[9fans] a question on style

[9fans] a question on style

[ron minnich]

For some time now, over 10 years anyway, the GNU world has been using
the linker to create initialized structs, viz:
static const struct cpu_driver driver __cpu_driver = {
        .ops = &cpu_dev_ops,
        .id_table = cpu_table,
};

(this from linuxbios)

The __cpu_driver is passed to the linker and you end up with an array
of all structs tagged with that name.

These are called linker sets. I first heard of them in 1995 or so when
freebsd moved away from generating arrays of structs from the BSD
config tool and used gld to create those same arrays.

Plan 9 takes the older style approach: a script parses a config file
and creates C code which initializes the struct arrays.

Now, I betray my age, I guess, but I prefer the plan 9 style. I am
curious: am I just out of touch with this modern world?

Why do I like the Plan 9 style?
1. you can figure out what source constitutes the
program/application/os/bios by looking at C code
2. program analysis tools can analyse C code; you don't need binaries
and linker scripts
3. Errors such as empty structs are easier to catch at build time;
we've seen cases where nobody
    built a binary with, e.g., any cpu drivers and we did not know
until the BIOS tried to start
4. If you are using, e.g., Eclipse or Kscope, you can see how code is
invoked and used, which is
    harder when initialized structs are created by the linker.

OK, am I just out of date or is there a real reason for linker
sets?This question just came up in linuxbios v3 and I am wondering if
I am a stubborn old coot (likely) or if there really is merit to my
dislike of linker sets.

ron

Re: [9fans] a question on style

[erik quanstrom]

> OK, am I just out of date or is there a real reason for linker
> sets?This question just came up in linuxbios v3 and I am wondering if
> I am a stubborn old coot (likely) or if there really is merit to my
> dislike of linker sets.

why can't both be true?

- erik

Re: [9fans] a question on style

[ron minnich]

On Thu, Jul 31, 2008 at 3:04 PM, erik quanstrom <quanstro@quanstro.net> wrote:
>> OK, am I just out of date or is there a real reason for linker
>> sets?This question just came up in linuxbios v3 and I am wondering if
>> I am a stubborn old coot (likely) or if there really is merit to my
>> dislike of linker sets.
>
> why can't both be true?
>

I suspect both are true ...

ron

Re: [9fans] a question on style

[Roman V. Shaposhnik]

On Thu, 2008-07-31 at 15:02 -0700, ron minnich wrote:
> OK, am I just out of date or is there a real reason for linker
> sets?This question just came up in linuxbios v3 and I am wondering if
> I am a stubborn old coot (likely) or if there really is merit to my
> dislike of linker sets.

I tend to dislike any kind of linker magic in user-space quite
profoundly. The most I can tolerate is PLT, I guess. But even
that is something I don't appreciate all that much when I need
to debug low level code. So I agree with you there 100%.

Now, on the other hand, I guess kernel is different though.
Isn't it supposed to be one huge piece of magic to begin with? ;-)

Thanks,
Roman.

Re: [9fans] a question on style

[erik quanstrom]

> Now, on the other hand, I guess kernel is different though.
> Isn't it supposed to be one huge piece of magic to begin with? ;-)

it's supposed to be neither.

- erik

Re: [9fans] a question on style

[Roman V. Shaposhnik]

On Fri, 2008-08-01 at 06:06 -0400, erik quanstrom wrote:
> > Now, on the other hand, I guess kernel is different though.
> > Isn't it supposed to be one huge piece of magic to begin with? ;-)
>
> it's supposed to be neither.

Yeah, but Ron was talking 'bout that kernel from Scandinavia,
right?

Thanks,
Roman.

Re: [9fans] a question on style

[erik quanstrom]

> >> OK, am I just out of date or is there a real reason for linker
> >> sets?This question just came up in linuxbios v3 and I am wondering if
> >> I am a stubborn old coot (likely) or if there really is merit to my
> >> dislike of linker sets.
> >
> > why can't both be true?
> >
>
> I suspect both are true ...

would i be rude to agree with you?

