Messages in this thread | | | Date | Sun, 15 Aug 1999 05:09:05 +1000 | From | Satanus Imperator <> | Subject | 2.2.11 -- AHA1542 as module seems broken? |
| |
I do apologise for posting to the Linux kernel list directly, without subscribing, but alas, we pay $0.20/MB, and I'm afraid we just can't accomodate the traffic such a list as this would generate. :((( I did check the post-August 9 linux-kernel archives, and found no mention of the following problem I'm going to detail.
I posted the following to the last named person in the aha1542.c driver (Mr. Fischer), but have received no word or reply.... and I thought I'd mention this to someone who might have ownership of this code before things move along too far to feature-freeze.
Please CC: me on any replies...
Thanks, =Rob=
PS: Please note that I cannot "live with" compiling in aha1542 support into the kernel (non-modularly)... and so I have not tried it.
------------------------------------
I have an ISA AHA-1542CF [it's properly setup, floppy is disabled, etc, etc] in an AMD K6-200 [IWILL w/on-motherboard SCSI AIC-7860].... with another PCI SCSI card [sym53c8xx] as well [lots of I/O channels, yes, but they are all setup with non-conflicting IRQs/DMAs/ioports].
[[ Kernel dudes: I really can't wait until I can fully utilise all three SCSI channels to their potential w/o essentially going all "serial". There are times I swear, during file-system intensive stuff on one channel, filesystems on other channels seem slower than they should be... but I need to do some synthetic benchmarks to measure this.]]
The Adaptec is configged as follows: (SCSI-ID 7) at IO:334, IRQ 9, DMA priority 5
Anyhow... I had recently gone from the rag-tag upgraded 2.0.37 kernel, to the 2.2.x series. My first version was 2.2.6, that was provided by Slackware 4.0. The 1542 worked fine with this system.
So, I downloaded 2.2.11, compiled it [I have IDE boot hard drives, and load ALL SCSI drivers as modules to control the "load order" [I don't want to have to worry about the scsi channel #s shifting around, messing up my /etc/fstab entries]. SCSI disk support I do build into the kernel, though (non-modularly).
So, the insmod aha1542 [module 2.2.11 version] fails [doesn't complete].... when I attempt to issue it again, I get a "resource or device busy" but nothing in /proc/scsi, and no notice from the cd-rom driver [which is attached to the aha1542].
So.... I took the aha1542.c and aha1542.h from the 2.2.7 [the latest source provided on Slack4.0], and punched it into the 2.2.11 driver directory, rebuilt that module, and installed it, and voila! It works just fine.
I have the "K6/Pentium" CPU option, as well as *NO* SMP enabled for my kernel build. I know things have been changing alot to improve SMP scalability... doing an object file size comparison shows about 280 bytes larger binary in the 2.2.11 driver.
I'm afraid I have no other details to go by.
Incidentally, I load the SCSI drivers in the following order:
aic7xxx sym53c8xx aha1542
Do you have an ideas or suggestions?
=Rob=
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |