Menu
▾
▴
linux-apus-devel
|
From: Ken T. <ke...@we...> - 2000-06-05 21:08:53
|
Hello, The latest Permedia 2 update as resulted in CONFIG_PCI always being true with no way to remove it - even if a) you have no need for PCI, b) no Permedia 2. What's the consensus on always having PCI support compiled in ? Should it be an option ? Ken. |
|
From: Nicholai B. <nic...@ch...> - 2000-06-05 23:13:33
|
Hello Ken Den 05-Jun-00, skrev Ken Tyler: KT> KT> Hello, KT> KT> The latest Permedia 2 update as resulted in CONFIG_PCI always being true KT> with no way to remove it - even if a) you have no need for PCI, b) no KT> Permedia 2. KT> KT> What's the consensus on always having PCI support compiled in ? KT> KT> Should it be an option ? What??? Even if you don't have a board? That sounds like a bug to me...not again :-( /Nicholai |
|
From: Michel <da...@re...> - 2000-06-06 08:11:20
|
Nicholai Benalal wrote: > KT> The latest Permedia 2 update as resulted in CONFIG_PCI always being true > KT> with no way to remove it - even if a) you have no need for PCI, b) no > KT> Permedia 2. > KT> > KT> What's the consensus on always having PCI support compiled in ? > KT> > KT> Should it be an option ? > > What??? Even if you don't have a board? That sounds like a bug to me...not > again :-( PCI is an architecture thing in the current kernel. Either an architecture has PCI, or it doesn't. If we try to make it dependent on pm2fb, we may run into some kind of chicken and egg problem - pm2fb depends on PCI support... Michel -- People who are wrong the most are wrong the loudest. ______________________________________________________________________________ Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS |
|
From: Alan B. <al...@ms...> - 2000-06-06 09:02:14
|
hi, > PCI is an architecture thing in the current kernel. Either an architecture has > PCI, or it doesn't. ah! of course. alan |
|
From: Alan B. <al...@ms...> - 2000-06-06 08:50:50
|
hi, > The latest Permedia 2 update as resulted in CONFIG_PCI always being true > with no way to remove it - even if a) you have no need for PCI, b) no > Permedia 2. 2) really? If you dont have a PermediaII then there shouldnt be PCI - though if you have a CSPPC or BPPC you do have that PCI stuff present and it could be used to autodetect a pm2 > What's the consensus on always having PCI support compiled in ? > > Should it be an option ? well, it should be an option if those without pm2's dont need PCI alan |
|
From: Sven L. <lu...@dp...> - 2000-06-06 08:57:25
|
On Tue, Jun 06, 2000 at 09:47:32AM +0100, Alan Buxey wrote: > hi, > > > The latest Permedia 2 update as resulted in CONFIG_PCI always being true > > with no way to remove it - even if a) you have no need for PCI, b) no > > Permedia 2. > > 2) really? If you dont have a PermediaII then there shouldnt be PCI - > though if you have a CSPPC or BPPC you do have that PCI stuff present and > it could be used to autodetect a pm2 Well, ... i guess is that you have to _compile_ the kernel with or without pci, so what happens is that in order for a kernel to be able to use, or not use pm2fb, it needs to be compiled with pci. > > What's the consensus on always having PCI support compiled in ? > > > > Should it be an option ? > > well, it should be an option if those without pm2's dont need PCI Well, that said, it should be possible to compile the kernel without pci nor pm2fb, but this should only be a configuration file change i think, ... That is unless i missed something, Michel ? Friendly, Sven |
|
From: Michel <da...@re...> - 2000-06-06 09:19:25
|
Sven LUTHER wrote: > i guess is that you have to _compile_ the kernel with or without pci, so > what happens is that in order for a kernel to be able to use, or not use > pm2fb, it needs to be compiled with pci. Exactly. pm2fb needs PCI to be compiled in. > > > What's the consensus on always having PCI support compiled in ? > > > > > > Should it be an option ? > > > > well, it should be an option if those without pm2's dont need PCI > > Well, that said, it should be possible to compile the kernel without pci nor > pm2fb, but this should only be a configuration file change i think, ... > > That is unless i missed something, Michel ? As I said, PCI support is currently not a configure option but depends on the architecture. If you want to change that, go ahead. I don't think moving even further from the 'official' kernel is a good idea though. Michel -- The Unix Guru's View of Sex: unzip; strip; touch; grep; finger; mount; fsck; more; yes; umount; sleep ______________________________________________________________________________ Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS |
|
From: Sven L. <lu...@dp...> - 2000-06-06 09:38:33
