Messages in this thread | | | Date | Fri, 23 Nov 2001 09:48:05 +0100 | From | Vojtech Pavlik <> | Subject | Re: slowdown on Via ide chipsets with 2.4.15? |
| |
On Thu, Nov 22, 2001 at 02:20:31PM -0500, safemode wrote: > something i just took notice while fixing my computers. > hdparm -t to test the speed of drives, i have in the past had about 30MB/s > across all my drives. > Now, my primary disk, which is the only one on it's channel, gets 19MB/s > (udma2 speeds). It's an udma4 drive and i've seen it get 30+ on average in > earlier kernels. My primary master on the promise card gets 30+ still, but > my slave on the same channel gets 19MB/s even though the master is not being > used. I know there is to be expected some performance drop, but that much is > a little disconcerting. I have my atapi devices on the secondary channel of > my via controller (motherboard) and the second channel on the promise card > will be used for my new drive. > The harddrives are configured exactly the way they used to be and the kernel > is compiled with the exact same options for ide/dma and such. The > changefile doesn't look like the via ide drivers were messed with. The only > other difference is that all my drives are now ext3. > > Controller (motherboard) VP_IDE: VIA vt82c686a (rev 22) IDE UDMA66 > controller on pci00:07.1 > Controller (card) PDC20262: (U)DMA Burst Bit ENABLED Primary PCI Mode > Secondary PCI Mode. > > All drives are set to UDMA4 (ATA66) hdparm isn't used, simple check shows > that everything is set the way it should be default. > > Also, the kernel displays UDMA(66) wouldn't it be ATA66. UDMA comes in > 1,2,4, and 5. that's like, ata16, ata33, ata66 and ata100.
Send me your /proc/ide/via and hdparm -i /dev/hd*, and I'll see if I see any problems there ...
-- Vojtech Pavlik SuSE Labs - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |