Eliot Moss wrote:
> On 2/3/2023 7:56 AM, Dan Williams wrote:
> > A memblock is an inclusive memory range. Bound the search by the last
> > address in the memory block.
> > 
> > Found by wondering why an offline 32-block (at 128MB == 4GB) range was
> > reported as 33 blocks with one online.
> > 
> > Signed-off-by: Dan Williams <[email protected]>
> > ---
> >   daxctl/lib/libdaxctl.c |    2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/daxctl/lib/libdaxctl.c b/daxctl/lib/libdaxctl.c
> > index 5703992f5b88..d990479d8585 100644
> > --- a/daxctl/lib/libdaxctl.c
> > +++ b/daxctl/lib/libdaxctl.c
> > @@ -1477,7 +1477,7 @@ static int memblock_in_dev(struct daxctl_memory *mem, 
> > const char *memblock)
> >             err(ctx, "%s: Unable to determine resource\n", devname);
> >             return -EACCES;
> >     }
> > -   dev_end = dev_start + daxctl_dev_get_size(dev);
> > +   dev_end = dev_start + daxctl_dev_get_size(dev) - 1;
> >   
> >     memblock_size = daxctl_memory_get_block_size(mem);
> >     if (!memblock_size) {
> 
> Might this address the bug I reported?

This one?

http://lore.kernel.org/r/[email protected]

I don't think so, that one seems to have something to do with the file
extent layout that causes fs/dax.c to fallback to 4K mappings.

Reply via email to