[lkml]   [2001]   [Nov]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH] driver: ide-floppy.c kernel >=2.4.7

I've just got around to testing this. I didn't have problems with the
current version of ide-floppy, but I thought I'd give it a go anyway. I'm
sorry to say that the performance as standard is atrocious. I was
previously using 2.4.12-ac3 with the preempt patch, for these tests I
tried 2.4.13-ac6. (All tests at the console, no other conscious use of the

hdparm -Tt /dev/hdd1 cache disk
2.4.12-ac3-preempt 52.67Mb/s 1.06Mb/s
2.4.13-ac6+your patch 52.89Mb/S 480.43kB/s
(above are worst of three, but there isn't wide variation)

time mke2fs /dev/hdd1 real user sys
2.4.12-ac3-preempt 0m4.252s 0m0.010s 0m0.090s
2.4.13-ac6+your patch 0m8.349s 0m0.000s 0m0.070s

Bonnie -s 85 -d /zip
---Sequential Output------- --Sequential Input--- Rnd Seek
Char Block Rewrite Char Block /sec
2.4.12-ac3 1246K/s 1504K/s 433K/s 702K/s 958K/s 30.0
2.4.13-ac6+ 575K/s 699K/s 211K/s 347K/s 447K/s 20.9

This is on a K6/500, and with your patch the disk feels _so_ slow. I
accept it can be tweaked through the /proc interface, but if this is the
default behaviour, it's dreadful. I do wonder if you've got a flakey
drive ? Sorry.


On Tue, 30 Oct 2001, Skip Gaede wrote:

> This patch fixes a lost interrupt problem with the Iomega ATAPI Zip 100 drive
> on an Asus A7V133 board (uses South Bridge VIA VT82C686 chip). The problem
> occurs when trying to format the drive using mke2fs /dev/hdx1. The patch
> introduces an adjustable delay between the time the drive asserts DRQ and
> deasserts BSY after issuing the packet command and before transferring the 12
> byte packet.. With delays of 3-5 ticks, the filesystem creation occurs
> without retries/resets. The delay can be adjusted through the proc interface
> by adjusting the value assigned to the parameter ticks. (Without the patch, I
> experienced 111 lost interrupts, resulting in an elapsed time of over 2 hours
> to format the drive. With the patch, mke2fs completed in <15 seconds.)
> I am looking for testing, by others who have the internal Zip drive and may
> have experienced the same problem, as well as comments.
> Thanks,
> --Skip

If a six turned out to be nine, I don't mind.

Home page :

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

 \ /
  Last update: 2005-03-22 13:12    [W:0.073 / U:0.080 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site