On Mon, 23 Sep 2024, Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx> wrote:
> On Mon, Sep 23, 2024 at 12:12:39PM +0300, Jani Nikula wrote:
>> On Fri, 20 Sep 2024, Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx> wrote:
>> > On Thu, Sep 12, 2024 at 03:15:52PM +0300, Jani Nikula wrote:
>> >> On Tue, 10 Sep 2024, Ville Syrjala <ville.syrjala@xxxxxxxxxxxxxxx> wrote:
>> >> > From: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx>
>> >> >
>> >> > Replace the three hand rolled "$VBT"s with a vbt_signature[]
>> >> > to avoid accidents.
>> >> >
>> >> > Signed-off-by: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx>
>> >> > ---
>> >> > drivers/gpu/drm/i915/display/intel_bios.c | 12 +++++++++---
>> >> > 1 file changed, 9 insertions(+), 3 deletions(-)
>> >> >
>> >> > diff --git a/drivers/gpu/drm/i915/display/intel_bios.c b/drivers/gpu/drm/i915/display/intel_bios.c
>> >> > index cbbda94c3dab..0b92b494117f 100644
>> >> > --- a/drivers/gpu/drm/i915/display/intel_bios.c
>> >> > +++ b/drivers/gpu/drm/i915/display/intel_bios.c
>> >> > @@ -2964,6 +2964,8 @@ static const struct bdb_header *get_bdb_header(const struct vbt_header *vbt)
>> >> > return _vbt + vbt->bdb_offset;
>> >> > }
>> >> >
>> >> > +static const char vbt_signature[4] = "$VBT";
>> >>
>> >> Always a bit scary to add strings without termination, but I guess it's
>> >> justified here.
>> >
>> > I guess we could make it look less like a string with eg.
>> > static const u8 vbt_signature[] = { '$', 'V', 'B', 'T' };
>> > ?
>>
>> Alternatively,
>>
>> static const char *vbt_signature = "$VBT";
>
> Or could just be something like:
>
> static const char vbt_signature[] = "$VBT";
> static const int vbt_signature_len = 4;
>
> BUILD_BUG_ON(vbt_signature_len != sizeof(vbt_signature) - 1);
> BUILD_BUG_ON(vbt_signature_len != sizeof(u32));
>
> and use vbt_signature_len instead sizeof(vbt_signature) everwhere.
Works for me, up to you.
BR,
Jani.
>
>>
>> >
>> >> Reviewed-by: Jani Nikula <jani.nikula@xxxxxxxxx>
>> >>
>> >>
>> >> > +
>> >> > /**
>> >> > * intel_bios_is_valid_vbt - does the given buffer contain a valid VBT
>> >> > * @display: display device
>> >> > @@ -2986,7 +2988,7 @@ bool intel_bios_is_valid_vbt(struct intel_display *display,
>> >> > return false;
>> >> > }
>> >> >
>> >> > - if (memcmp(vbt->signature, "$VBT", 4)) {
>> >> > + if (memcmp(vbt->signature, vbt_signature, sizeof(vbt_signature))) {
>>
>> But then this would need strlen()
>>
>> >> > drm_dbg_kms(display->drm, "VBT invalid signature\n");
>> >> > return false;
>> >> > }
>> >> > @@ -3082,9 +3084,11 @@ static struct vbt_header *spi_oprom_get_vbt(struct intel_display *display,
>> >> > oprom_offset = intel_uncore_read(&i915->uncore, OROM_OFFSET);
>> >> > oprom_offset &= OROM_OFFSET_MASK;
>> >> >
>> >> > + BUILD_BUG_ON(sizeof(vbt_signature) != sizeof(u32));
>>
>> And maybe this could be sizeof(vbt_signature) < sizeof(u32)
>>
>> All of the alternatives are a bit disappointing, so maybe let's just
>> roll with the patch you already have here?
>>
>> BR,
>> Jani.
>>
>> >> > +
>> >> > for (count = 0; count < oprom_size; count += 4) {
>> >> > data = intel_spi_read32(&i915->uncore, oprom_offset + count);
>> >> > - if (data == *((const u32 *)"$VBT")) {
>> >> > + if (data == *((const u32 *)vbt_signature)) {
>> >> > found = oprom_offset + count;
>> >> > break;
>> >> > }
>> >> > @@ -3144,9 +3148,11 @@ static struct vbt_header *oprom_get_vbt(struct intel_display *display,
>> >> > if (!oprom)
>> >> > return NULL;
>> >> >
>> >> > + BUILD_BUG_ON(sizeof(vbt_signature) != sizeof(u32));
>> >> > +
>> >> > /* Scour memory looking for the VBT signature. */
>> >> > for (i = 0; i + 4 < size; i += 4) {
>> >> > - if (ioread32(oprom + i) != *((const u32 *)"$VBT"))
>> >> > + if (ioread32(oprom + i) != *((const u32 *)vbt_signature))
>> >> > continue;
>> >> >
>> >> > p = oprom + i;
>> >>
>> >> --
>> >> Jani Nikula, Intel
>>
>> --
>> Jani Nikula, Intel
--
Jani Nikula, Intel
