Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.bugs.dist > #1176304 > unrolled thread

Bug#1056933: raspi-firmware: CMA=0 not working as intended

Back to article view | Back to linux.debian.bugs.dist

Contents

  Bug#1056933: raspi-firmware: CMA=0 not working as intended Thorsten Glaser <tg@debian.org> - 2023-11-27 00:20 +0100
    Bug#1056933: [Pkg-raspi-maintainers] Bug#1056933: raspi-firmware: CMA=0 not working as intended Cyril Brulebois <cyril@debamax.com> - 2023-11-27 00:50 +0100
      Bug#1056933: [Pkg-raspi-maintainers] Bug#1056933: raspi-firmware: CMA=0 not working as intended Thorsten Glaser <tg@debian.org> - 2023-11-27 01:30 +0100
        Bug#1056933: [Pkg-raspi-maintainers] Bug#1056933: raspi-firmware: CMA=0 not working as intended Cyril Brulebois <cyril@debamax.com> - 2023-11-27 02:00 +0100
          Bug#1056933: [Pkg-raspi-maintainers] Bug#1056933: raspi-firmware: CMA=0 not working as intended Thorsten Glaser <tg@debian.org> - 2023-11-27 02:40 +0100

#1176304 — Bug#1056933: raspi-firmware: CMA=0 not working as intended

Package: raspi-firmware
Version: 1.20230405+ds-2

When I set CMA=0 the cma=… argument in cmdline.txt is omitted. This
leads to:

[    0.000000] Memory: 413440K/507904K available (8192K kernel code, 990K rwdata
, 2024K rodata, 1024K init, 252K bss, 28928K reserved, 65536K cma-reserved)


However, when I manually add cma=0 to cmdline.txt I get this instead:

[    0.000000] Memory: 479104K/507904K available (8192K kernel code, 990K rwdata
, 2024K rodata, 1024K init, 252K bss, 28800K reserved, 0K cma-reserved)


This is Debian bullseye/armel (with raspi-firmware from sid)
on an RPi B+ (rev. 1.2).

bye,
//mirabilos

[toc] | [next] | [standalone]

#1176306 — Bug#1056933: [Pkg-raspi-maintainers] Bug#1056933: raspi-firmware: CMA=0 not working as intended

[Multipart message — attachments visible in raw view] — view raw

Thorsten Glaser <tg@debian.org> (2023-11-26):
> Package: raspi-firmware
> Version: 1.20230405+ds-2
> 
> When I set CMA=0 the cma=… argument in cmdline.txt is omitted.

This is consistent with the documentation in that file.

> However, when I manually add cma=0 to cmdline.txt I get this instead:
> 
> [    0.000000] Memory: 479104K/507904K available (8192K kernel code, 990K rwdata
> , 2024K rodata, 1024K init, 252K bss, 28800K reserved, 0K cma-reserved)

Sounds like you want CMA=0M then?


Cheers,
-- 
Cyril Brulebois -- Debian Consultant @ DEBAMAX -- https://debamax.com/

[toc] | [prev] | [next] | [standalone]

#1176308 — Bug#1056933: [Pkg-raspi-maintainers] Bug#1056933: raspi-firmware: CMA=0 not working as intended

Cyril Brulebois dixit:

>Thorsten Glaser <tg@debian.org> (2023-11-26):
>> Package: raspi-firmware
>> Version: 1.20230405+ds-2
>>
>> When I set CMA=0 the cma=… argument in cmdline.txt is omitted.
>
>This is consistent with the documentation in that file.

It’s not. The file documents:

# To disable CMA allocation entirely, f.e. for a headless setup, set
# CMA=0

But CMA=0 in the file leads to no cma=0 on the kernel command line,
which makes the kernel use the default CMA allocation of 64 MiB.

>Sounds like you want CMA=0M then?

Perhaps, I just threw it here for now:

$ cat /etc/default/raspi-extra-cmdline
TZ=:Europe/Berlin nofb nomodeset cma=0

Meow,
//mirabilos
-- 
15:39⎜«mika:#grml» mira|AO: "mit XFree86® wär’ das nicht passiert" - muhaha
15:48⎜<thkoehler:#grml> also warum machen die xorg Jungs eigentlich alles
kaputt? :)    15:49⎜<novoid:#grml> thkoehler: weil sie als Kinder nie den
gebauten Turm selber umschmeissen durften?	-- ~/.Xmodmap wonders…

[toc] | [prev] | [next] | [standalone]

#1176311 — Bug#1056933: [Pkg-raspi-maintainers] Bug#1056933: raspi-firmware: CMA=0 not working as intended

[Multipart message — attachments visible in raw view] — view raw

Thorsten Glaser <tg@debian.org> (2023-11-27):
> It’s not. The file documents:
> 
> # To disable CMA allocation entirely, f.e. for a headless setup, set
> # CMA=0

Well, that's still the intent behind the commit that introduced support
for that. And since that went into a stable release, I don't see how we
could safely move away from CMA=0 means no cma= settings at all.

> But CMA=0 in the file leads to no cma=0 on the kernel command line,
> which makes the kernel use the default CMA allocation of 64 MiB.
> 
> >Sounds like you want CMA=0M then?
> 
> Perhaps, I just threw it here for now:
> 
> $ cat /etc/default/raspi-extra-cmdline
> TZ=:Europe/Berlin nofb nomodeset cma=0

If you don't want to verify a proposed solution to the needs you
expressed, I suppose this bug report can be closed?
-- 
Cyril Brulebois -- Debian Consultant @ DEBAMAX -- https://debamax.com/

[toc] | [prev] | [next] | [standalone]

#1176318 — Bug#1056933: [Pkg-raspi-maintainers] Bug#1056933: raspi-firmware: CMA=0 not working as intended

Cyril Brulebois dixit:
>Thorsten Glaser <tg@debian.org> (2023-11-27):
>> It’s not. The file documents:
>>
>> # To disable CMA allocation entirely, f.e. for a headless setup, set
>> # CMA=0
>
>Well, that's still the intent behind the commit that introduced support
>for that.

But it totally mismatches the documentation!

>And since that went into a stable release, I don't see how we
>could safely move away from CMA=0 means no cma= settings at all.

Granted, but…

>expressed, I suppose this bug report can be closed?

… definitely not.

Currently, CMA=0 is documented as “disable CMA allocation entirely”,
but it DOES NOT DO THAT. Instead it removes the cma= parameter from
the kernel command line INSTEAD of disabling CMA, which makes the
kernel use its default NONZERO allocation.

At the *very* leaast, that comment must be changed to e.g.

# To disable CMA allocation entirely, e.g. for a headless setup,
# set CMA=0M; setting CMA=0 makes the kernel use the default CMA
# size by omitting the cma parameter from the command line.

bye,
//mirabilos
-- 
21:41⎜«Tonnerre:#nosec» Do at least one thing every day which makes
     ⎜    inspirational quotes lovers sad

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.bugs.dist

csiph-web