On Sat, Jun 10, 2006 at 03:54:25AM +0200, Arjan van de Ven wrote: > On Fri, 2006-06-09 at 22:25 +0200, Guennadi Liakhovetski wrote: > > + > > + if (cmd->use_sg) { > > + struct scatterlist* sg = (struct scatterlist *)cmd->request_buffer; > > + size_t offset = 0, len = sizeof(struct ScsiInqData); > > + > > + local_irq_save(flags); > > + base = scsi_kmap_atomic_sg(sg, cmd->use_sg, &offset, &len); > > + ptr = (struct ScsiInqData *)(base + offset); > > + } else > > + ptr = (struct ScsiInqData *)(cmd->request_buffer); > > hmm I wonder if the non-use-sg case can ever happen? > (and if that irq_save isn't leaking here) It can still happen in when invoked through the EH code, but I plan to fix that for 2.6.18. - : 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