Hello, I would like to know if someone can confirm/deny if usb ohci Linux driver is multicore or fast CPU safe. I have just bought my first multicore CPU (Phenom 9550) and new mainboard for it. Because of this big hw change and troubles with usb ohci I do not know if this is: -mainboard problem, -fast CPU problem, -multicore problem (this is my first multicore CPU, before I used only one single core CPU). PROBLEM: USB OHCI controller on Nvidia MCP78S (Geforce 8200) randomly hangs during big transfer on fast 12Mbit usb device (usb ADSL modem, usb hard drive, usb irda dongle, pendrive, printer). Problem discovered on Asrock K10N78FullHD-hSLI R3.0 mainboard. All bios releases checked, including current one, no IRQ sharing happens on Linux. After hang communication with usb ohci controller/driver is lost till reboot. Pluging in usb hard drive can cause usb hang on plugin moment (device is detected, partition table is read but device files in /dev are not created or sometimes created but hangs on file copy). I do not know if this bug happens because of new mainboard or because this is my first multicore CPU I have. Slow 1Mbit USB OHCI devices are not affected by this bug (keyboards, trackballs, mice). USB EHCI works perfect (and all other devices). Windows XP SP3 is not affected by this bug so this is not hardware issue (using reference drivers provided with OS, Nvidia driver was NOT installed for testing). All my USB devices works perfect with previous mainboard: Asus A8N- VM CSM (Nvidia MCP51 (Geforce6150/nForce430)) and the older one: Abit NF7-S v2.0 (nForce2/MCP-T chipset). I always used single core CPU then. Workaround: noapic or acpi=noirq make this hang much less frequent, then usb ohci hangs randomly on only huge transfers: copying 1GiB file to/from usb drive or downloading CentOs DVD iso (via torrent usb hang is almost instant, via ftp it hangs after about 300MiB downloaded). Without noapic or acpi=noirq usb ohci hangs after downloading about 50% of e2fsprogs-1.41.7.tar.gz (4,3MiB size) Here is my bug report with dmesg,interrupts,acpi... dumps: http://bugzilla.kernel.org/show_bug.cgi?id=13405 Here is my topic on nvnews.net: http://www.nvnews.net/vbulletin/showthread.php?t=135022 and topic on kernel-devel mailing list: http://lkml.org/lkml/2009/6/27/99 have a nice day, Zbigniew Luszpinski -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
