Page MenuHomeFreeBSD
Differential D52402

stand: Remove Spleen 32x64 fonts from INDEX
ClosedPublic
Actions

Authored by jrm on Sep 5 2025, 7:40 PM.
Tags
None
Referenced Files
Unknown Object (File)
Fri, Oct 10, 12:37 PM
Unknown Object (File)
Fri, Oct 10, 5:48 AM
Unknown Object (File)
Fri, Oct 10, 5:48 AM
Unknown Object (File)
Fri, Oct 10, 5:48 AM
Unknown Object (File)
Fri, Oct 10, 12:12 AM
Unknown Object (File)
Wed, Oct 1, 3:18 PM
Unknown Object (File)
Fri, Sep 26, 11:14 PM
Unknown Object (File)
Wed, Sep 24, 2:17 AM
View All 24 Files
Subscribers
bz
cperciva
dab
kevans
mark_freebsdfoundation.org

Details

Reviewers
emaste
imp
mhorne
tsoome
ziaee
manu
Group Reviewers
releng
Commits
rGf834384fdad9: stand: Remove Spleen 32x64 fonts from INDEX
rG5e82eeccd252: stand: Remove Spleen 32x64 fonts from INDEX
Summary

The Spleen 32x64 fonts cause some systems such as the AMD Framework 16
to boot extremely slowly. Adding screen.font="16x32" to
/boot/loader.conf is a partial workaround, but text displayed before the
boot menu still takes ~30 seconds to render on that laptop. With this
change, the same text renders immediately.

By leaving the font files in place but removing them from the INDEX,
users who want to use them can still load them manually. Once pending
work is committed to allow normal booting with these fonts, they can be
added back to the INDEX.

PR: 289007

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Skipped
Unit
Tests Skipped
Build Status
Buildable 66854
Build 63737: arc lint + arc unit

Event Timeline

jrm created this revision.Sep 5 2025, 7:40 PM
Herald added a reviewer: manu. · View Herald TranscriptSep 5 2025, 7:40 PM
Herald added a subscriber: dab. · View Herald Transcript
Harbormaster completed remote builds in B66852: Diff 161617.Sep 5 2025, 7:40 PM
jrm requested review of this revision.Sep 5 2025, 7:40 PM
jrm edited the summary of this revision. (Show Details)Sep 5 2025, 7:45 PM
tsoome added a comment.Sep 5 2025, 7:46 PM

one option is to leave font file in place but remove from INDEX. This way it is still possible to load it manually for those who want to use it, but it will not picked automatically.

jrm added a comment.Sep 5 2025, 7:48 PM

Ok. I'll do that instead.

jrm retitled this revision from Revert "stand: Enable Spleen 32x64 font in the bootloader" to stand: Remove Spleen 32x64 fonts from INDEX.Sep 5 2025, 8:03 PM
jrm updated this revision to Diff 161619.Sep 5 2025, 8:04 PM

Follow tsoome's suggestion.

Harbormaster completed remote builds in B66854: Diff 161619.Sep 5 2025, 8:04 PM
jrm edited the summary of this revision. (Show Details)Sep 5 2025, 8:12 PM
tsoome accepted this revision.Sep 5 2025, 8:14 PM

Thanks.

This revision is now accepted and ready to land.Sep 5 2025, 8:14 PM
ziaee accepted this revision.Sep 5 2025, 8:18 PM

Everyone has to loose features because something doesn't work on one specific model of laptop? Since when do we do that? Shouldn't the people using that specific model of laptop fix their issue?

Is forcing everyone else to have the loader crammed into a corner by removing the font the logic wants to select reasonable ?

tsoome added a comment.Sep 5 2025, 9:05 PM

The problem is, we do not really know exact blast zone. While preparing the possible work around, the font file is still there and can be scripted to be used with loadfont command.

jrm added a comment.EditedSep 5 2025, 9:06 PM
In D52402#1196630, @ziaee wrote:

Everyone has to loose features because something doesn't work on one specific model of laptop? Since when do we do that? Shouldn't the people using that specific model of laptop fix their issue?

Is forcing everyone else to have the loader crammed into a corner by removing the font the logic wants to select reasonable ?

