Date: Mon, 15 May 2023 08:41:19 -0400 From: Oleg Lelchuk <oleglelchuk@gmail.com> To: Toomas Soome <tsoome@me.com> Cc: Warner Losh <imp@bsdimp.com>, Ed Maste <emaste@freebsd.org>, Emmanuel Vadot <manu@bidouilliste.com>, FreeBSD Current <freebsd-current@freebsd.org> Subject: Re: Why doesn't the EFI boot loader want to display the graphical orb logo in its boot menu on an Asus Prime 7590-P motherboard? Message-ID: <CA%2BGqWgs9-H%2BZpOTRe=kXd%2Bx1WRSfzkCDfKe2F-2ZV0vN=NE87Q@mail.gmail.com> In-Reply-To: <4AC4F6EE-CE18-4D67-A7F7-9328DAB3E1AB@me.com> References: <CA%2BGqWgudMq%2BeV5OJzuC_zR1Osdrak70PmOpFJqLMk3aDE0wy8w@mail.gmail.com> <3B658415-3AD0-4E8B-8CBE-F13FA70CBDC8@me.com> <20230512070557.859671981b7c616c0da7d666@bidouilliste.com> <CA%2BGqWguYpkm17xTQKN3qgyACpG7MkHPK3jJz=f9tx=zovvaNMA@mail.gmail.com> <4F0D21B1-58B6-413D-8499-11AF0E338C78@me.com> <CA%2BGqWguPT8uxtGhbWBw4kmR34jC7kG-DSMT0SuEm02smaigoEA@mail.gmail.com> <B8A678CB-8E2C-4A4C-852C-3E66C2C3B4B4@me.com> <CA%2BGqWgta5%2B-POUk88Z%2B-_FtcJouvLZpu5vYQ6kfX1DGwwjeCXw@mail.gmail.com> <CAPyFy2A4cW4SnFLEypM31JMDJT=A=vMh%2BW5g%2BO2gigf5OWDKrA@mail.gmail.com> <CA%2BGqWgsviBD_B_OQHERZd4J6S0chT-F=Rz0kExtUWq9t6QV_dA@mail.gmail.com> <CA%2BGqWgumZtyVJTbyfMf2bLypK1Ndw1L1AcQ7LTjnCtvRwAtJtA@mail.gmail.com> <CANCZdforzzd6kwQZ7RVctB5MG1=BU86ot7vBGydCek%2BUV44JOw@mail.gmail.com> <CA%2BGqWgsfVxvEduT_ncD%2BMKWDw9r8LqADF0wGsuZS8xt8V0p_-w@mail.gmail.com> <CA%2BGqWgucZuH5jc7xGUs_GPmebPzacHdH%2BWdh7qn2X8qnBm6uLQ@mail.gmail.com> <4AC4F6EE-CE18-4D67-A7F7-9328DAB3E1AB@me.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--0000000000003221d105fbbac64a
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
I got it.
On Mon, May 15, 2023, 8:32 AM Toomas Soome <tsoome@me.com> wrote:
>
>
> On 15. May 2023, at 15:22, Oleg Lelchuk <oleglelchuk@gmail.com> wrote:
>
> Adding screen.font=3D"16=C3=9732" to loader.conf fixed that tiny issue me=
ntioned
> in the previous email message... I find it a bit surprising that I only h=
ad
> to make one tiny change to the source code of stand to make the graphical
> logo appear, to start playing with the EFI resolution, and etc.
>
>
> The font size/resolution is difficult topic. The implementation itself ca=
n
> choose =E2=80=9Cgood enough=E2=80=9D variant and then some people are hap=
py and some people
> are unhappy.
>
> The current loader UI is built on terminal dimensions (which depend on
> glyph size and resolution), and there the traditional assumption is that =
we
> have 80x24 terminal. With different fonts and depending on how much scree=
n
> space we want to leave unused, we can get different dimensions for termin=
al.
