Douglas Gilbert wrote:
> > Well, the error report could of course be added to sanei_scsi_req_enter,
> > but a backend should check for this error in sanei_scsi_req_wait too: If
> > you queue two or more commands and the low level driver does not support
> > command queueing, sanei_scsi_req_enter puts the second and following
> > commands into its internal queue, so it cannot return an error from the
> > SG driver.
>
> The reserved buffer is only good for one SCSI command
> at a time (per file descriptor) and only if it is big
> enough to hold the size of transfer being requested.
>
> Thus when commands are queued there is a risk that
> subsequent commands will fail with ENOMEM. This "out of
> memory" situation is more likely to occur with ISA
> SCSI adapters which can only use the bottom 16 MB of
> memory for DMA. This is not a hard error and should
> not be treated as such [IMO]. The reserve buffer
> more or less guarantees SANE can queue 1 command.
> If memory is not available to queue 2 or more commands
> then wait ....
Doug,
thanks for helping my memory :) I knew there were even more reasons to
check ENOMEM late, but I could not remember them while I wrote my mail.
Abel
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
This archive was generated by hypermail 2b29 : Mon Jul 16 2001 - 03:04:42 PDT