On Mon, May 8, 2017 at 2:17 PM, Verma, Vishal L
On Mon, 2017-05-08 at 10:00 -0700, Dan Williams wrote:
> On Wed, Apr 26, 2017 at 3:43 PM, Vishal Verma <vishal.l.verma(a)intel.c
> om> wrote:
> > If we had badblocks/poison in the metadata area of a BTT,
> > recreating the
> > BTT would not clear the poison in all cases, notably the flog area.
> > This
> > is because rw_bytes will only clear errors if the request being
> > sent
> > down is 512B aligned and sized.
> > Make sure that when writing the map and info blocks, the rw_bytes
> > being
> > sent are of the correct size/alignment. For the flog, instead of
> > doing
> > the smaller log_entry writes only, first do a 'wipe' of the entire
> > area
> > by writing zeroes in large enough chunks so that errors get
> > cleared.
> These eventually nsio_rw_bytes() while the namespace is claimed by a
> btt instance, so this collides with the hack to disable error
> for btt. If we want to fix this up for 4.12 I think we need to pass a
> 'flags' parameter to the ->rw_bytes() routine to indicate that we are
> in atomic context (NVDIMM_IO_ATOMIC), rather than assuming that all
> BTT I/O is atomic. Care to code that up? I think we can include it in
> a pull request before the merge window closes.
Ah good point. You mean a flag to say that we're in process context
right?We want to skip the clearing in atomic context..
Either way, a flag to indicate atomic vs process so that we can decide
whether to skip vs allow error clearing.