>-----Original Message----- >From: Mike Frysinger [mailto:[email protected]] >Sent: Tuesday, May 26, 2009 10:38 AM >To: Cai, Cliff >Cc: [email protected]; >[email protected] >Subject: Re: [Linux-kernel-commits] [6469] >trunk/drivers/usb/musb/blackfin.c: bug[#4120]Sometimes, the >DMA complete interrupt never happens, this will cause the >polling to dead loop > >On Mon, May 25, 2009 at 22:28, Cai, Cliff wrote: >> >> >>>-----Original Message----- >>>From: Mike Frysinger [mailto:[email protected]] >>>Sent: Tuesday, May 26, 2009 10:18 AM >>>To: [email protected] >>>Cc: [email protected] >>>Subject: Re: [Linux-kernel-commits] [6469] >>>trunk/drivers/usb/musb/blackfin.c: bug[#4120]Sometimes, the DMA >>>complete interrupt never happens, this will cause the >polling to dead >>>loop >>> >>>On Mon, May 25, 2009 at 05:16, <[email protected]> wrote: >>>> -#if defined(CONFIG_MUSB_DMA_POLL) >>>> +/* Sometimes,the DMA complete interrupt never happens, >>>> + * this will cause the polling to dead loop. >>>> + */ >>>> +#if 0 >>>> u8 epnum = hw_ep->epnum; >>>> u16 dma_reg = 0; >>> >>>uhh, this is already an ifdef for a reason -- so it can be >turned off. >>> why does it have to be replaced with an '#if 0' ? >> >> Just for reserving the code here,maybe this problem will be fixed in >> the future. > >is this an anomaly for only *reading* the fifo ? or did you >debug to this problem to this piece of code ? this option is >already Blackfin-specific and i imagine Bryan added it in the >first place because he wanted a kconfig to tweak quickly to >test for changes in silicon rather than mucking in the source code. >-mike
I don't know if the writing side also has this problem(no pressure test available),but reading side does have it. Cliff _______________________________________________ Linux-kernel-commits mailing list [email protected] https://blackfin.uclinux.org/mailman/listinfo/linux-kernel-commits
