powerpc-ieee1275 and grub-mkfont dependency

powerpc-ieee1275 and grub-mkfont dependency

[Aaro Koskinen]

Hi,

When upgrading from 2.00 to 2.02~beta3, I noticed that on powerpc it's
no longer possible to build with --disable-grub-mkfont: ./configure
forces it and complains about missing fontconfig etc.

However, when I delete those checks from the configure script, I'm able
to build and install a working bootloader for my G4 & G5 Macs (with the
same functionality as with 2.00, using either OF console or serial).

So I think those configure checks should be relaxed. I think it's overkill
to require fontconfig etc. stuff just to build a serial console bootloader
for G5 server.

A.

Re: powerpc-ieee1275 and grub-mkfont dependency

[Daniel Kiper]

Hi Aaro,

On Wed, Nov 09, 2016 at 12:23:59AM +0200, Aaro Koskinen wrote:
> Hi,
>
> When upgrading from 2.00 to 2.02~beta3, I noticed that on powerpc it's
> no longer possible to build with --disable-grub-mkfont: ./configure
> forces it and complains about missing fontconfig etc.
>
> However, when I delete those checks from the configure script, I'm able
> to build and install a working bootloader for my G4 & G5 Macs (with the
> same functionality as with 2.00, using either OF console or serial).
>
> So I think those configure checks should be relaxed. I think it's overkill
> to require fontconfig etc. stuff just to build a serial console bootloader
> for G5 server.

No problem. Could you send us a fix for this issue?

Daniel

Re: powerpc-ieee1275 and grub-mkfont dependency

[Aaro Koskinen]

Hi,

On Wed, Nov 09, 2016 at 07:46:54PM +0100, Daniel Kiper wrote:
> On Wed, Nov 09, 2016 at 12:23:59AM +0200, Aaro Koskinen wrote:
> > When upgrading from 2.00 to 2.02~beta3, I noticed that on powerpc it's
> > no longer possible to build with --disable-grub-mkfont: ./configure
> > forces it and complains about missing fontconfig etc.
> >
> > However, when I delete those checks from the configure script, I'm able
> > to build and install a working bootloader for my G4 & G5 Macs (with the
> > same functionality as with 2.00, using either OF console or serial).
> >
> > So I think those configure checks should be relaxed. I think it's overkill
> > to require fontconfig etc. stuff just to build a serial console bootloader
> > for G5 server.
> 
> No problem. Could you send us a fix for this issue?

I'm using the below patch.

A.

---8<---

From beb8ca2c32e1720686a476b61d492ad2b5ed2c3a Mon Sep 17 00:00:00 2001
From: Aaro Koskinen <aaro.koskinen@iki.fi>
Date: Wed, 9 Nov 2016 23:26:43 +0200
Subject: [PATCH] configure.ac: don't require build time grub-mkfont on
 powerpc-ieee1275

Don't require build time grub-mkfont on powerpc-ieee1275.

Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
---
 configure.ac | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/configure.ac b/configure.ac
index 9ddfc53..3d25510 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1555,11 +1555,11 @@ if test x"$grub_build_mkfont_excuse" = x ; then
 else
   enable_build_grub_mkfont=no
 fi
-if test x"$enable_build_grub_mkfont" = xno && ( test "x$platform" = xqemu || test "x$platform" = xloongson || test "x$platform" = xqemu_mips || test "x$target_cpu"-"$platform" = xpowerpc-ieee1275 || test "x$platform" = xcoreboot ); then
+if test x"$enable_build_grub_mkfont" = xno && ( test "x$platform" = xqemu || test "x$platform" = xloongson || test "x$platform" = xqemu_mips || test "x$platform" = xcoreboot ); then
   if test x"$grub_build_mkfont_excuse" = x ; then
-    AC_MSG_ERROR([qemu, powerpc-ieee1275, coreboot and loongson ports needs build-time grub-mkfont])
+    AC_MSG_ERROR([qemu, coreboot and loongson ports need build-time grub-mkfont])
   else
-    AC_MSG_ERROR([qemu, powerpc-ieee1275, coreboot and loongson ports needs build-time grub-mkfont ($grub_build_mkfont_excuse)])
+    AC_MSG_ERROR([qemu, coreboot and loongson ports need build-time grub-mkfont ($grub_build_mkfont_excuse)])
   fi
 fi
 
@@ -1629,11 +1629,11 @@ if test x"$enable_build_grub_mkfont" = xno ; then
   FONT_SOURCE=
 fi
 
