Good to know someone shares the problem. I posted on pretty much all the mailng list I know, without any "me too" effect. I have not fixed this issue yet. I know it is related to USB host driver (ehci-omap.c) because USB mass storage - panda being the host, usb stick being the device - also has transfer issue (Write). An interrupt is not fired properly in TX, RX is fine. I will resume my investigation not before the end of october.
Post in here if you make any progress in the meantime. Thanks, Emeric On Friday, September 28, 2012 4:19:22 AM UTC-4, Sagar Shinde wrote: > > Hi, > We are facing the same issue. > Did you get any solution for this? > > Thanks in advance. > BR > Sagar > sangi.hs_gmail.com > > On Thursday, July 12, 2012 10:26:32 PM UTC+3, m-ric wrote: >> >> Hi, >> >> I run AOSP-4.0.4 on a pandaboard rev A3. I got linux kernel from >> remotes/android-omap/android-omap-panda-3.0 branch, 52f4764 commit >> "Revert gpu: pvr: Revert to 1.8@550175". >> >> With this software: >> 1. Ethernet bandwidth is very low, barely usable for file transfers. >> TX gets stuck, I measured 20KBps. RX is quite slow too. >> 2. USB storage on USB host port suffers same limitations. I measured >> ~100KBps. >> >> Problem is known though, fedora- and ubuntu guys saw it: >> http://article.gmane.org/gmane.linux.redhat.fedora.arm/1266 >> https://bugs.launchpad.net/linux-linaro/+bug/709245 >> >> My kernel embeds Ming Lei's barrier patch in ehci-q.c: >> usb: ehci: make HC see up-to-date qh/qtd descriptor ASAP >> and Will Deacon's ARM errata CONFIG_PL310_ERRATA_769419: >> ARM: 7161/1: errata: no automatic store buffer drain >> I checked that asm on this latter commit is fine (dsb instr.). >> >> Still above bugs are still present. For fedora and ubuntu guys though, >> above fixes have solved the problem. >> I am reluctant to change my (entire) kernel because everything else >> that I need works fine. I am in a maturation phase, not in a >> development phase anymore. >> >> Does anyone experience that on pandaboard and aosp? >> Did you find any workaround? >> >> I have been on this bug for 4 weeks now. I am kind of desperate right >> now. > > -- unsubscribe: android-kernel+unsubscr...@googlegroups.com website: http://groups.google.com/group/android-kernel