>
> And since there is quite a variation of displays, the challenge is to get
> decent enough visual on most commonly used displays - so there can be
> pressure to use fixed resolution etc. And this is also the reason, why yo=
u
> see very simple boot screens with something like spinning wheel on some
> other systems.
>
> rgds,
> toomas
>
>
> On Sun, May 14, 2023, 8:58 AM Oleg Lelchuk <oleglelchuk@gmail.com> wrote:
>
>> Okay, so I edited /usr/src/stand/efi/loader/main.c , and I replaced
>> ConOut with ConIn in this line: rv =3D efi_global_getenv("ConIn", buf, &=
sz);
>> . Now I am able to see the beautiful graphical logo in the efi boot menu=
!
>> But why are the boot menu and the logo shown in the top left corner of m=
y
>> computer screen? My monitor is 1080p and the setting
>> efi_max_resolution=3D1080p in loader.conf only affects what happens afte=
r the
>> kernel starts booting up, but it doesn't affect what happens before it: =
the
>> boot menu and the logo remain in the top left corner of the screen. Why =
is
>> this the case? You can see the photo in the provided attachment... And
>> thank you, guys, for your work!
>>
>> On Sat, May 13, 2023 at 9:35=E2=80=AFAM Warner Losh <imp@bsdimp.com> wro=
te:
>>
>>>
>>>
>>> On Sat, May 13, 2023, 6:26 AM Oleg Lelchuk <oleglelchuk@gmail.com>
>>> wrote:
>>>
>>>> I've been reading the documentation for loader.efi and it says this:
>>>> "If there is no ConOut variable, both serial and video are attempted.
>>>> loader.efi uses the "efi" console for the video (which may or may
>>>> not
>>>> work) and "comconsole" for the serial on COM1 at the default baud
>>>> rate.
>>>> The kernel will use a dual console, with the video console primar=
y
>>>> if a
>>>> UEFI graphics device is detected, or the serial console as primar=
y
>>>> if
>>>> not."
>>>> I find this language confusing because I don't know what is meant by "=
a
>>>> UEFI graphics device". In my situation, is my Intel Integrated Graphic=
s
>>>> card an UEFI graphics device? Does it mean that once i915kms is loaded=
, I
>>>> no longer deal with UEFI graphics? I think lots of people whose native
>>>> language is English will find the documentation describing loader.efi
>>>> confusing. The documentation page also mentions this: "BUGS
>>>> Systems that do not have a ConOut variable set are not conformant
>>>> with
>>>> the standard, and likely have unexpected results." But I think yo=
u
>>>> guys already implied that the UEFI specification doesn't mandate havin=
g
>>>> such a variable.
>>>>
>>>
>>> That's unclear. The standard refers to it many times. Earlier versions
>>> especially. It doesn't say it's optional, unlike some other variables. =
Yet
>>> later versions don't say it's mandatory. I've yet to own or use a syst=
em
>>> without it... such systems exist but they are quite new...
>>>
>>> Warner
>>>
>>> On Fri, May 12, 2023 at 7:55=E2=80=AFPM Oleg Lelchuk <oleglelchuk@gmail=
.com>
>>>> wrote:
>>>>
>>>>> I got it. Thanks.
>>>>>
>>>>> On Fri, May 12, 2023 at 7:45=E2=80=AFPM Ed Maste <emaste@freebsd.org>=
wrote:
>>>>>
>>>>>> On Fri, 12 May 2023 at 09:26, Oleg Lelchuk <oleglelchuk@gmail.com>
>>>>>> wrote:
>>>>>> >
>>>>>> > I don't want to go through the hassle of filling a bug with my
>>>>>> vendor. I will just wait for you, guys, to update the stand implemen=
tation.
>>>>>> Thank you for explaining to me what causes this issue.
>>>>>>
>>>>>> This issue is tracked in PR 265980 if you want to follow it.