-if test "x$FONT_SOURCE" = x && ( test "x$platform" = xqemu || test "x$platform" = xloongson || test "x$platform" = xqemu_mips || test "x$target_cpu"-"$platform" = xpowerpc-ieee1275 || test "x$platform" = xcoreboot ); then
+if test "x$FONT_SOURCE" = x && ( test "x$platform" = xqemu || test "x$platform" = xloongson || test "x$platform" = xqemu_mips || test "x$platform" = xcoreboot ); then
   if test x"$grub_build_mkfont_excuse" = x ; then
-    AC_MSG_ERROR([qemu, powerpc-ieee1275, coreboot and loongson ports need unifont])
+    AC_MSG_ERROR([qemu, coreboot and loongson ports need unifont])
   else
-    AC_MSG_ERROR([qemu, powerpc-ieee1275, coreboot and loongson ports need unifont ($grub_build_mkfont_excuse)])
+    AC_MSG_ERROR([qemu, coreboot and loongson ports need unifont ($grub_build_mkfont_excuse)])
   fi
 fi
 
-- 
2.9.2

Re: powerpc-ieee1275 and grub-mkfont dependency

[Andrei Borzenkov]

I do not like it, sorry. If this platform supports framebuffer we
should build grub with framebuffer. If you do not (want to) use it -
fine, nobody forces you to do it. But on this path we get
proliferation of slightly different incompatible binaries in the wild.

On Thu, Nov 10, 2016 at 12:52 AM, Aaro Koskinen <aaro.koskinen@iki.fi> wrote:
> Hi,
>
> On Wed, Nov 09, 2016 at 07:46:54PM +0100, Daniel Kiper wrote:
>> On Wed, Nov 09, 2016 at 12:23:59AM +0200, Aaro Koskinen wrote:
>> > When upgrading from 2.00 to 2.02~beta3, I noticed that on powerpc it's
>> > no longer possible to build with --disable-grub-mkfont: ./configure
>> > forces it and complains about missing fontconfig etc.
>> >
>> > However, when I delete those checks from the configure script, I'm able
>> > to build and install a working bootloader for my G4 & G5 Macs (with the
>> > same functionality as with 2.00, using either OF console or serial).
>> >
>> > So I think those configure checks should be relaxed. I think it's overkill
>> > to require fontconfig etc. stuff just to build a serial console bootloader
>> > for G5 server.
>>
>> No problem. Could you send us a fix for this issue?
>
> I'm using the below patch.
>
> A.
>
> ---8<---
>
> From beb8ca2c32e1720686a476b61d492ad2b5ed2c3a Mon Sep 17 00:00:00 2001
> From: Aaro Koskinen <aaro.koskinen@iki.fi>
> Date: Wed, 9 Nov 2016 23:26:43 +0200
> Subject: [PATCH] configure.ac: don't require build time grub-mkfont on
>  powerpc-ieee1275
>
> Don't require build time grub-mkfont on powerpc-ieee1275.
>
> Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
> ---
>  configure.ac | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)
>
> diff --git a/configure.ac b/configure.ac
> index 9ddfc53..3d25510 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -1555,11 +1555,11 @@ if test x"$grub_build_mkfont_excuse" = x ; then
>  else
>    enable_build_grub_mkfont=no
>  fi
> -if test x"$enable_build_grub_mkfont" = xno && ( test "x$platform" = xqemu || test "x$platform" = xloongson || test "x$platform" = xqemu_mips || test "x$target_cpu"-"$platform" = xpowerpc-ieee1275 || test "x$platform" = xcoreboot ); then
> +if test x"$enable_build_grub_mkfont" = xno && ( test "x$platform" = xqemu || test "x$platform" = xloongson || test "x$platform" = xqemu_mips || test "x$platform" = xcoreboot ); then
>    if test x"$grub_build_mkfont_excuse" = x ; then
> -    AC_MSG_ERROR([qemu, powerpc-ieee1275, coreboot and loongson ports needs build-time grub-mkfont])
> +    AC_MSG_ERROR([qemu, coreboot and loongson ports need build-time grub-mkfont])
>    else
> -    AC_MSG_ERROR([qemu, powerpc-ieee1275, coreboot and loongson ports needs build-time grub-mkfont ($grub_build_mkfont_excuse)])
> +    AC_MSG_ERROR([qemu, coreboot and loongson ports need build-time grub-mkfont ($grub_build_mkfont_excuse)])
>    fi
>  fi
>
> @@ -1629,11 +1629,11 @@ if test x"$enable_build_grub_mkfont" = xno ; then
>    FONT_SOURCE=
>  fi
>
> -if test "x$FONT_SOURCE" = x && ( test "x$platform" = xqemu || test "x$platform" = xloongson || test "x$platform" = xqemu_mips || test "x$target_cpu"-"$platform" = xpowerpc-ieee1275 || test "x$platform" = xcoreboot ); then
> +if test "x$FONT_SOURCE" = x && ( test "x$platform" = xqemu || test "x$platform" = xloongson || test "x$platform" = xqemu_mips || test "x$platform" = xcoreboot ); then
>    if test x"$grub_build_mkfont_excuse" = x ; then
> -    AC_MSG_ERROR([qemu, powerpc-ieee1275, coreboot and loongson ports need unifont])
> +    AC_MSG_ERROR([qemu, coreboot and loongson ports need unifont])
>    else
> -    AC_MSG_ERROR([qemu, powerpc-ieee1275, coreboot and loongson ports need unifont ($grub_build_mkfont_excuse)])
> +    AC_MSG_ERROR([qemu, coreboot and loongson ports need unifont ($grub_build_mkfont_excuse)])
>    fi
>  fi
>
> --
> 2.9.2
>
>
> _______________________________________________
> Grub-devel mailing list
> Grub-devel@gnu.org
> https://lists.gnu.org/mailman/listinfo/grub-devel

