On Mon, Feb 24, 2025 at 11:05:24AM +0100, Lorenzo Bianconi wrote: > > On Sat, Feb 22, 2025 at 11:43:45AM +0100, Lorenzo Bianconi wrote: > > [...] > > > > + > > > + entry = resource_list_first_type(&host->windows, IORESOURCE_MEM); > > > + if (!entry) > > > + return -EINVAL; > > > > -ENODEV or -ENOMEM > > ack, I will fix it in v4 > > > > > > + > > > + addr = entry->res->start - entry->offset; > > > + err = regmap_write(pbus_regmap, args[0], lower_32_bits(addr)); > > > + if (err) > > > > MMIO write is not supposed to fail. > > ack, I will fix it in v4 > > > > > > + return err; > > > + > > > + size = lower_32_bits(resource_size(entry->res)); > > > + mask = size ? GENMASK(31, __fls(size)) : 0; > > > > Size of MEM region could be 0? > > I added this check since we consider just lower_32_bits(). > Do you think we should remove it? > Yes please. > > > > > + err = regmap_write(pbus_regmap, args[1], mask); > > > + if (err) > > > > MMIO write is not supposed to fail. > > ack, I will fix it in v4 > > BTW I will remove your Reviwed-by tag since the patch has changed > with respect to the one you added it. Please let me know if you > prepfer to keep it. > Sure, fine with me. - Mani -- மணிவண்ணன் சதாசிவம்
