cGull wrote:
>>The problem with assembly language is that it is processor specific.
>>Since Linux runs on a _variety_ of processors, it makes no sense to
>>write drivers specific to one platform. The other problem is
>>
>>
>
>I totally disagree with you and Dirk. The "problem" lies firmly with
>your attitude to linux and not with processor specifics.
>
>
With Linux , you have the option to agree or disagree, well that's the
beauty of Linux..
Well, i did not mean anything other than for my own usage. Everyone is
free to use their choice. The Linux kernel coding style is basically
limited to "C". You can do a google to get an idea of "how even the code
should follow certain styles in C." Not even C++.
First of all, i will make it clear that i am more of an assmebly based
coder rather than "C". But "when you are in Rome, one has to be a Roman"
otherwise it will be difficult.
>Secondly you guys are just too **** serious. If someone wants to write
>assembler drivers shut up and let them get on with having fun.
>The whole point of linux to me is to have as much fun as I can and a
>bunch of pedantic old bores aren't going to stop me.
>
>
I wasn't stopping you from writing any drivers in asm. Well, what i was
saying is that , first of all if i write my drivers in asm.
1) Linus and gang will _not_ accept it under any circumstance for
inclusion in the mainline kernel.
Well, a LUG itself has so many flamewar's for _nothing_. Imagine
LKML with a lot of kernel developers.. (the ego's are still bigger
there). Imagine Linux users having a big ego, think what the developers
would be having then ;-)
2) I will _not_ be able to use any subsystem, other than the asm-lib
routines included alonwith the assembler.
Suppose i want to use a subsystem call rather than a generic
syscall. Do you mean to say that, you will write that entire subsyatem
in assembly... ? ie, write your driver as a complete Block or Character
driver.. ?
3) My maintenance issue will be greater, as i will have to work around
with processor based quirks/issues too other than my block driver
issues. The time where i can devote to fixing driver bugs/issues, i will
use it fix the subsystem that i rewrote in asm.. ?
Another issue is that when your code goes mainline, the most important
aspect is that , there should be no duplication. Imagine all that
duplication you do in asm.
Developers will have _only_ objections for your code to go mainline.
Basically one can say for certain that it will not go mainline. Well we
can see the thread for "I request the inclusion of SAS". :-) You will
get a basic idea of what it sounds in reality.
But none of these hold good, if one is writing drivers for fun, ie. use
by a single person, ie, not distributed in any manner.
4) Writing a decent driver in asm under Linux, for a complex subsystem
(Networking, SCSI, DVB, V4L etc), do you know how much overhead it can
bring .. ? Imagine writing a network stack with all the ASN.1 style in
ASM. Wow ! or maybe some CA stack in asm ..
BTW, DVB and V4L subsystems got merged ..
5) Now when you write code, and get paid for it, The people who pays you
will expect some sort of seriousness, rather than simple fun. Nobody
pays you for "fun".
If you get some standard drivers under Linux ported from "C" to "asm" i
would love to see it. Please do let us know how you get along .. ;-)
Manu
------------------------ Yahoo! Groups Sponsor --------------------~-->
Get fast access to your favorite Yahoo! Groups. Make Yahoo! your home page
http://us.click.yahoo.com/dpRU5A/wUILAA/yQLSAA/0XFolB/TM
--------------------------------------------------------------------~->
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/dubailug/
<*> To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