Re: powerpc-ieee1275 and grub-mkfont dependency

[Aaro Koskinen]

Hi,

On Thu, Nov 10, 2016 at 04:53:23PM +0300, Andrei Borzenkov wrote:
> I do not like it, sorry. If this platform supports framebuffer we
> should build grub with framebuffer. If you do not (want to) use it -
> fine, nobody forces you to do it. But on this path we get
> proliferation of slightly different incompatible binaries in the wild.

What's then the point of having --disable-grub-mkfont (or any
--disable/enable-* option)?

A.

Re: powerpc-ieee1275 and grub-mkfont dependency

[Andrei Borzenkov]

15.11.2016 01:13, Aaro Koskinen пишет:
> Hi,
> 
> On Thu, Nov 10, 2016 at 04:53:23PM +0300, Andrei Borzenkov wrote:
>> I do not like it, sorry. If this platform supports framebuffer we
>> should build grub with framebuffer. If you do not (want to) use it -
>> fine, nobody forces you to do it. But on this path we get
>> proliferation of slightly different incompatible binaries in the wild.
> 
> What's then the point of having --disable-grub-mkfont (or any
> --disable/enable-* option)?
> 

Non of them change boot time code. They affect only user space tools.
User space varies greatly, some simply miss support for some features
(like having device-mapper libraries on native Windows build). But you
always get identical boot time code.

Re: powerpc-ieee1275 and grub-mkfont dependency

[Daniel Kiper]

On Tue, Nov 15, 2016 at 06:45:58AM +0300, Andrei Borzenkov wrote:
> 15.11.2016 01:13, Aaro Koskinen ??????????:
> > Hi,
> >
> > On Thu, Nov 10, 2016 at 04:53:23PM +0300, Andrei Borzenkov wrote:
> >> I do not like it, sorry. If this platform supports framebuffer we
> >> should build grub with framebuffer. If you do not (want to) use it -
> >> fine, nobody forces you to do it. But on this path we get
> >> proliferation of slightly different incompatible binaries in the wild.
> >
> > What's then the point of having --disable-grub-mkfont (or any
> > --disable/enable-* option)?
> >
>
> Non of them change boot time code. They affect only user space tools.
> User space varies greatly, some simply miss support for some features
> (like having device-mapper libraries on native Windows build). But you
> always get identical boot time code.

Why we need identical boot time code?

Daniel

Re: powerpc-ieee1275 and grub-mkfont dependency

[Aaro Koskinen]

Hi,

On Tue, Nov 15, 2016 at 06:45:58AM +0300, Andrei Borzenkov wrote:
> 15.11.2016 01:13, Aaro Koskinen пишет:
> > On Thu, Nov 10, 2016 at 04:53:23PM +0300, Andrei Borzenkov wrote:
> >> I do not like it, sorry. If this platform supports framebuffer we
> >> should build grub with framebuffer. If you do not (want to) use it -
> >> fine, nobody forces you to do it. But on this path we get
> >> proliferation of slightly different incompatible binaries in the wild.
> > 
> > What's then the point of having --disable-grub-mkfont (or any
> > --disable/enable-* option)?
> 
> Non of them change boot time code. They affect only user space tools.
> User space varies greatly, some simply miss support for some features
> (like having device-mapper libraries on native Windows build). But you
> always get identical boot time code.

