsystem driver Revision: 1.00
sym53c8xx: at PCI bus 1, device 4, function 0
sym53c8xx: setting PCI_COMMAND_PARITY...(fix-up)
sym53c8xx: 53c875 detected with Symbios NVRAM
sym53c875-0: rev 0x4 on pci bus 1 device 4 function 0 irq 19
sym53c875-0: Symbios format NVRAM, ID 7, Fast-20, Parity Checking
scsi0 : sym53c8xx-1.7.3c-20010512
blk: queue fffffc000026cad0, no I/O memory limit
Vendor: TOSHIBA Model: CD-ROM XM-6401TA Rev: 1009
Type: CD-ROM ANSI SCSI revision: 02
blk: queue fffffc000026ccd0, no I/O memory limit
Vendor: IBM Model: DDRS-39130W Rev: S97B
Type: Direct-Access ANSI SCSI revision: 02
blk: queue fffffc000026ced0, no I/O memory limit
Vendor: IBM Model: DDRS-39130W Rev: S97B
Type: Direct-Access ANSI SCSI revision: 02
blk: queue fffffc000026d0d0, no I/O memory limit
Vendor: IBM Model: DDRS-39130W Rev: S97B
Type: Direct-Access ANSI SCSI revision: 02
blk: queue fffffc000026d2d0, no I/O memory limit
Vendor: IBM Model: DDRS-39130W Rev: S97B
Type: Direct-Access ANSI SCSI revision: 02
blk: queue fffffc000026d4d0, no I/O memory limit
Vendor: SONY Model: SDT-9000 Rev: 0400
Type: Sequential-Access ANSI SCSI revision: 02
blk: queue fffffc000026d6d0, no I/O memory limit
Vendor: IBM Model: DDRS-39130W Rev: S97B
Type: Direct-Access ANSI SCSI revision: 02
blk: queue fffffc000026d8d0, no I/O memory limit
Vendor: IBM Model: DDRS-39130W Rev: S97B
Type: Direct-Access ANSI SCSI revision: 02
blk: queue fffffc000026dad0, no I/O memory limit
ncr53c8xx: at PCI bus 1, device 4, function 0
ncr53c8xx: IO region 0x8000[0..127] is in use
scsi1 : SCSI host adapter emulation for IDE ATAPI devices
st: Version 20030406, bufsize 32768, max init. bufs 4, s/g segs 16
Attached scsi tape st0 at scsi0, channel 0, id 5, lun 0
Attached scsi disk sda at scsi0, channel 0, id 1, lun 0
Attached scsi disk sdb at scsi0, channel 0, id 2, lun 0
Attached scsi disk sdc at scsi0, channel 0, id 3, lun 0
Attached scsi disk sdd at scsi0, channel 0, id 4, lun 0
Attached scsi disk sde at scsi0, channel 0, id 6, lun 0
Attached scsi disk sdf at scsi0, channel 0, id 8, lun 0
sym53c875-0-<1,*>: FAST-20 WIDE SCSI 40.0 MB/s (50.0 ns, offset 15)
SCSI device sda: 17850000 512-byte hdwr sectors (9139 MB)
Partition check:
blah...

Hope this helps someone...

.-=gp=-.
"John" < XXXX@XXXXX.COM > wrote in message
news: XXXX@XXXXX.COM ...
setup=disc:y,specf:3,tags:8,sync:10,burst:7,wide:y,diff:0,revprob:n,b
setup=mpar:y,spar:y,fsn=n,verb:2,debug:0x0,led:n,settle:2,irqm:0x0,nv


 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by jalyv » Tue, 03 Feb 2004 13:04:51


I did try that. It doesn't work either. The symptoms change,
basically it goes into dead mode after reset, abort, timeout, etc...

I went to the trouble of diff'ing the source between 2.4.18 and the hp
version, and the hp-version and 2.4.24. There doesn't seem to be a
lot that changed, just revisions to allow lots more scsi devices, and
perhaps an error message or error situations that might be handled
differently. The hardware does work since 2.4.18-27.7.hp or whatever
it is called does in fact boot.

I did spend a lot of time reviewing past postings on alphalinux,
google, kernel.org, and so forth, back to 1999. There were 3 or 4
people that reported similar issues (reset loops or just eventual
hangs) but no resolutions.

Next choice: incremental source changes plus rebuilds....
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by Michal Jae » Wed, 04 Feb 2004 04:57:24


Just to be sure you can drop driver sources which did work for you into
2.4.24 (most likely this does not need any adjustments) and try that.
If this does not work then something else changed (generic SCSI layer?,
handling PCI?, ...) which affects you. In such case you should likely
post either to XXXX@XXXXX.COM , or
XXXX@XXXXX.COM mailing lists, or both. If reverting
driver changes does resolve the issue then you should post to lists
above anyway.

I assume that you excluded possible cable troubles and things like bad
termination or other hardware issues? If these are marginal then slight
change in timings may be a difference between "mostly working" and "not
working at all".