>>>>>> https://bugs.freebsd.org/265980
>>>>>>
>>>>>
>
--0000000000003221d105fbbac64a
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
<div dir=3D"auto">I got it.</div><br><div class=3D"gmail_quote"><div dir=3D=
"ltr" class=3D"gmail_attr">On Mon, May 15, 2023, 8:32 AM Toomas Soome <<=
a href=3D"mailto:tsoome@me.com">tsoome@me.com</a>> wrote:<br></div><bloc=
kquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #cc=
c solid;padding-left:1ex"><div style=3D"line-break:after-white-space"><br><=
div><br><blockquote type=3D"cite"><div>On 15. May 2023, at 15:22, Oleg Lelc=
huk <<a href=3D"mailto:oleglelchuk@gmail.com" target=3D"_blank" rel=3D"n=
oreferrer">oleglelchuk@gmail.com</a>> wrote:</div><br><div><div dir=3D"a=
uto">Adding screen.font=3D"16=C3=9732" to loader.conf fixed that =
tiny issue mentioned in the previous email message... I find it a bit surpr=
ising that I only had to make one tiny change to the source code of stand t=
o make the graphical logo appear, to start playing with the EFI resolution,=
and etc.</div></div></blockquote><div><br></div><div>The font size/resolut=
ion is difficult topic. The implementation itself can choose =E2=80=9Cgood =
enough=E2=80=9D variant and then some people are happy and some people are =
unhappy.</div><div><br></div><div>The current loader UI is built on termina=
l dimensions (which depend on glyph size and resolution), and there the tra=
ditional assumption is that we have 80x24 terminal. With different fonts an=
d depending on how much screen space we want to leave unused, we can get di=
fferent dimensions for terminal.</div><div><br></div><div>And since there i=
s quite a variation of displays, the challenge is to get decent enough visu=
al on most commonly used displays - so there can be pressure to use fixed r=
esolution etc. And this is also the reason, why you see very simple boot sc=
reens with something like spinning wheel on some other systems.</div><div><=
br></div><div>rgds,</div><div>toomas</div><br><blockquote type=3D"cite"><di=
v><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">On S=
un, May 14, 2023, 8:58 AM Oleg Lelchuk <<a href=3D"mailto:oleglelchuk@gm=
ail.com" target=3D"_blank" rel=3D"noreferrer">oleglelchuk@gmail.com</a>>=
wrote:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8=
ex;border-left:1px #ccc solid;padding-left:1ex"><div dir=3D"ltr">Okay, so I=
edited /usr/src/stand/efi/loader/main.c , and I replaced ConOut with ConIn=
in this line:=C2=A0rv =3D efi_global_getenv("ConIn", buf, &s=
z); . Now I am able to see the beautiful graphical logo in the efi boot men=
u! But why are the boot menu and the logo shown in the top left corner of m=
y computer screen? My monitor is 1080p and the setting efi_max_resolution=
=3D1080p in loader.conf only affects what happens after the kernel starts b=
ooting up, but it doesn't affect what happens before it: the boot menu =
and the logo remain in the top left corner of the screen. Why is this the c=
ase? You can see the photo in the provided attachment... And thank you, guy=
s, for your work!</div><br><div class=3D"gmail_quote"><div dir=3D"ltr" clas=
s=3D"gmail_attr">On Sat, May 13, 2023 at 9:35=E2=80=AFAM Warner Losh <<a=
href=3D"mailto:imp@bsdimp.com" rel=3D"noreferrer noreferrer" target=3D"_bl=
ank">imp@bsdimp.com</a>> wrote:<br></div><blockquote class=3D"gmail_quot=
e" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204)=
;padding-left:1ex"><div dir=3D"auto"><div><br><br><div class=3D"gmail_quote=
"><div dir=3D"ltr" class=3D"gmail_attr">On Sat, May 13, 2023, 6:26 AM Oleg =
Lelchuk <<a href=3D"mailto:oleglelchuk@gmail.com" rel=3D"noreferrer nore=
ferrer" target=3D"_blank">oleglelchuk@gmail.com</a>> wrote:<br></div><bl=
ockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-lef=
t:1px solid rgb(204,204,204);padding-left:1ex"><div dir=3D"ltr">I've be=
en reading the documentation for loader.efi and it says this: "If ther=
e is no ConOut variable, both serial and video are attempted.<br>=C2=A0 =C2=
=A0 =C2=A0loader.efi uses the "efi" console for the video (which =
may or may not<br>=C2=A0 =C2=A0 =C2=A0work) and "comconsole" for =
the serial on COM1 at the default baud rate.<br>=C2=A0 =C2=A0 =C2=A0The ker=
nel will use a dual console, with the video console primary if a<br>=C2=A0 =
=C2=A0 =C2=A0UEFI graphics device is detected, or the serial console as pri=
mary if<br>=C2=A0 =C2=A0 =C2=A0not."<div>I find this language confusin=
g because I don't know what is meant by "a UEFI graphics device&qu=
ot;. In my situation, is my Intel Integrated Graphics card an UEFI graphics=
device? Does it mean that once i915kms is loaded, I no longer deal with UE=
FI graphics? I think lots of people whose native language is English will f=
ind the documentation describing loader.efi confusing. The documentation pa=
ge also mentions this: "BUGS</div>=C2=A0 =C2=A0 =C2=A0Systems that do =
not have a ConOut variable set are not conformant with<br>=C2=A0 =C2=A0 =C2=
=A0the standard, and likely have unexpected results." But I think you =
guys already implied that the UEFI specification doesn't mandate having=
such a variable.</div></blockquote></div></div><div dir=3D"auto"><br></div=
><div dir=3D"auto">That's unclear. The standard refers to it many times=
. Earlier versions especially. It doesn't say it's optional, unlike=
some other variables. Yet later versions don't say it's mandatory.=
=C2=A0 I've yet to own or use a system without it... such systems exist=
but they are quite new...</div><div dir=3D"auto"><br></div><div dir=3D"aut=
o">Warner</div><div dir=3D"auto"><br></div><div dir=3D"auto"><div class=3D"=
gmail_quote"><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px =
0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div class=
=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">On Fri, May 12, 2023=
at 7:55=E2=80=AFPM Oleg Lelchuk <<a href=3D"mailto:oleglelchuk@gmail.co=
m" rel=3D"noreferrer noreferrer noreferrer" target=3D"_blank">oleglelchuk@g=
mail.com</a>> wrote:<br></div><blockquote class=3D"gmail_quote" style=3D=
"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-le=
ft:1ex"><div dir=3D"ltr">I got it. Thanks.</div><br><div class=3D"gmail_quo=
te"><div dir=3D"ltr" class=3D"gmail_attr">On Fri, May 12, 2023 at 7:45=E2=
=80=AFPM Ed Maste <<a href=3D"mailto:emaste@freebsd.org" rel=3D"noreferr=
er noreferrer noreferrer" target=3D"_blank">emaste@freebsd.org</a>> wrot=
e:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0=
.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On Fri, 12 Ma=
y 2023 at 09:26, Oleg Lelchuk <<a href=3D"mailto:oleglelchuk@gmail.com" =
rel=3D"noreferrer noreferrer noreferrer" target=3D"_blank">oleglelchuk@gmai=
l.com</a>> wrote:<br>
><br>
> I don't want to go through the hassle of filling a bug with my ven=
dor. I will just wait for you, guys, to update the stand implementation. Th=
ank you for explaining to me what causes this issue.<br>
<br>
This issue is tracked in PR 265980 if you want to follow it.<br>
<a href=3D"https://bugs.freebsd.org/265980" rel=3D"noreferrer noreferrer no=
referrer noreferrer" target=3D"_blank">https://bugs.freebsd.org/265980</a><=
br>
</blockquote></div>
</blockquote></div>
</blockquote></div></div></div>
</blockquote></div>
</blockquote></div>
</div></blockquote></div><br></div></blockquote></div>
--0000000000003221d105fbbac64a--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CA%2BGqWgs9-H%2BZpOTRe=kXd%2Bx1WRSfzkCDfKe2F-2ZV0vN=NE87Q>