Re: [ata] 0568e61225: stress-ng.copy-file.ops_per_sec -15.0% regression

[John Garry]

On 09/08/2022 15:57, Damien Le Moal wrote:
> > > As far as I can see, this patch should not make a difference unless the
> > > ATA shost driver is setting the max_sectors value unnecessarily low.
> > For __ATA_BASE_SHT, we don't set max_sectors. As such, we default
> > shost->max_sectors = SCSI_DEFAULT_MAX_SECTORS (=1024) in
> > scsi_host_alloc(). I assume no shost dma mapping limit applied.
> >
> > Then - for example - we could select dev->max_sectors =
> > ATA_MAX_SECTORS_LBA48 (=65535) in ata_dev_configure().
> >
> > So with commit 0568e6122574 we would have final max sectors = 1024, as
> > opposed to 65535 previously. I guess that the problem is something like
> > this.
> >
> > If so, it seems that we would need to apply the shost dma mapping limit
> > separately in ata_scsi_dev_config() and not use shost->max_sectors.
> OK. Will have a look at that.

We may need to introduce something like shost->max_hw_sectors, which is 
set according to sht max sectors and dma mapping limits. That could be 
also used in USB scsiglue slave_configure()

Or else set max_sectors value for __ATA_BASE_SHT, but I don't know a 
sane value there considering ATA_MAX_SECTORS_LBA48 gives max_sectors of 
65535.

Damien, please let me know if you need help now. I am just waiting for 
you to test to prove this theory about dev->max_sectors being capped. I 
don't have an AHCI setup readily-available for testing - just SAS cards 
or QEMU.

Thanks,
John