Hello, I have a problem with the last stable version of linux (2.6.15.x) : it hangs at boot while loading ehci_hcd. The cursor is always blinking, but everything is stopped. There was no problem with the kernel 2.6.14 and previous versions. Here are the two last lines that appear on the screen : ehci_hcd 0000:00:1d.7: EHCI Host Controller ehci_hcd 0000:00:1d.7: debug port 1 Of course, I use vanilla sources with no patch. My motherboard is a MSI 865PE Neo2-P, with an Intel 865PE chipset. I have the newest BIOS. There is something strange because I have exactly the same problem with FreeBSD 6 which also hangs while loading ehci... It seems that I'm not the only one who have this problem : http://lkml.org/lkml/2006/1/4/170 But it hasn't been resolved. I also tried with the 2.6.16-rc4 kernel and it's worse, it hangs before (I don't remember where exactly). Thanks. Fred (sorry for my english)
Where exactly does it hang in the 2.6.16-rc4 kernel? And if you enable CONFIG_USB_DEBUG, does it help show more information?
With CONFIG_USB_DEBUG, and when I add "debug" in grub : - With 2.6.15 : ehci_hcd 0000:00:1d.7: EHCI Host Controller ehci_hcd 0000:00:1d.7: reset hcs_params 0x104208 dbg=1 cc=4 pcc=2 ordered !ppc ports=8 ehci_hcd 0000:00:1d.7: reset hcc_params 6871 thresh 7 uframes 1024 64 bit addr ehci_hcd 0000:00:1d.7: debug port 1 ehci_hcd 0000:00:1d.7: capability 0001 at 68 - With 2.6.16-rc4 : 0000:00:1d.0: uhci_check_and_reset_hc: legsup = 0x0030 0000:00:1d.0: Performing full reset 0000:00:1d.1: uhci_check_and_reset_hc: legsup = 0x0030 0000:00:1d.1: Performing full reset 0000:00:1d.2: uhci_check_and_reset_hc: legsup = 0x0030 0000:00:1d.2: Performing full reset 0000:00:1d.3: uhci_check_and_reset_hc: legsup = 0x0030 0000:00:1d.3: Performing full reset
Check and see if a BIOS update is available for your machine. This type of problem is caused by a BIOS bug.
I have already done the latest BIOS update. There was no problem with Linux 2.6.14 and older versions.
Re Comment #2, you appear to be saying that UHCI didn't come up on 2.6.15 while EHCI did (partly), and on 2.6.16-rc4 it was EHCI that didn't while UHCI did. That doesn't seem right; I suspect you edited important information out of the boot logs.
Created attachment 7468 [details] experimental patch against 2.6.14-rc4 Try this against RC4. This should pretty much finish morphing the previous "early handoff" code for EHCI so it matches the previous EHCI driver code, with the only notable difference being the SMI now always being forcibly disabled. (That previous "early handoff" code was pretty buggy.)
This patch works ! Thanks !
The patch went into Linus' tree and will therefore be in 2.6.16-rc6.