Michal
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by jalyv » Sat, 07 Feb 2004 15:20:55

> I assume that you excluded possible cable troubles and things like bad

yah, it's all pure Storageworks hardware so it should be fine, plus or
minus a cable.

looking over linux-scsi it seems there is a bunch of changes in scsi
mid-layer that may have affected things. hard to say. most of the
current work is on fibre channel, not legacy stuff like 2gb disks ;{)

rebuilding and rebooting is kinda boring, so i may try a gentoo or
debian distro just to see....

wish me luck


John
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by jalyv » Mon, 23 Feb 2004 09:23:57

> wish me luck

Well, it wasn't luck, it was a very useful hint from Adrian. Thanks!

The trick is: build a Generic kernel (CONFIG_ALPHA_GENERIC=y), do not
tailor it for Avanti. SCSI works fine then.

I'm going to post an email to kernel developers.


John
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by jukk » Tue, 24 Feb 2004 00:57:31

ohn wrote:


Nice to hear that you found a solution. However, I'm still having problems, not
sure if they're related though. I'm using 53C896 board on Miata, and it simply
doesn't work with 2.4.20 or later kernels. 2.4.19 and earlier are ok. I've tried
different drivers and options (also CONFIG_ALPHA_GENERIC=y) to no avail. BTW I
don't think this is the infamous Pyxis bug though, at least there is no data
corruption involved. Google found several hits with my error msg, so it
looks like I'm not alone with this. Any ideas?


This is from Debian testing kernel 2.6.2 (generic):

aboot> b 1/vmlinuz ro root=/dev/sda1 console=ttyS0
sym53c8xx=nvram:n,revprob:y,buschk:0x2,verb:2,irqm:1,led:0,de
bug:0x200
aboot: loading uncompressed vmlinuz...
aboot: loading compressed vmlinuz...
aboot: zero-filling 168664 bytes at 0xfffffc000073ca30
aboot: starting kernel vmlinuz with arguments ro root=/dev/sda1 console=ttyS0
sym53c8xx=nvram:n,revprob:y,buschk
:0x2,verb:2,irqm:1,led:0,debug:0x200
Linux version 2.6.2-generic (root@partner) (gcc version 3.3.3 20040125
(prerelease) (Debian)) #2 Sun Feb 22 11:5
3:27 EET 2004
Booting GENERIC on Miata using machine vector Miata from SRM
Major Options: LEGACY_START VERBOSE_MCHECK MAGIC_SYSRQ
Command line: ro root=/dev/sda1 console=ttyS0
sym53c8xx=nvram:n,revprob:y,buschk:0x2,verb:2,irqm:1,led:0,debug:0
x200
memcluster 0, usage 1, start 0, end 236
memcluster 1, usage 0, start 236, end 98303
memcluster 2, usage 1, start 98303, end 98304
freeing pages 236:384
freeing pages 949:98303
reserving pages 949:951
pci: cia revision 1 (pyxis)
On node 0 totalpages: 98303
DMA zone: 98303 pages, LIFO batch:8
Normal zone: 0 pages, LIFO batch:1
HighMem zone: 0 pages, LIFO batch:1
Building zonelist for node : 0
Kernel command line: ro root=/dev/sda1 console=ttyS0
sym53c8xx=nvram:n,revprob:y,buschk:0x2,verb:2,irqm:1,led:0,
debug:0x200
PID hash table entries: 4096 (order 12: 65536 bytes)
HWRPB cycle frequency bogus. Estimated 433162719 Hz
Using epoch = 2000
Console: colour VGA+ 80x25
Memory: 771648k/786424k available (2800k kernel code, 12536k reserved, 650k
data, 224k init)
Calibrating delay loop... 858.04 BogoMIPS
Dentry cache hash table entries: 131072 (order: 7, 1048576 bytes)
Inode-cache hash table entries: 65536 (order: 6, 524288 bytes)
Mount-cache hash table entries: 512 (order: 0, 8192 bytes)
POSIX conformance testing by UNIFIX
NET: Registered protocol family 16
EISA bus registered
pci: passed tb register update test
pci: passed sg loopback i/o read test
pci: passed pte write cache snoop test
pci: failed valid tag invalid pte reload test (mcheck; workaround available)
pci: passed pci machine check test
pci: tbia workaround enabled
pci: enabling save/restore of SRM state
PCI: Bus 1, bridge: 0000:00:14.0
IO window: disabled.
MEM window: disabled.
PREFETCH window: disabled.
PCI: Setting latency timer of device 0000:00:14.0 to 64
Linux Plug and Play Support v0.97 (c) Adam Belay
SCSI subsystem initialized
srm_env: version 0.0.5 loaded successfully
ikconfig 0.7 with /proc/config*
udf: registering filesystem
isapnp: Scanning for PnP cards...
isapnp: No Plug & Play device found
pty: 256 Unix98 ptys configured
lp: driver loaded but no devices found
rtc: SRM (post-2000) epoch (2000) detected
Real Time Clock Driver v1.12
[drm] Initialized tdfx 1.0.0 20010216 on minor 0
[drm] Initialized r128 2.5.0 20030725 on mi
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by Torsten Zi » Fri, 27 Feb 2004 17:13:00

Hi.

jukka < XXXX@XXXXX.COM > writes:



[deletia]

Well, did you or anone found a solution, why only the generic build
is working and not the specialized ones like e.g. Miata. I still run a
2.4.19, since this seems to be the last version which works in the
specialized build. I have a miata with a 53C875 chip. I tried the
NCR53c8xx, SYM53C8xx and SYM53C8xx2 driver, but all died with the
behavour described in the previous posts of this thread.

