On Tue, Feb 25, 2020 at 3:23 PM Dan Williams <[email protected]> wrote:
>
> On Wed, Feb 19, 2020 at 10:04 AM Jeff Moyer <[email protected]> wrote:
> >
> > Dan Williams <[email protected]> writes:
> >
> > >
> > > +NDCTL_EXPORT unsigned long ndctl_region_get_align(struct ndctl_region 
> > > *region)
> > > +{
> > > +     return region->align;
> > > +}
> > > +
> > > +NDCTL_EXPORT int ndctl_region_set_align(struct ndctl_region *region,
> > > +             unsigned long align)
> > > +{
> > > +     struct ndctl_ctx *ctx = ndctl_region_get_ctx(region);
> > > +     char *path = region->region_buf;
> > > +     int len = region->buf_len, rc;
> > > +     char buf[SYSFS_ATTR_SIZE];
> > > +
> > > +     if (snprintf(path, len, "%s/align", region->region_path) >= len) {
> > > +             err(ctx, "%s: buffer too small!\n",
> > > +                             ndctl_region_get_devname(region));
> > > +             return -ENXIO;
> > > +     }
> > > +
> > > +     sprintf(buf, "%#lx\n", align);
> > > +     rc = sysfs_write_attr(ctx, path, buf);
> > > +     if (rc < 0)
> > > +             return rc;
> > > +
> > > +     region->align = align;
> > > +     return 0;
> > > +}
> > > +
> >
> > Missing doctext.  Specifically, there should be a big, fat warning
> > against changing the region alignment.
>
> I don't mind adding one, but is this the right place to document an
> API warning? If the audience is future ndctl developers that should be
> warned to keep the status quo of not plumbing this capability into
> "create-namespace" that's one message. If it's to stop other libndctl
> application developers, they'll likely never see this source file.
>
> As it stands there is

Whoops, meant to delete this before sending...
_______________________________________________
Linux-nvdimm mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to