OK, I think I understand. And that makes sense from maintainer point
of view. But as an user I don't want to build stuff I will never
need. Especially when it forces me to build/install/maintain 5-10
additional packages thanks to dependencies.

And you're right, nobody forces me to use grub - I just realized that
openfirmware can boot Linux directly. On an old G4 Mac there are size
limitations but I can probably craft a minimal kernel and chain load
with kexec... This way I wouldn't need to worry about how to build grub
in the future.

Thanks,

A.

Re: powerpc-ieee1275 and grub-mkfont dependency

[Daniel Kiper]

On Tue, Nov 15, 2016 at 11:44:16AM +0100, Daniel Kiper wrote:
> On Tue, Nov 15, 2016 at 06:45:58AM +0300, Andrei Borzenkov wrote:
> > 15.11.2016 01:13, Aaro Koskinen ??????????:
> > > Hi,
> > >
> > > On Thu, Nov 10, 2016 at 04:53:23PM +0300, Andrei Borzenkov wrote:
> > >> I do not like it, sorry. If this platform supports framebuffer we
> > >> should build grub with framebuffer. If you do not (want to) use it -
> > >> fine, nobody forces you to do it. But on this path we get
> > >> proliferation of slightly different incompatible binaries in the wild.
> > >
> > > What's then the point of having --disable-grub-mkfont (or any
> > > --disable/enable-* option)?
> > >
> >
> > Non of them change boot time code. They affect only user space tools.
> > User space varies greatly, some simply miss support for some features
> > (like having device-mapper libraries on native Windows build). But you
> > always get identical boot time code.
>
> Why we need identical boot time code?

Ping?

Daniel

Re: powerpc-ieee1275 and grub-mkfont dependency

[Andrei Borzenkov]

22.11.2016 12:00, Daniel Kiper пишет:
> On Tue, Nov 15, 2016 at 11:44:16AM +0100, Daniel Kiper wrote:
>> On Tue, Nov 15, 2016 at 06:45:58AM +0300, Andrei Borzenkov wrote:
>>> 15.11.2016 01:13, Aaro Koskinen ??????????:
>>>> Hi,
>>>>
>>>> On Thu, Nov 10, 2016 at 04:53:23PM +0300, Andrei Borzenkov wrote:
>>>>> I do not like it, sorry. If this platform supports framebuffer we
>>>>> should build grub with framebuffer. If you do not (want to) use it -
>>>>> fine, nobody forces you to do it. But on this path we get
>>>>> proliferation of slightly different incompatible binaries in the wild.
>>>>
>>>> What's then the point of having --disable-grub-mkfont (or any
>>>> --disable/enable-* option)?
>>>>
>>>
>>> Non of them change boot time code. They affect only user space tools.
>>> User space varies greatly, some simply miss support for some features
>>> (like having device-mapper libraries on native Windows build). But you
>>> always get identical boot time code.
>>
>> Why we need identical boot time code?
> 
> Ping?
> 

OK, I misunderstood this, sorry. This is just about generating font, not
about changing code.

I will commit it.

Re: powerpc-ieee1275 and grub-mkfont dependency

[Daniel Kiper]

On Tue, Nov 22, 2016 at 08:37:56PM +0300, Andrei Borzenkov wrote:
> 22.11.2016 12:00, Daniel Kiper ??????????:
> > On Tue, Nov 15, 2016 at 11:44:16AM +0100, Daniel Kiper wrote:
> >> On Tue, Nov 15, 2016 at 06:45:58AM +0300, Andrei Borzenkov wrote:
> >>> 15.11.2016 01:13, Aaro Koskinen ??????????:
> >>>> Hi,
> >>>>
> >>>> On Thu, Nov 10, 2016 at 04:53:23PM +0300, Andrei Borzenkov wrote:
> >>>>> I do not like it, sorry. If this platform supports framebuffer we
> >>>>> should build grub with framebuffer. If you do not (want to) use it -
> >>>>> fine, nobody forces you to do it. But on this path we get
> >>>>> proliferation of slightly different incompatible binaries in the wild.
> >>>>
> >>>> What's then the point of having --disable-grub-mkfont (or any
> >>>> --disable/enable-* option)?
> >>>>
> >>>
> >>> Non of them change boot time code. They affect only user space tools.
> >>> User space varies greatly, some simply miss support for some features
> >>> (like having device-mapper libraries on native Windows build). But you
> >>> always get identical boot time code.
> >>
> >> Why we need identical boot time code?
> >
> > Ping?
>
> OK, I misunderstood this, sorry. This is just about generating font, not
> about changing code.
>
> I will commit it.

Great! Thanks a lot!

Daniel