Thread: [Ndiswrapper-general] Problems upgrading kernel
Status: Beta
Brought to you by:
pgiri
From: Bill W. <woe...@co...> - 2004-06-27 02:20:56
|
I just installed Fedora Core 2 and ndiswrapper. I was FLOORED how easy it was. It worked fine... until I ran "yum update", which grabbed the latest kernel, 2.6.6-1.435. Upon reboot, ndiswrapper didn't work (of course). So I recompiled it. Even then, it didn't work. I kept getting: ndiswrapper: Windows driver couldn't initialize the device mrv8k51.sys: probe of 0000:01:0a.0 failed with error -22 Lucky for me, Fedora keeps the old kernels around when it updates. I booted back to the old kernel (2.6.5-1.358) and everything worked fine. *shrug* So then, I grabbed ndiswrapper from CVS. It still didn't work (same errors). On a whim, I grabbed the vanilla kernel source for 2.6.7 and built my own kernel. Same problem. I finally decided to mail the list about this. I recompiled ndiswrapper with DEBUG defined. The output is enormous (~3000 lines), so I won't put it in this email. Instead, it's on my web page. Here's a direct link: http://members.cox.net/woessner/messages About my system: Home-built Athlon XP 3000 with 400 MHz FSB 512 MB RAM D-Link DWL-G510 (Marvell (Libertas) chipset) Brand new Fedora Core 2 installation For now, I'll stick with the 2.5.5-1.358 kernel. But I think this problem is bigger than just a weird Fedora kernel. Strangely enough, ndiswrapper worked with the kernel that shipped with Fedora Core 2, despite the fact that it was compiled with 4k stacks. Go figure. Any help you could offer would be greatly appreciated. And please feel free to contact me with any questions. Thanks in advance. --Bill |
From: Giridhar P. <gi...@lm...> - 2004-06-27 02:36:08
|
See the FAQ (about error -22). Something else (probably pcmcia) is using the same device? See "cat /proc/interrupts" to figure out what is going on. Of late, there seem to be lot of ACPI issues interfering with ndiswrapper. For example, in the log that you sent, I see the following messages: Jun 26 16:53:29 papageno kernel: ehci_hcd 0000:00:02.2: new USB bus registered, assigned bus number 1 Jun 26 16:53:29 papageno kernel: irq 12: nobody cared! (screaming interrupt?) Jun 26 16:53:29 papageno kernel: Stack pointer is garbage, not printing trace Jun 26 16:53:29 papageno kernel: handlers: Jun 26 16:53:29 papageno kernel: [<022169f4>] (usb_hcd_irq+0x0/0x4b) Jun 26 16:53:29 papageno kernel: Disabling IRQ #12 If ndiswrapper gets the same IRQ, then it obviously won't work. All this points to bigger issues with that kernel. -- Giri |
From: Bill W. <woe...@co...> - 2004-06-27 20:20:54
|
> Something else (probably pcmcia) is using the same device? See > "cat /proc/interrupts" to figure out what is going on. Of late, there > seem to be lot of ACPI issues interfering with ndiswrapper. I checked out the interrupt situation. This is a desktop, so I don't have a PCMCIA subsystem. Just to be safe, I stopped PCMCIA from starting at boot time. That didn't help, either. Here's the output from "cat /proc/interrupts" under 2.6.5-1.358 (the one that works): CPU0 0: 112820 XT-PIC timer 2: 0 XT-PIC cascade 5: 252 XT-PIC ohci_hcd 8: 1 XT-PIC rtc 9: 0 XT-PIC acpi 11: 6 XT-PIC ohci_hcd, NVidia nForce2 12: 907 XT-PIC ehci_hcd, ndiswrapper 14: 8211 XT-PIC ide0 15: 1814 XT-PIC ide1 NMI: 0 ERR: 0 Now the output under 2.6.6-1.435 (the one that doesn't work): CPU0 0: 161403 XT-PIC timer 2: 0 XT-PIC cascade 5: 619 XT-PIC ohci_hcd 8: 1 XT-PIC rtc 9: 0 XT-PIC acpi 11: 37 XT-PIC ohci_hcd, NVidia nForce2 12: 100000 XT-PIC ehci_hcd 14: 8993 XT-PIC ide0 15: 2360 XT-PIC ide1 NMI: 0 ERR: 0 Another interesting fact that may help: When I boot the 2.6.6 kernel, my USB laser printer is detected. If I boot back under the 2.6.5 kernel, the laser printer never gets detected. Thanks again for all your help. --Bill |
From: Giridhar P. <gi...@lm...> - 2004-06-27 20:44:07
|
I mentioned it, but you seemed to miss it. The problem is not ndiswrapper, but USB. As I said, USB disabled interrupt 12, so anything that uses that interrupt will not work. You can try another kernel. -- Giri |
From: Giridhar P. <gi...@lm...> - 2004-06-27 02:39:02
|
It turns out that ndiswrapper indeed gets IRQ 12 (which has been disabled due to USB problems), so that is at least one of the problems. Later in the log: Jun 26 16:56:22 papageno kernel: NdisMQueryAdapterResources handle: 103fa220. buf: 0f62fbd8, len: 344. IRQ:12 -- Giri |