Re: [RFC PATCH v3 07/11] iomap: Add DSYNC support to RWF_WRITETHROUGH

[Ojaswin Mujoo]

On Mon, Aug 17, 2026 at 03:26:18PM +0200, Pankaj Raghav (Samsung) wrote:
> >  		return -EINVAL;
> >  
> > +	/*
> > +	 * If we realise that cache flush is necessary (eg FUA is not present
> > +	 * or we need metadata updates) then we turn off the optimization.
> > +	 */
> > +	if (wt_ctx->use_fua) {
> > +		if (iter->iomap.type != IOMAP_MAPPED ||
> > +		    (iter->iomap.flags &
> > +		     (IOMAP_F_NEW | IOMAP_F_SHARED | IOMAP_F_DIRTY)) ||
> 
> > +		    (bdev_write_cache(iter->iomap.bdev) &&
> > +		     !bdev_fua(iter->iomap.bdev)))
> Shouldn't this be similar to DIO code: !bdev_write_cache() || bdev_fua()
> ?
> 
> As long as bdev_fua() is supported, it does not matter if the
> bdev_write_cache is there or not right?

Hey Pankaj, sorry i somehow thought i replied to this already :/

So yeah we do keep the logic same as dio. In dio we enable FUA if:

    (!bdev_write_cache(iomap->bdev) || bdev_fua(iomap->bdev)))

and here we __disable__ if 

    (bdev_write_cache(iter->iomap.bdev) && !bdev_fua(iter->iomap.bdev)))

So its the exact inverse of the dio check.

Regards,
ojaswin

> 
> > +			wt_ctx->use_fua = false;
> > +	}
> > +
> 
> --
> Pankaj