[Announce] Emulex lpfcdriver v8.0.9 available

[Announce] Emulex lpfcdriver v8.0.9 available

[James.Smart]

Christoph,

We posted the 8.0.9 lpfc driver on SourceForge last week. 

Of the more notable changes that have occurred over the last few drops:
- Applied Christoph's patches for proc info, memory allocation routines, and bld_scsi_cmd dead code.
- Removed lun handling. Removed code that replicated midlayer functions, issued report luns, did partial VSA, had unnecessary data structures, etc.
- Reworked the clock/timer handling per Christoph's comments.
- Reworked driver parameters per Christoph's comments.
- Use the SCSI host lock instead of our own adapter lock
- Fixed the reset_bus and reset_device routines. Send "Reset LUN" for reset_device requests.
- More cleanup: condensing routines, removing casts and dead code, etc.

The ChangeLogs have more detail.

To date, the known todos are:
- Use mid-layer I/O timers
- No Peeking into timer structures
- Need to get approvals on all the clock changes, parameter changes, etc.
- Add calls to scan functions on new device detection, remove_device on device loss.
- Add interfaces to support managability.

Thanks

-- James S


The tarball is here:
http://prdownloads.sourceforge.net/lpfcxxxx/lpfcdriver-2.6-8.0.9.tar.gz?download

The full ChangeLog is on SourceForge:
http://sourceforge.net/forum/forum.php?forum_id=398596

Re: [Announce] Emulex lpfcdriver v8.0.9 available

[Christoph Hellwig]

On Mon, Aug 16, 2004 at 04:07:51PM -0400, James.Smart@Emulex.Com wrote:
> - Removed lun handling. Removed code that replicated midlayer functions, issued report luns, did partial VSA, had unnecessary data structures, etc.

this looks very good.

> - Reworked the clock/timer handling per Christoph's comments.

Nope.  Rezad the statements about timer_pending() and mod_timer() in
my mail again.

> - Reworked driver parameters per Christoph's comments.

No really.

> To date, the known todos are:
> - Use mid-layer I/O timers
> - No Peeking into timer structures
> - Need to get approvals on all the clock changes, parameter changes, etc.
> - Add calls to scan functions on new device detection, remove_device on device loss.
> - Add interfaces to support managability.


 - streamline I/O submission path
 - sort out memory allocation issues

Re: [Announce] Emulex lpfcdriver v8.0.9 available

[Christoph Hellwig]

On Mon, Aug 16, 2004 at 09:27:06PM +0100, Christoph Hellwig wrote:
> > To date, the known todos are:
> > - Use mid-layer I/O timers
> > - No Peeking into timer structures
> > - Need to get approvals on all the clock changes, parameter changes, etc.
> > - Add calls to scan functions on new device detection, remove_device on device loss.
> > - Add interfaces to support managability.
> 
> 
>  - streamline I/O submission path
>  - sort out memory allocation issues

- get rid of persistant lun binding support in the driver