resources.h patch / RLIM_INFINITY __KERNEL__ depend ?

resources.h patch / RLIM_INFINITY __KERNEL__ depend ?

[Florian Lohoff]

Hi,
again - short patch - RLIM_INFINITY - This is also defined in
both glibc 2.0 and glibc 2.1 headers so it should be __KERNEL__
dependend - Shouldnt it ?


Index: resource.h
===================================================================
RCS file: /cvs/linux/include/asm-mips/resource.h,v
retrieving revision 1.5
diff -u -r1.5 resource.h
--- resource.h	2000/02/04 07:40:53	1.5
+++ resource.h	2000/03/30 12:19:42
@@ -25,13 +25,13 @@
 
 #define RLIM_NLIMITS 10			/* Number of limit flavors.  */
 
+#ifdef __KERNEL__
+
 /*
  * SuS says limits have to be unsigned.
  * Which makes a ton more sense anyway.
  */
 #define RLIM_INFINITY	0x7fffffffUL
-
-#ifdef __KERNEL__
 
 #define INIT_RLIMITS					\
 {							\



Flo
-- 
Florian Lohoff		flo@rfc822.org		      	+49-5241-470566
"Technology is a constant battle between manufacturers producing bigger and
more idiot-proof systems and nature producing bigger and better idiots."

Re: resources.h patch / RLIM_INFINITY __KERNEL__ depend ?

[Ralf Baechle]

On Thu, Mar 30, 2000 at 02:27:05PM +0200, Florian Lohoff wrote:

> again - short patch - RLIM_INFINITY - This is also defined in
> both glibc 2.0 and glibc 2.1 headers so it should be __KERNEL__
> dependend - Shouldnt it ?

Glibc 2.1 doesn't include <asm/resource.h>, and 2.0 shouldn't do as well
as per convention.  So the fix is to copy the necessary definitions from
kernel headers into glibc headers.

Thanks for reporting this.  I'm just cooking  a new glibc release ...

  Ralf

Re: resources.h patch / RLIM_INFINITY __KERNEL__ depend ?

[Florian Lohoff]

On Thu, Mar 30, 2000 at 06:47:15PM +0200, Ralf Baechle wrote:
> On Thu, Mar 30, 2000 at 02:27:05PM +0200, Florian Lohoff wrote:
> 
> > again - short patch - RLIM_INFINITY - This is also defined in
> > both glibc 2.0 and glibc 2.1 headers so it should be __KERNEL__
> > dependend - Shouldnt it ?
> 
> Glibc 2.1 doesn't include <asm/resource.h>, and 2.0 shouldn't do as well
> as per convention.  So the fix is to copy the necessary definitions from
> kernel headers into glibc headers.
> 
> Thanks for reporting this.  I'm just cooking  a new glibc release ...

My glibc does include asm/resource.h - This is default hardhat (little endian)

resourcebits.h:#include <asm/resource.h>

This is what caused a lot of userspace builts to fail.

Flo
-- 
Florian Lohoff		flo@rfc822.org		      	+49-5241-470566
"Technology is a constant battle between manufacturers producing bigger and
more idiot-proof systems and nature producing bigger and better idiots."

Re: resources.h patch / RLIM_INFINITY __KERNEL__ depend ?

[Ralf Baechle]

On Fri, Mar 31, 2000 at 10:29:52AM +0200, Florian Lohoff wrote:

> > > again - short patch - RLIM_INFINITY - This is also defined in
> > > both glibc 2.0 and glibc 2.1 headers so it should be __KERNEL__
> > > dependend - Shouldnt it ?
> > 
> > Glibc 2.1 doesn't include <asm/resource.h>, and 2.0 shouldn't do as well
> > as per convention.  So the fix is to copy the necessary definitions from
> > kernel headers into glibc headers.
> > 
> > Thanks for reporting this.  I'm just cooking  a new glibc release ...
> 
> My glibc does include asm/resource.h - This is default hardhat (little endian)
> 
> resourcebits.h:#include <asm/resource.h>
> 
> This is what caused a lot of userspace builts to fail.

I know.  It's something that used to work with kernel 2.2 but breaks with
2.3.  There are a few more problems of this kind with glibc 2.0 and I'm
cooking up a patch that will make glibc usable and compilable with 2.3
as well.

  Ralf