Skip to contentSkip to conversation

yacc dependency of aic7xxx driver

DISCUSSION J�rn Nettingsmeier 6 participants 10 messages Mar 7, 2001 – Mar 8, 2001

lore
Share & tools
lore
#0 lore raw

hello justin !

i have just tried to install the latest 2.4.3pre3 kernel with your driver.
it failed with yacc: file not found.
while i could install yacc, i have never had to use it before. i was assuming that the newer bison could do the same thing (which is what i have installed).
so far, the kernel has not relied on yacc, which is why i'd like to ask you if it's possible to make it work with bison.

sorry if this has been dealt with before, but i didn't find anything in the lkml archive.

regards,

jörn

please keep me cc:ed, i'm not on lkml. thanks.

Signature · 4 lines
Jörn Nettingsmeier     
home://Kurfürstenstr.49.45138.Essen.Germany      
phone://+49.201.491621
http://www.folkwang-hochschule.de/~nettings/
#1 lore raw
Quoted 10 lines

hello justin !

i have just tried to install the latest 2.4.3pre3 kernel with your driver.
it failed with yacc: file not found.
while i could install yacc, i have never had to use it before. i was assuming that the newer bison could do the same thing (which is what i have installed).
so far, the kernel has not relied on yacc, which is why i'd like to ask you if it's possible to make it work with bison.

The assembler makefile doesn't reference yacc, but instead relies on gmake's built in rules to figure out how to generate a .c from a .y. I'm somewhat surprised that bison doesn't create a link to yacc or that gmake doesn't try to look for bison.

Oh well. We'll just have to be more careful in how future patches are generated so that the dependency between the generated firmware files and the firmware source only triggers if you are actually performing firmware development. Trying to build this simple assmebler on everyone's systems is turning out to be just too hard.

--
Justin

#2 lore raw
Quoted 4 lines · In article <[email protected]> you wrote:

The assembler makefile doesn't reference yacc, but instead relies on gmake's built in rules to figure out how to generate a .c from a .y. I'm somewhat surprised that bison doesn't create a link to yacc or that gmake doesn't try to look for bison.

Yepp, gmake is kindof stupid in that respect - it does also not try to use flex if there is no lex, etc... I'd really vote for a BSDish sys.mk.

Quoted 6 lines

Oh well. We'll just have to be more careful in how future patches are generated so that the dependency between the generated firmware files and the firmware source only triggers if you are actually performing firmware development. Trying to build this simple assmebler on everyone's systems is turning out to be just too hard.

What about simply removing the firmware source and assembler from the kernel tree? We have lots of firmware in the kernel tree for which there isn't even firmware avaible...

The few people that want to do firware development could download it from your ftp site.

Christoph

Signature · 1 line
Of course it doesn't work. We've performed a software upgrade.
#3 lore raw
Quoted 3 lines

What about simply removing the firmware source and assembler from the kernel tree? We have lots of firmware in the kernel tree for which there isn't even firmware avaible...

What, and not allow others to fix my bugs for me? :-)

Lots of people have embedded this driver just because it is completely open source. I'd like to have all distributions be "complete" distributions.

--
Justin

#4 lore raw
Quoted 4 lines · Justin T. Gibbs
Quoted 3 lines

What about simply removing the firmware source and assembler from the kernel tree? We have lots of firmware in the kernel tree for which there isn't even firmware avaible...

Quoted 8 lines

What, and not allow others to fix my bugs for me? :-)

Lots of people have embedded this driver just because it is completely open source. I'd like to have all distributions be "complete" distributions.

--
Justin

alias yacc='bison -y'

If you don't have one of these, you don't have the tools necessary to do kernel development.

Cheers,
Dick Johnson

Penguin : Linux version 2.4.1 on an i686 machine (799.53 BogoMips).

"Memory is like gasoline. You use it up when you are running. Of course you get it all back when you reboot..."; Actual explanation obtained from the Micro$oft help desk.

#5 lore raw
In reply to Christoph Hellwig
Quoted 9 lines · Justin T. Gibbs
Quoted 3 lines

What about simply removing the firmware source and assembler from the kernel tree? We have lots of firmware in the kernel tree for which there isn't even firmware avaible...

What, and not allow others to fix my bugs for me? :-)

Lots of people have embedded this driver just because it is completely open source. I'd like to have all distributions be "complete" distributions.