|
On Tue, Jun 06, 2000 at 11:15:56AM +0200, Michel Dänzer wrote: > Sven LUTHER wrote: > > > i guess is that you have to _compile_ the kernel with or without pci, so > > what happens is that in order for a kernel to be able to use, or not use > > pm2fb, it needs to be compiled with pci. > > Exactly. pm2fb needs PCI to be compiled in. > > > > > > What's the consensus on always having PCI support compiled in ? > > > > > > > > Should it be an option ? > > > > > > well, it should be an option if those without pm2's dont need PCI > > > > Well, that said, it should be possible to compile the kernel without pci nor > > pm2fb, but this should only be a configuration file change i think, ... > > > > That is unless i missed something, Michel ? > > As I said, PCI support is currently not a configure option but depends on the > architecture. If you want to change that, go ahead. I don't think moving even > further from the 'official' kernel is a good idea though. What difference are there if you enable PCI or not ? will enabling PCI on all apus kernel be causing a problem ? making the kernel bigger ? I think the best guess would be to have a global option to enable a pci kernel or not, if not we go for the old kernel style, and if yes we go for a pci kernel. If the kernel is pci less, just don't propose pm2fb ? I suppose this would be no more changes than adding a configure option to the configuration stuff ? Friendly, Sven LUTHER |
|
From: Geert U. <ge...@li...> - 2000-06-06 09:44:47
|
On Tue, 6 Jun 2000, Sven LUTHER wrote:
> On Tue, Jun 06, 2000 at 11:15:56AM +0200, Michel Dänzer wrote:
> > Sven LUTHER wrote:
> > As I said, PCI support is currently not a configure option but depends on the
> > architecture. If you want to change that, go ahead. I don't think moving even
> > further from the 'official' kernel is a good idea though.
Some archs do have CONFIG_PCI questions, though. Just like we have a
CONFIG_ZORRO question.
> What difference are there if you enable PCI or not ?
>
> will enabling PCI on all apus kernel be causing a problem ? making the kernel
> bigger ?
It will make the kernel bigger.
> I think the best guess would be to have a global option to enable a pci kernel
> or not, if not we go for the old kernel style, and if yes we go for a pci
> kernel.
Yep, just like CONFIG_ZORRO.
> If the kernel is pci less, just don't propose pm2fb ?
>
> I suppose this would be no more changes than adding a configure option to the
> configuration stuff ?
Just remove the test for CONFIG_AMIGA. People need to be not so stupid to
disable CONFIG_ZORRO if they want to use Zorro devices also, so this sounds OK
to me.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ge...@li...
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
|
|
From: Michel <da...@re...> - 2000-06-06 10:16:27
|
Geert Uytterhoeven wrote: > > > As I said, PCI support is currently not a configure option but depends > > > on the architecture. If you want to change that, go ahead. I don't think > > > moving even further from the 'official' kernel is a good idea though. > > Some archs do have CONFIG_PCI questions, though. Just like we have a > CONFIG_ZORRO question. Well, no problem then :) > > What difference are there if you enable PCI or not ? > > > > will enabling PCI on all apus kernel be causing a problem ? making the > > kernel bigger ? > > It will make the kernel bigger. Can you give an estimate how much? > > I think the best guess would be to have a global option to enable a pci > > kernel or not, if not we go for the old kernel style, and if yes we go for > > a pci kernel. > > Yep, just like CONFIG_ZORRO. OK, I'll do that. Or Ken, do you have time? > > If the kernel is pci less, just don't propose pm2fb ? > > > > I suppose this would be no more changes than adding a configure option to > > the configuration stuff ? > > Just remove the test for CONFIG_AMIGA. People need to be not so stupid to > disable CONFIG_ZORRO if they want to use Zorro devices also, so this sounds > OK to me. What about automatically enabling PCI if the user wants pm2fb? Otherwise, we will have people ask "Why has pm2fb disappeared?" :) Michel -- It's not a bug, it's tradition! ______________________________________________________________________________ Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS |
|
From: Geert U. <Gee...@so...> - 2000-06-06 10:52:59
|
On Tue, 6 Jun 2000, Michel [iso-8859-1] Dänzer wrote:
> Geert Uytterhoeven wrote:
> > > > As I said, PCI support is currently not a configure option but depends
> > > > on the architecture. If you want to change that, go ahead. I don't think
> > > > moving even further from the 'official' kernel is a good idea though.
