is there (or can there be) a prescribed standard for assertions in
kernel code? as it stands, include/asm-generic/bug.h defines the
following debugging directives:
BUG()
BUG_ON(condition)
WARN_ON(condition)
WARN_ON_ONCE(condition)
which, you would think, would be sufficient to cover the majority of
needs. and yet, loads of people insist on defining some variation of
an assertion for their code:
$ grep -r "define.*assert(" *
in some cases, an assertion is defined directly in terms of BUG_ON():
net/tipc/core.h:#define assert(i) BUG_ON(!(i))
in other cases, it might as well have been:
arch/cris/arch-v32/drivers/cryptocop.c:#define assert(s) do{if (!(s)) panic(#s);} while(0);
so there's two possible issues here:
1) if an assertion is exactly equivalent to one of the existing
directives, should it be removed and that directive be used directly
in the code?
2) if an assertion *isn't* exactly equivalent to an existing
directive, should a new directive be added? what about just adding a
new header file "assert.h" with a universal assert() macro? wouldn't
that make more sense than having all these separate definitions?
rday
--
========================================================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry
Waterloo, Ontario, CANADA
http://fsdev.net/wiki/index.php?title=Main_Page
========================================================================
--
To unsubscribe from this list: send an email with
"unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx
Please read the FAQ at http://kernelnewbies.org/FAQ
