Re: [PATCH v5 3/4] zram: validate parameters in each backend's setup_params

On Tue, Aug 4, 2026 at 1:48=E2=80=AFPM Sergey Senozhatsky
<[email protected]> wrote:
>
> On (26/08/03 22:12), Haoqin Huang wrote:
> [..]
> >  static int setup_params_842(struct zcomp_params *params)
> >  {
> > +     if (params->dict_sz) {
> > +             pr_err("842: dictionary is not supported\n");
> > +             return -EOPNOTSUPP;
> > +     }
> > +     if (params->level !=3D ZCOMP_PARAM_NOT_SET) {
> > +             pr_err("842: compression level is not supported\n");
> > +             return -EOPNOTSUPP;
> > +     }
> >       return 0;
> >  }
> [..]
> >  static int deflate_setup_params(struct zcomp_params *params)
> >  {
> > +     if (params->dict_sz) {
> > +             pr_err("deflate: dictionary is not supported\n");
> > +             return -EOPNOTSUPP;
> > +     }
> > +
> >       if (params->level =3D=3D ZCOMP_PARAM_NOT_SET)
> >               params->level =3D Z_DEFAULT_COMPRESSION;
> > +     else if (params->level < Z_DEFAULT_COMPRESSION ||
> > +              params->level > Z_BEST_COMPRESSION) {
> > +             pr_err("deflate: invalid compression level %d\n", params-=
>level);
> > +             return -EINVAL;
> > +     }
>
> This is purely optional, if you add per-backend pr_fmt() with backend
> name e.g. "zstd:","lz4:" and so on (in a separate patch) then you don't
> need to explicitly prefix every pr_err().

I considered adding pr_fmt, but the existing winbits message already
has "deflate" inline:

    pr_err("invalid deflate winbits: %d\n", wb);

which would become "deflate: invalid deflate winbits:" and look redundant.
I can add pr_fmt in a separate patch and clean that up to just
"invalid winbits %d" at the same time though, if you prefer.