Hi, Actually, what is the status of ur problem now ?? Have u solved it ?? Karthik bala guru
-----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Anil Yadav Sent: Sunday, June 05, 2005 2:37 PM To: David Brownell; [email protected]; [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: [Linux-usb-users] Re:Re:Problem with EHCI Sorry for using different e-mail id, I tried to send the reply but my mail is getting bounced from that account. Thanks David, I tried the audio class driver of kernel 2.6.9 and found the host controller sending ISO tokens at 125 us. I used same application to get the audio sample in user space but surprisingly it is sending all zeros.I did check the usb_audio_read() of both kernel versions, there's not much difference in the reading mechanism. My application is based on OSS api's. On kernel 2.4 for full speed there wasn't any problem as such.I was able to capture and playback. But this time I am only getting zeros and that too at a damn slow rate. I don't think i am able to do the read. Even i tried alsa libs for capture but it didn't work for me. Then I modified my video capture driver for kernel 2.6.9 and I have been able to get the video continuously with the same URB which I mentioned in my last query in kernel 2.4. The difference being 2.6 doesn't hang.I capture the video for around 30-35 minutes. The only problem with my implementation is I didn't find any difference between the frame rates I got with both the kernels. Its same for both as the analyzer output shows. I guess the device doesn't fetch the data at high speed usb's expectation.. But I am able to get much higher capture rate on MS windows. Thanks in Advance Regards Anil On Sunday 29 May 2005 6:03 am, Anil Yadav wrote: > I tried almost all possible combinations to frame urb. I checked with > kernels from 2.4.20 to 2.4.30. I'd not recommend a 2.4 kernel for your application. Try the latest 2.6 code, possibly with the patch from: Also, make sure your ISO code works with for example OHCI on 2.6; you may have problems in your ISO logic, apart from any potential EHCI issues. (Differences at high speed will include data rate, polling interval -- units in uframes, not frames -- and possibly per-period data size of up to 3 KB.) And the behavior for ISO changed in 2.6; a pointless "automagic resubmit" mechanism is gone. (Which was never implemented for EHCI in 2.4 fwiw ... the 2.6 approach works just fine on 2.4 though, with all HCDs.) > Is there any driver available in the kernel ( Isochronous) for high > speed device. By the way my kernel version is 2.4.20-8. Definitely do NOT use EHCI in any version before 2.4.22 ... there are some 2.4 versions with ISO support that may work for high speed ISO, but I don't know which ones they'd be. Certainly 2.4.20 is strongly discouraged. As I said, 2.6.recent ... plus maybe the patch above. Once it's working on 2.6, the iso code may work on newish 2.4 kernels. - dave ------------------------------------------------------- This SF.Net email is sponsored by: NEC IT Guy Games. How far can you shotput a projector? How fast can you ride your desk chair down the office luge track? If you want to score the big prize, get to know the little guy. Play to win an NEC 61" plasma display: http://www.necitguy.com/?r _______________________________________________ [email protected] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-users ------------------------------------------------------- This SF.Net email is sponsored by: NEC IT Guy Games. How far can you shotput a projector? How fast can you ride your desk chair down the office luge track? If you want to score the big prize, get to know the little guy. Play to win an NEC 61" plasma display: http://www.necitguy.com/?r=20 _______________________________________________ [email protected] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-users