I presume the aicasm (or whatever) can do its meager DB needs
by simpler ubiquitous things like GDBM. SleepyCat DBx is good,
but in this case as serious overkill as e.g. using Oracle ;) (No, I didn't read the source of the aicasm.)

There really should be some convenient way to compile the firmware from source when wanted, but the normal case should not be dependent of FIRMWARE SOURCE - the integer/byte array include file should be enough for normal users.

$ (cd drivers/xx/yy; make aic7xxx.fw )

What HOSTCC et.al. settings are needed, that I can't say, and that 'make' method doesn't get top-level definitions in...

Simultaneously (as 'aic7xxx.fw' would not be the real filename), aic7xxx.o file's dependencies would contain the real firmware files so that "make bzImage" will recompile it correctly.

When people like Justin are sending the driver update to Linus, they just need to be carefull at sending all relevant files/diffs.

We are placing a lot of trust to these firmware makers to supply us working firmware, why not trust them to do it right so that compiling kernel won't need to generate firmware assembler program at first...

Quoted 2 lines

--
Justin

/Matti Aarnio

#6 lore raw
Quoted 4 lines

I presume the aicasm (or whatever) can do its meager DB needs
by simpler ubiquitous things like GDBM. SleepyCat DBx is good,
but in this case as serious overkill as e.g. using Oracle ;) (No, I didn't read the source of the aicasm.)

People keep saying stuff like this. I think one of the best things a developer can realize is that being lazy is often a good thing. I already was experience with using db, it took me about 5 minutes to utilize it for this project when rolling my own would have taken much longer, and it happens to be available on almost any platform of interest (including windows which is used to host some of the RTOS environments that also use this driver). More importantly, it works... so why fix it?

People who really want to assemble the firmware will install the required packages. We'll just ship precompiled firmware and make sure the dependencies don't fire even if only a few of the files in the dependency chain are updated.

--
Justin

#7 lore raw
In reply to Christoph Hellwig
Quoted 9 lines · Justin T. Gibbs
Quoted 3 lines

What about simply removing the firmware source and assembler from the kernel tree? We have lots of firmware in the kernel tree for which there isn't even firmware avaible...

What, and not allow others to fix my bugs for me? :-)

Lots of people have embedded this driver just because it is completely open source. I'd like to have all distributions be "complete" distributions.

Not having it in the kernel tree doesn't have to mean not having it in the distributions. If you are really concerned having iy in the kernel tree it might also be a good idea to have it in scripts.

Christoph

Signature · 1 line
Of course it doesn't work. We've performed a software upgrade.
#8 lore raw
In reply to Justin T. Gibbs
Quoted 17 lines · Justin T. Gibbs
Quoted 11 lines

hello justin !

i have just tried to install the latest 2.4.3pre3 kernel with your driver.
it failed with yacc: file not found.
while i could install yacc, i have never had to use it before. i was assuming that the newer bison could do the same thing (which is what i have installed).
so far, the kernel has not relied on yacc, which is why i'd like to ask you if it's possible to make it work with bison.

The assembler makefile doesn't reference yacc, but instead relies on gmake's built in rules to figure out how to generate a .c from a .y. I'm somewhat surprised that bison doesn't create a link to yacc or that gmake doesn't try to look for bison.

Quoted 6 lines

Oh well. We'll just have to be more careful in how future patches are generated so that the dependency between the generated firmware files and the firmware source only triggers if you are actually performing firmware development. Trying to build this simple assmebler on everyone's systems is turning out to be just too hard.

i might also be SuSE 7.1 related, since this was the first kernel i compiled on the new distro.
but since the problem arose only with the new aic driver, i thought it might be that you had a slightly different development environment...?

anyway, robbert muller sent me the following simple workaround:

<quote>
Just create a shell script called yacc with the following content
-------------------
#!/bin/sh
bison --yacc $*
-------------------
i ran into the same problem with a school proiject here yesterday </quote>

regards,

jörn

Signature · 4 lines
Jörn Nettingsmeier     
home://Kurfürstenstr.49.45138.Essen.Germany      
phone://+49.201.491621
http://www.folkwang-hochschule.de/~nettings/
#9 lore raw
In reply to Justin T. Gibbs

[Jörn Nettingsmeier, quoting Robert Muller]

Quoted 6 lines

Just create a shell script called yacc with the following content
-------------------
#!/bin/sh
bison --yacc $*
-------------------
i ran into the same problem with a school proiject here yesterday

Why does nobody learn shell anymore? (Just curious.)

#!/bin/sh
exec bison -y "$@"

Peter