According to reports I've heard, it's not just one model. From my conversation with @tsoome, excerpts of which are in the bug report, it's more of limitations with our implementations that he plans to fix. In the meantime, can you add screen.font=32x64 to /boot/loader.conf?

jrm added a comment.EditedSep 5 2025, 9:08 PM

Oh, sorry. One can't add screen.font=32x64 to /boot/loader.conf to get the Spleen fonts? I really can't tell much of a difference between the fonts on my system.

kevans added a subscriber: kevans.EditedSep 6 2025, 2:37 AM
In D52402#1196665, @jrm wrote:

Oh, sorry. One can't add screen.font=32x64 to /boot/loader.conf to get the Spleen fonts? I really can't tell much of a difference between the fonts on my system.

no, you'd probably want a loader.conf like:

exec="loadfont /boot/fonts/32x64.fnt.gz"
screen.font="32x64"

maybe

jrm added a comment.Sep 6 2025, 3:38 AM

Committed in 5e82eeccd252e96370e8b73603b609622497000c. (Forgot to put the darn review URL in the commit log.)

jrm closed this revision.Sep 6 2025, 3:39 AM
jrm added a commit: rG5e82eeccd252: stand: Remove Spleen 32x64 fonts from INDEX.
mark_freebsdfoundation.org added a subscriber: mark_freebsdfoundation.org.Sep 9 2025, 6:29 PM

This has fixed the painfully slow boot on my Framework 13 AMD too. Thanks @jrm !

jrm added a commit: rGf834384fdad9: stand: Remove Spleen 32x64 fonts from INDEX.Sep 11 2025, 2:08 AM
mark_freebsdfoundation.org added a comment.Sep 15 2025, 10:58 AM

Hmm, I've just installed 15-ALPHA2 - and the installer correctly didn't use these slow, large, fonts on my Framework 13 AMD test machine. But once installed the system is back to using these infuriatingly slow fonts.

jrm added a subscriber: cperciva.Sep 15 2025, 1:26 PM
In D52402#1199726, @mark_freebsdfoundation.org wrote:

Hmm, I've just installed 15-ALPHA2 - and the installer correctly didn't use these slow, large, fonts on my Framework 13 AMD test machine. But once installed the system is back to using these infuriatingly slow fonts.

Looking at the stable/15 commit log, f834384fdad9 landed before 99cc54090e1c. @cperciva, was ALPHA2 built from 99cc54090e1c? Do you know why f834384fdad9 (seemingly) wasn't included?

cperciva added a comment.Sep 15 2025, 4:22 PM

@jrm ALPHA2 was built from da62a299de31 and has this change.

Mark, did you install from disc1 or from the bootonly image? What does uname say? The bootonly image downloads pkgbase bits from pkg.freebsd.org, which was a bit behind last time I checked. Putting the pkgbase bits in a separate repository so the installer can actually get the right version is coming soon.

mark_freebsdfoundation.org added a comment.Thu, Sep 18, 6:15 PM

Sorry @cperciva, I've been a bit slow to see this update. I did the install from the 'disc1' image. I've since done a pkg upgrade, so I'm guessing the second question might not be answerable?

cperciva added a comment.Thu, Sep 18, 6:19 PM
In D52402#1201553, @mark_freebsdfoundation.org wrote:

Sorry @cperciva, I've been a bit slow to see this update. I did the install from the 'disc1' image. I've since done a pkg upgrade, so I'm guessing the second question might not be answerable?

Ok, well, I'm not sure exactly what happened then, but can you test with ALPHA3 over the weekend and confirm that the problem is gone there?

bz added a subscriber: bz.Thu, Sep 18, 6:57 PM
In D52402#1201557, @cperciva wrote:
In D52402#1201553, @mark_freebsdfoundation.org wrote:

Sorry @cperciva, I've been a bit slow to see this update. I did the install from the 'disc1' image. I've since done a pkg upgrade, so I'm guessing the second question might not be answerable?

Ok, well, I'm not sure exactly what happened then, but can you test with ALPHA3 over the weekend and confirm that the problem is gone there?

I left a comment on the bug; for me on main with a fw.16 I am still seeing slow scrolling after the first page fills. I am not sure if that is the very same problem that we are trying to fix here:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=289007#c10

mark_freebsdfoundation.org added a comment.Fri, Sep 19, 8:52 AM
In D52402#1201557, @cperciva wrote:
In D52402#1201553, @mark_freebsdfoundation.org wrote:

Sorry @cperciva, I've been a bit slow to see this update. I did the install from the 'disc1' image. I've since done a pkg upgrade, so I'm guessing the second question might not be answerable?

Ok, well, I'm not sure exactly what happened then, but can you test with ALPHA3 over the weekend and confirm that the problem is gone there?

Will do @cperciva 👍

mark_freebsdfoundation.org added a comment.Mon, Sep 22, 4:06 PM

@cperciva Just tested install of ALPHA3 on the Framework 13, and all is good 👍

Revision Contents
Changeset List

PathSize
stand/
fonts/
3 lines

Diff 161619

View Options

stand/fonts/INDEX.fonts

Loading...