- erik

Re: [9fans] a question on style

[ron minnich]

On Thu, Jul 31, 2008 at 3:17 PM, erik quanstrom <quanstro@quanstro.net> wrote:
>> >> OK, am I just out of date or is there a real reason for linker
>> >> sets?This question just came up in linuxbios v3 and I am wondering if
>> >> I am a stubborn old coot (likely) or if there really is merit to my
>> >> dislike of linker sets.
>> >
>> > why can't both be true?
>> >
>>
>> I suspect both are true ...
>
> would i be rude to agree with you?
>

I have kids. Trust me, you're not being rude.

ron

Re: [9fans] a question on style

[Steve Simon]

> OK, am I just out of date or is there a real reason for linker
> sets?

I see it this way:

	using linker sets means you have to learn and understand the linkers language
	to understand how the system is configured (when trying to track down a problem)

	using a bit of script to generate some tables of C code you just need to
	know C.

This combined with the theroy that there is too much to learn and the less new stuff
I need to learn to get a job done the better makes the choice easy (IMHO).

-Steve

Re: [9fans] a question on style

[Pietro Gagliardi]

On Aug 1, 2008, at 11:37 AM, Steve Simon wrote:

>> OK, am I just out of date or is there a real reason for linker
>> sets?
>
> I see it this way:
>
> 	using linker sets means you have to learn and understand the
> linkers language
> 	to understand how the system is configured (when trying to track
> down a problem)
>
> 	using a bit of script to generate some tables of C code you just
> need to
> 	know C.
>
> This combined with the theroy that there is too much to learn and
> the less new stuff
> I need to learn to get a job done the better makes the choice easy
> (IMHO).
>
> -Steve
>

Plus, linker sets are nonportable. I don't see 8l supporting them, so
I don't think we'll be using them... but all the world uses GCC
nowadays, so why bother with portability anymore?

Re: [9fans] a question on style

[Charles Forsyth]

> but all the world uses GCC
> nowadays, so why bother with portability anymore?

and then when you look at compiling for particular platforms you get to ask
"which gcc?" because code can be non-portable between them
(for the same processor), and the options that work are different too.  such fun.

Re: [9fans] a question on style

[erik quanstrom]

> > but all the world uses GCC
> > nowadays, so why bother with portability anymore?
>
> and then when you look at compiling for particular platforms you get to ask
> "which gcc?" because code can be non-portable between them
> (for the same processor), and the options that work are different too.  such fun.

hey!  less portable than assembly language.  what a feat.

- erik

Re: [9fans] a question on style

[bblochl]

Pietro Gagliardi schrieb:
> On Aug 1, 2008, at 11:37 AM, Steve Simon wrote:
>
>>> OK, am I just out of date or is there a real reason for linker
>>> sets?
>>
>> I see it this way:
>>
>> using linker sets means you have to learn and understand the linkers
>> language
>> to understand how the system is configured (when trying to track down
>> a problem)
>>
>> using a bit of script to generate some tables of C code you just need to
>> know C.
>>
>> This combined with the theroy that there is too much to learn and the
>> less new stuff
>> I need to learn to get a job done the better makes the choice easy
>> (IMHO).
>>
>> -Steve
>>
>
> Plus, linker sets are nonportable. I don't see 8l supporting them, so
> I don't think we'll be using them... but all the world uses GCC
> nowadays, so why bother with portability anymore?
>
>
>
Beside all that different opinions - I don`t want to meddle - there is a
port of gcc for plan9 on http://plan9.bell-labs.com/sources/extra/gcc/.
Never tried it.

Regards

BB

Re: [9fans] a question on style

[erik quanstrom]

> > OK, am I just out of date or is there a real reason for linker
> > sets?
>
> I see it this way:
>
> 	using linker sets means you have to learn and understand the linkers language
> 	to understand how the system is configured (when trying to track down a problem)
>
> 	using a bit of script to generate some tables of C code you just need to
> 	know C.
>
> This combined with the theroy that there is too much to learn and the less new stuff
> I need to learn to get a job done the better makes the choice easy (IMHO).

oddly similar to my argument against using a magic file with file(1).

- erik

Re: [9fans] a question on style

[erik quanstrom]

> > > Now, on the other hand, I guess kernel is different though.
> > > Isn't it supposed to be one huge piece of magic to begin with? ;-)
> >
> > it's supposed to be neither.
>
> Yeah, but Ron was talking 'bout that kernel from Scandinavia,
> right?

uff-da.

- erik