> >
> > Some archs do have CONFIG_PCI questions, though. Just like we have a
> > CONFIG_ZORRO question.
>
> Well, no problem then :)
>
> > > What difference are there if you enable PCI or not ?
> > >
> > > will enabling PCI on all apus kernel be causing a problem ? making the
> > > kernel bigger ?
> >
> > It will make the kernel bigger.
>
> Can you give an estimate how much?
callisto$ ls -l drivers/pci/*[.oa]
-rw-r--r-- 1 geert users 2792 Jun 5 21:52 drivers/pci/compat.o
-rw-r--r-- 1 geert users 358400 Jun 5 21:52 drivers/pci/names.o
-rw-r--r-- 1 geert users 405592 Jun 5 21:52 drivers/pci/pci.a
-rw-r--r-- 1 geert users 18528 Jun 5 21:52 drivers/pci/pci.o
-rw-r--r-- 1 geert users 29549 Jun 5 21:52 drivers/pci/pci_core.o
-rw-r--r-- 1 geert users 8380 Jun 5 21:52 drivers/pci/proc.o
-r--r--r-- 1 geert users 8213 Jun 5 21:35 drivers/pci/quirks.c
-rw-r--r-- 1 geert users 4824 Jun 5 21:52 drivers/pci/quirks.o
-rw-r--r-- 1 geert users 2480 Jun 5 21:52 drivers/pci/setup-bus.o
-rw-r--r-- 1 geert users 1692 Jun 5 21:52 drivers/pci/setup-irq.o
-rw-r--r-- 1 geert users 3272 Jun 5 21:52 drivers/pci/setup-res.o
-rw-r--r-- 1 geert users 3416 Jun 5 21:52 drivers/pci/syscall.o
Notes:
- I compiled with debugging on
- The name database will be freed after booting (you can also disable it
completely with CONFIG_PCI_NAMES=n)
After using objdump and summing up, I get about 113 KB of initdata (freed
afterwards) and 18 KB permanent (plus a bit for data structures that are
kmalloc'ed).
Not that bad for a nice /proc/bus/pci interface, right? :-)
> > > If the kernel is pci less, just don't propose pm2fb ?
> > >
> > > I suppose this would be no more changes than adding a configure option to
> > > the configuration stuff ?
> >
> > Just remove the test for CONFIG_AMIGA. People need to be not so stupid to
> > disable CONFIG_ZORRO if they want to use Zorro devices also, so this sounds
> > OK to me.
>
> What about automatically enabling PCI if the user wants pm2fb? Otherwise, we
> will have people ask "Why has pm2fb disappeared?" :)
That's much more difficult, since CONFIG_PCI must be y before you can see the
pm2fb question. Besides, my argument was that you must say CONFIG_ZORRO=y as
well if you want to use Zorro graphics cards. So let's put in the FAQ that the
pm2 is a PCI video card (technically it is).
