[lkml]   [1998]   [Oct]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: IDE tape and SMP
On Fri, 23 Oct 1998, Andre M. Hedrick wrote:

> I have back checked the ide-driver parts that are pre-me and post-me,
> that is 2.1.121 and 2.1.122. In both cases using an offboard IDE
> controller (Promise Ultra/33) and calling tape commands via
> generic 'tar' commands "tar cvvf /dev/ht0".

In my case, it's under SMP mode, using the Supermicro P6DNF's onboard IDE.
I do get the same hang and DSC timeouts as well as complaints about IRQs
being shafted or something (I can dig out the logs from my home computer's
mail database if desired)

I tested this particular configuration under UP - backing up to tape works
and runs very quickly - a hour for 1.5GB worth of data is not too bad (If
I'd turned on bus-mastering it'd go a lot faster - but it was off by
default in BIOS - I just turned it on before I left for work and will
attempt to do another backup tonight). Unfortunately there is a nasty
problem with verifying the tapes - when it starts reading back the data
off the tape, the throughput drops to 6K/s (yes, 6KB a second!) I'm at a
loss as to why this happens since this never happened with my old i486
machine. It's a shame since it knocks out my machine for 5-6 hours instead
of 2 hours previously which was just right for watching a movie ;o)

> Since there are reports that UP ide-tape is functional and behaving (not
> verified personally), there may be bigger issues to follow-up.

This is correct - under UP it all works for backing up. Has anyone with
the same problems tried verifying their backups? (for bru, use -ivvvv,
this just verifies the data is good without needing to compare to the
files on the hard disks). I'd like to know if I am unique with the
apparent slowness of the data verification phase.

This give you any good data to work with, Andre? I'm open for patches to
test out ;o) (But let me back up first...!)

I only made the move to SMP thanks to my new computer with 2.1.125 so I
can't verify whether older kernels had this problem though.


/\_/\ Legalise cannabis now!
( o.o ) Grow some cannabis today!
> ^ < Peace, Love, Unity and Respect to all.

Check out
Linux lo-pc3035a 2.1.125 #6 Fri Oct 9 13:53:00 EDT 1998
One Intel Pentium 75+ processor, 66.36 total bogomips, 16M RAM
System library 5.4.44

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:45    [W:0.036 / U:5.496 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site