Mike, I applied "patch 1" posted by Boaz Harrosh to my SLES 10 kernel. So far IO's are running for 45 minutes which is the longest in my numerous re-tries. We were not aware that "scsi_execute_async" was temperory workaround. Where is the depricated warning that you mentioned about ? I tried to look in scsi_lib.c and scsi_device.h Regards, Sudhir -----Original Message----- From: Mike Christie [mailto:michaelc@xxxxxxxxxxx] Sent: Friday, March 02, 2007 4:00 PM To: Dachepalli, Sudhir Cc: Benny Halevy; Jens Axboe; Boaz Harrosh; linux-scsi@xxxxxxxxxxxxxxx; James Bottomley Subject: Re: Possible bug in scsi_lib.c:scsi_req_map_sg() Mike Christie wrote: > Dachepalli, Sudhir wrote: >> scsi_req_map_sg::i=2,len=1024,data_len=3072,off=2048,PAGE_SIZE=4096,b >> yte >> s=1024,nr_vecs=0, nr_pages=0 > > >> if (bio_add_pc_page(q, bio, page, bytes, off) != >> bytes) { >> printk("scsi_req_map_sg:: calling >> bio_put \n"); >> >> printk("scsi_req_map_sg::i=%d,len=%d,data_len=%d,off=%d,PAGE_SIZE=%ld >> ,by >> tes=%d,nr_vecs=%d, nr_pages=%d\n", >> >> i,len,data_len,off,PAGE_SIZE,bytes,nr_vecs,nr_pages); >> if( bio->bi_io_vec == NULL ) > > > I think Boaz's first patch in this thread that counts the offsets > correctly should be merged. Just one clarification. When I wrote scsi_execute_async, it was meant as a temp hack so we would kill scsi_request and fix scatterlists for drivers like iscsi_tcp and ib_iser, but in the original patches and the patches I am sending now I modify the blk helpers and have sg use them directly. scsi_execute_async was meant to be temporary only supported what sg and st and other mainline drivers were sending at the time, so it did not support something like: sg[0].offset 0; sg[0].length = 4096; sg[1].offset = 1024; sg[1].offset = 3072; because sg and st can only have a offset for the first sg element (offsets for the sg element is supported). If we are going to support whatever scsi_do_req supported the we should merge Boaz's patch. If scsi_execute_async is going to be limited to what is in mainline until I can kill it, then we may not want to merge Boaz's patch and just have people convert the code to use blk_get_request, blk_rq_map_kern or blk_rq_map_user and blk_execute_rq_nowait. - To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
