Hi David, I applied your most recent patch to the 2.4.21-pre5 Kernel and found that I can write as much as I like to the EHCI, but reading is still an issue (although I can read more before it hangs - something that was introduced with the usb-storage patch).
I was going through the Kernel logs and noticed something that is displayed just after ehci-hcd is loaded, regarding the VIA USB 2.0 controller and PCI: Mar 6 11:59:27 Webdesktop kernel: ehci-hcd 00:10.3: VIA Technologies, Inc. USB 2.0 Mar 6 11:59:27 Webdesktop kernel: ehci-hcd 00:10.3: irq 12, pci mem d28c4d00 Mar 6 11:59:27 Webdesktop kernel: usb.c: new USB bus registered, assigned bus number 4 Mar 6 11:59:27 Webdesktop kernel: ehci-hcd 00:10.3: ehci_start hcs_params 0x3206 dbg=0 cc=3 pcc=2 ordered !ppc ports=6 Mar 6 11:59:27 Webdesktop kernel: ehci-hcd 00:10.3: ehci_start hcc_params 6872 thresh 7 uframes 256/512/1024 Mar 6 11:59:27 Webdesktop kernel: ehci-hcd 00:10.3: capability 0001 at 68 Mar 6 11:59:27 Webdesktop kernel: ehci-hcd 00:10.3: reset command 080002 (park)=0 ithresh=8 period=1024 Reset HALT ***Mar 6 11:59:27 Webdesktop kernel: PCI: 00:10.3 PCI cache line size set incorrectly (32 bytes) by BIOS/FW. ***Mar 6 11:59:27 Webdesktop kernel: PCI: 00:10.3 cache line size too large - expecting 16. Mar 6 11:59:27 Webdesktop kernel: ehci-hcd 00:10.3: init command 010009 (park)=0 ithresh=1 period=256 RUN 00:10.3 is the VT8235's USB 2.0 controller. There were no options in the CMOS to configure the PCI cache line size, so I haven't been able to test other values. Could it be that this issue is causing problems with the EHCI (particularly reading data from the storage device)? > It's been quite a while since VIA would "lock up" hardware > in my testing. Lately it's not even hiccupping for me, won't > even wedge a single process. (Though more VT6202 testing would > surely be good). So I'm surprised you imply nothing changed on > the 2.5 kernels. The 2.5 Kernels aren't something I've been testing lately. I recall you mentioning that they are kinder with the VT8235's in general, so I'll take your advice and do some more testing on 2.5. > I've had it running with a bunch of "usbtest" cases for the > last day ... those put a mix of read and write loads (high > speed in this case). It ran several hours of disk testing > before that (including "mkfs -c"). Those things have been > intermittently troublesome on VIA. That's very good to hear, my experiences are similar with USB 2.0 hardware on the VT8235, but reading from the device hasn't improved anywhere close to that of writing to the device. I'll keep you updated. Thanks, Jonathan Thorpe ------------------------------------------------------- This SF.net email is sponsored by: Etnus, makers of TotalView, The debugger for complex code. Debugging C/C++ programs can leave you feeling lost and disoriented. TotalView can help you find your way. Available on major UNIX and Linux platforms. Try it free. www.etnus.com _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
