On Tue, 10 Jan 2006, Steve Finney wrote: > I am running into an odd intermittent symptom using the Samsung S3C2410 > ARM SOC with the 2.6.14.3 kernel (it seems to happen occasionally, but > less > often, with the 2.6.10 kernel). This chip has built-in OHCI, which > has virtually no documentation in the manual except for the addresses of > the OHCI controller registers. The gross symptom is that intermittently > (say, 50% of the time in the worst configuration of U-Boot and kernel > versions) none of the USB devices get enumerated (i.e., lsusb > only sees the root hub); this appears to be due to an -ETIMEDOUT error > trying to read a descriptor. A more detailed symptom (based on turning > on > both CONFIG_USB_DEBUG and OHCI_VERBOSE_DEBUG on ohci-hcd.c, and adding > some additional ohci_dbg() messages) is that (consistently) in the cases > that successfully enumerate devices, the controller is sending > SOF's (based on the SF bit in the InterruptStatus register) pretty > much as soon as OHCI_USB_OPER is written to HcControl in ohci_run(), > whereas in the unsuccessful cases, SF only starts appearing later on > (after some more register accesses). The former behavior appears to be > correct based on documentation of the HCFS field in the OHCI spec.
This is a naive question from someone who doesn't know anything about OHCI. Can you modify the routine that writes OHCI_USB_OPER to make it wait until SF is set? If you do, does that help? Or does it just hang? Alan Stern ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
