[syslinux] ivy bridge and custom font

[Gene Cumm]

On Fri, Oct 10, 2014 at 7:18 AM, Ady <ady-sf at hotmail.com> wrote:
>
>> On Oct 10, 2014 4:00 AM, "Ady" <ady-sf at hotmail.com> wrote:
>> >
>> >
>> > > Hello,
>> > >
>> > > I've encountered a strange problem with loading a font file.
>> > > The environment:
>> > > motherboard GA-Z68MA-D2H-B3, processor i7-3770, onboard graphics,
>> > > extlinux 5.01.
>> > >
>> > > After loading any font I get a mess on a screen. The only rule I could
>> > > discover is that: characters with odd code are not displayed at all,
>> > > and for characters with even code there is taken a bitmap from position
>> > >   ord(character) / 2.
>> > >
>> > > However, there is neither problem with the builtin font nor console
>> fonts
>> > > loaded undre system control (Linux).
>> > >
>> > >
>> > > Could you help me to fix the problem?
>> > >
>> > > P.O.
>> > >
>> >
>> > This one is tricky.
>> >
>> > There are revisions of such mainboard with BIOS firmware, and others
>> > with UEFI firmware. Updating to the "latest" firmware is frequently
>> > suggested, but in this case the procedure might not be
>> > straightforward, nor the selection of which version of firmware is
>> > the most appropriate. Add beta version(s) of firmware updates...
>> >
>> > Booting in UEFI mode could also have some screen-resolution issues
>> > that might affect what you see on screen.
>> >
>> > Anyway, I would suggest either testing with Syslinux (extlinux) 4.xx
>> > or with 6.03. Forget about 5.xx.
>> >
>> > Please remember that official distribution archives for version 6.xx
>> > include the binary files under respective "bios/", "efi32/" and
>> > "efi64/" directories.
>>
>> The better question is do these fonts work at all with Syslinux 5.01, is it
>> an interaction with this system or perhaps a regression in 5.01?
>>
>> Did you try these same fonts with Syslinux 5.01 on another system?
>>
>> Did you try these fonts on this system with Syslinux 4.04/4.05/4.07/3.86?
>>
>> On another system?
>>
>> --Gene
>
> FWIW, 5.00 introduced a regression regarding the .psf font loader,
> which was later fixed in 5.10+ (so 5.01 still has this problem).
>
> Considering that 6.03 has so many improvements over 5.xx, IMHO using
> 5.xx is not worth the time (unless there is some specific test in
> mind).
>
> These are the reasons for me to suggest the font test in 4.xx and/or
> 6.03, skipping altogether the test in 5.xx.

Thanks.  I didn't recall the regression.

-- 
-Gene