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven ------------- Sony Software Development Center Europe (SDCE)
Gee...@so... ------------------- Sint-Stevens-Woluwestraat 55
Voice +32-2-7248638 Fax +32-2-7262686 ---------------- B-1130 Brussels, Belgium
|
|
From: Michel <da...@re...> - 2000-06-06 12:13:16
|
Geert Uytterhoeven wrote: > > > > will enabling PCI on all apus kernel be causing a problem ? making the > > > > kernel bigger ? > > > > > > It will make the kernel bigger. > > > > Can you give an estimate how much? > > After using objdump and summing up, I get about 113 KB of initdata (freed > afterwards) and 18 KB permanent (plus a bit for data structures that are > kmalloc'ed). > > Not that bad for a nice /proc/bus/pci interface, right? :-) No, but if it's empty... > > What about automatically enabling PCI if the user wants pm2fb? Otherwise, > > we will have people ask "Why has pm2fb disappeared?" :) > > That's much more difficult, since CONFIG_PCI must be y before you can see > the pm2fb question. I take it you are talking about make config. With menuconfig, it would be possible to check for CONFIG_PCI or CONFIG_AMIGA and then set CONFIG_PCI if it isn't before, wouldn't it? > Besides, my argument was that you must say CONFIG_ZORRO=y as well if you > want to use Zorro graphics cards. So let's put in the FAQ that the pm2 is a > PCI video card (technically it is). Agreed. Michel -- UNIX is like Sex: If you don't know it, you don't miss it. But if you know it, you'll need it. ______________________________________________________________________________ Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS |
|
From: Geert U. <Gee...@so...> - 2000-06-06 13:09:06
|
On Tue, 6 Jun 2000, Michel [iso-8859-1] Dänzer wrote:
> Geert Uytterhoeven wrote:
> > > What about automatically enabling PCI if the user wants pm2fb? Otherwise,
> > > we will have people ask "Why has pm2fb disappeared?" :)
> >
> > That's much more difficult, since CONFIG_PCI must be y before you can see
> > the pm2fb question.
>
> I take it you are talking about make config. With menuconfig, it would be
> possible to check for CONFIG_PCI or CONFIG_AMIGA and then set CONFIG_PCI if it
> isn't before, wouldn't it?
Nope. That's not the way to go. It might work with menuconfig, but not with
plain config. Perhaps with ESR's new config scheme proposal?
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven ------------- Sony Software Development Center Europe (SDCE)
Gee...@so... ------------------- Sint-Stevens-Woluwestraat 55
Voice +32-2-7248638 Fax +32-2-7262686 ---------------- B-1130 Brussels, Belgium
|
|
From: Michel <da...@re...> - 2000-06-06 13:39:34
|
Geert Uytterhoeven wrote:
> > > > What about automatically enabling PCI if the user wants pm2fb?
> > > > Otherwise, we will have people ask "Why has pm2fb disappeared?" :)
> > >
> > > That's much more difficult, since CONFIG_PCI must be y before you can
> > > see the pm2fb question.
> >
> > I take it you are talking about make config. With menuconfig, it would be
> > possible to check for CONFIG_PCI or CONFIG_AMIGA and then set CONFIG_PCI
> > if it isn't before, wouldn't it?
>
> Nope. That's not the way to go. It might work with menuconfig, but not with
> plain config. Perhaps with ESR's new config scheme proposal?
What is it about?
However, we're talking about 2.2, so I guess we have to stick to
{menu,,x}config.
Michel
--
It's not a bug, it's tradition!
______________________________________________________________________________
Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS
|
|
From: Geert U. <ge...@li...> - 2000-06-06 13:45:35
|
On Tue, 6 Jun 2000, Michel [iso-8859-1] Dänzer wrote: > Geert Uytterhoeven wrote: > > > > > What about automatically enabling PCI if the user wants pm2fb? > > > > > Otherwise, we will have people ask "Why has pm2fb disappeared?" :) > > > > > > > > That's much more difficult, since CONFIG_PCI must be y before you can > > > > see the pm2fb question. > > > > > > I take it you are talking about make config. With menuconfig, it would be > > > possible to check for CONFIG_PCI or CONFIG_AMIGA and then set CONFIG_PCI > > > if it isn't before, wouldn't it? > > > > Nope. That's not the way to go. It might work with menuconfig, but not with > > plain config. Perhaps with ESR's new config scheme proposal? > > What is it about? About a new configuration scheme with better dependencies and multiple interfaces. Read all about it at linux-kernel or at ESR's homepage (http://www.tuxedo.org/~esr/ IIRC). Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ge...@li... In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds |
|
From: Michel <da...@re...> - 2000-06-06 19:50:40
|
Michel Dänzer wrote: > > > I think the best guess would be to have a global option to enable a pci > > > kernel or not, if not we go for the old kernel style, and if yes we go > > > for a pci kernel. > > > > Yep, just like CONFIG_ZORRO. > > OK, I'll do that. Committed. I've also added something to Configure.help. Michel -- Me? A skeptic? Can you prove it? ______________________________________________________________________________ Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS |
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X