TIA

Torsten

--
s-mail: Neckarstr. 120 Torsten Zirzlaff
70190 Stuttgart
Germany e-mail : XXXX@XXXXX.COM
phone: +49 711 285 19 40 XXXX@XXXXX.COM
data: +49 711 286 40 57/9 XXXX@XXXXX.COM
mobil: +49 174 857 56 77

+-----------------------------------------------------------------------------+
| The shortest unit of time in the universe is the New York Second, defined |
| as the period of time between the traffic lights turning green and the cab |
| behind you honking. -- Terry Pratchett 'Lords and Ladies' |
+-----------------------------------------------------------------------------+
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by jukk » Sat, 28 Feb 2004 22:19:46


So do you mean that you can get 2.4.20-> kernels working as generic? How? Just
asking, because I can't get even generic kernels working. 2.4.19 or earlier are
ok though, regardless of build options and drivers. Maybe this should be posted
to linux-scsi or kernel mailinglists?
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by Jay Maynar » Sat, 28 Feb 2004 22:47:38


I've now successfully built and run both generic and system-specific (in my
case, Noritake) kernels from the Gentoo alpha-sources package, which is a
somewhat patched 2.4.21. I don't have anything higher than a 53C810, though,
so I can't speak to how well it might work with the newer chips.
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by Tom Gravga » Mon, 01 Mar 2004 06:04:37

On 26 Feb 2004 09:13:00 +0100, Torsten Zirzlaff



I haven't had any problems with 2.4.20+ kernels built for the miata
machinetype. I have an Asus SC875 SCSI controller in my PWS433a.

I'm currently running 2.4.24 and before that I had 2.4.21 running. All
kernels built from pristine sources taken from a kernel.org mirror.
I used gcc 2.96-112 from RH7.2 while binutils was taken from RH9
(src.rpm rebuild)

I've put my config up here:
http://www.yqcomputer.com/

-tgc
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by Torsten Zi » Tue, 02 Mar 2004 04:49:19

Hi.

Jay Maynard < XXXX@XXXXX.COM > writes:



Unfortunately, no. I now get the following error for each kernel >
2.4.20 I build for myself. I tried it with gcc-2.95 and gcc-3.3


I even tried the posted config here and it failed to start with the
follwing error:

sym.1.8.0 setting PCI_COMMAND_INVALIDATE
sym0:<875> rev 0x3 on pci bus 1 device 8 function 0 irq 36
sym0: No NVRAM, ID7, Fast-20 SE, parity checking
CACHETEST FAILED: DMA error (estat=0x81).sym0:CACHE INCORRECTLY CONFIGURED

Have anybody here an idea how to get a more modern kernel running on
a Miata with debian testing with a Symbios 53C875 ?

I do not know what might be wrong, since the system works with a
2.4.19 and I'm writting from it.

TIA

Torsten
--
s-mail: Neckarstr. 120 Torsten Zirzlaff
70190 Stuttgart
Germany e-mail : XXXX@XXXXX.COM
phone: +49 711 285 19 40 XXXX@XXXXX.COM
mobil: +49 174 857 56 77
data: +49 711 286 40 57/9

+-----------------------------------------------------------------------------+
| The shortest unit of time in the universe is the New York Second, defined |
| as the period of time between the traffic lights turning green and the cab |
| behind you honking. -- Terry Pratchett 'Lords and Ladies' |
+-----------------------------------------------------------------------------+
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by jukk » Wed, 03 Mar 2004 07:05:31


OK, I think I've found a solution. I think the problem is Debian kernel patches.
Everything seems to be OK with vanilla kernel from kernel.org (tested 2.4.25 and
2.6.3), but nothing above 2.4.19 works with Debian patched kernels. Thanks a lot
Tom, hope this helps others too.
 
 
 

kernel 2.4.24 seems to break NCR53C8XX scsi driver (scsi resets)

Post by jalyv » Fri, 05 Mar 2004 12:23:44

Torsten Zirzlaff < XXXX@XXXXX.COM > wrote


This seems to be a SYM error not NCR. You could try a newer kernel
because I seem to remember problems mentioned around 2.4.20 or 2.4.22
on Alpha. Try avoiding[1] the SYM driver and using SYM..._2, which
seems to be newer and better supported. If you read the driver code
carefully it may tell you if your card is even supposed to work.

The literal interpretation of the message is "PCI card not compatible
with your hardware", but this is clearly mistaken as 2.4.19, as you
say, works.

[1] I mean do not configure it as y or m, but enable only the
sym53c8xx_2 SCSI driver, as y or m.



--John