Ok - I'll look at this path further... -- james
Ma, Steve wrote: > > With the latest patch, and I set > > job->reply->reply_payload_rcv_len = len; > > It crashed at the BUG_ON in fc_bsg_jobdone. > > -Steve > > > > ------------------------------------------------------------------------ > > *From:* James Smart [mailto:[email protected]] > *Sent:* Monday, April 20, 2009 12:20 PM > *To:* Ma, Steve > *Cc:* [email protected] > *Subject:* Re: [Open-FCoE] [RFC PATCH] [open-fcoe] Add FC pass-through > support > > > > > > Ma, Steve wrote: > >>> + pp, len); >>> + job->reply->reply_payload_rcv_len = job->reply_payload.payload_len; >>> >>> >> Why is this not: "job->reply->reply_payload_rcv_len = len;" ? >> > > I have found that in case the user-space code specified a bigger response > buffer size than len. i.e. when job->reply_payload.payload_len > len, it will > consider the reply data is not fulfilled, and it hangs. The way I do is to > allow receiving less data than expected. > > If this is true, it's a bug. Given the other statements in your reply, > you may need to test with the latest patch. See: > http://marc.info/?l=linux-scsi&m=123808882309133&w=2 > <http://marc.info/?l=linux-scsi&m=123808882309133&w=2> > There was one patch on top of it > (http://marc.info/?l=linux-scsi&m=123980043221985&w=2 > <http://marc.info/?l=linux-scsi&m=123980043221985&w=2>) > > -- james > _______________________________________________ devel mailing list [email protected] http://www.open-fcoe.org/mailman/listinfo/devel
