* [PATCH][2.6] constant_test_bit doesn't like my gcc
@ 2003-10-16 3:13 Zwane Mwaikambo
2003-10-16 4:10 ` Andrew Morton
0 siblings, 1 reply; 8+ messages in thread
From: Zwane Mwaikambo @ 2003-10-16 3:13 UTC (permalink / raw)
To: Linux Kernel; +Cc: Andrew Morton
[-- Attachment #1: Type: TEXT/PLAIN, Size: 2229 bytes --]
cpu_has_foo and friends don't work at all with my gcc 3.2.2-5 and i'm
branching off into all sorts of tests (which is another story...). i also
took the liberty of removing the const volatile...
Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/3.2.2/specs
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
--infodir=/usr/share/info --enable-shared --enable-threads=posix
--disable-checking --with-system-zlib --enable-__cxa_atexit
--host=i386-redhat-linux
Thread model: posix
gcc version 3.2.2 20030222 (Red Hat Linux 3.2.2-5)
0x08048328 <main+0>: push %ebp
0x08048329 <main+1>: mov %esp,%ebp
0x0804832b <main+3>: sub $0x8,%esp
0x0804832e <main+6>: mov 0xfffffffc(%ebp),%eax
0x08048331 <main+9>: and $0xfffffff0,%esp
0x08048334 <main+12>: shr $0x3,%eax
0x08048337 <main+15>: sub $0x8,%esp
0x0804833a <main+18>: and $0x1,%eax
0x0804833d <main+21>: push %eax
0x0804833e <main+22>: push $0x8048400
0x08048343 <main+27>: movl $0xff,0xfffffffc(%ebp)
0x0804834a <main+34>: call 0x8048268 <printf>
0x0804834f <main+39>: xor %eax,%eax
0x08048351 <main+41>: leave
Index: linux-2.6.0-test7-mm1/include/asm-i386/bitops.h
===================================================================
RCS file: /build/cvsroot/linux-2.6.0-test7-mm1/include/asm-i386/bitops.h,v
retrieving revision 1.1.1.1
diff -u -p -B -r1.1.1.1 bitops.h
--- linux-2.6.0-test7-mm1/include/asm-i386/bitops.h 15 Oct 2003 09:02:10 -0000 1.1.1.1
+++ linux-2.6.0-test7-mm1/include/asm-i386/bitops.h 16 Oct 2003 03:04:34 -0000
@@ -239,12 +239,12 @@ static __inline__ int test_and_change_bi
static int test_bit(int nr, const volatile void * addr);
#endif
-static __inline__ int constant_test_bit(int nr, const volatile unsigned long * addr)
+static __inline__ int constant_test_bit(int nr, const unsigned long * addr)
{
- return ((1UL << (nr & 31)) & (((const volatile unsigned int *) addr)[nr >> 5])) != 0;
+ return ((1UL << (nr & 31)) & (((const unsigned long *) addr)[nr >> 5])) != 0;
}
-static __inline__ int variable_test_bit(int nr, const volatile unsigned long * addr)
+static __inline__ int variable_test_bit(int nr, const unsigned long * addr)
{
int oldbit;
[-- Attachment #2: Type: TEXT/PLAIN, Size: 368 bytes --]
#include <stdio.h>
static int __inline__ constant_test_bit(int nr, const volatile unsigned long * addr)
{
return ((1UL << (nr & 31)) & (((const volatile unsigned int *) addr)[nr >> 5])) != 0;
}
int main()
{
unsigned long foo[2];
*foo = 0xff;
foo[1] = 0xff;
printf("%d %d\n", constant_test_bit(3, foo), constant_test_bit(33, foo));
return 0;
}
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH][2.6] constant_test_bit doesn't like my gcc
2003-10-16 3:13 [PATCH][2.6] constant_test_bit doesn't like my gcc Zwane Mwaikambo
@ 2003-10-16 4:10 ` Andrew Morton
2003-10-16 4:12 ` Zwane Mwaikambo
0 siblings, 1 reply; 8+ messages in thread
From: Andrew Morton @ 2003-10-16 4:10 UTC (permalink / raw)
To: Zwane Mwaikambo; +Cc: linux-kernel
Zwane Mwaikambo <zwane@arm.linux.org.uk> wrote:
>
> cpu_has_foo and friends don't work at all with my gcc 3.2.2-5 and i'm
> branching off into all sorts of tests (which is another story...). i also
> took the liberty of removing the const volatile...
The volatile is rather important.
static inline int constant_test_bit(int nr, const unsigned long *addr)
{
return ((1UL << (nr & 31)) & (((const unsigned long *) addr)[nr >> 5])) != 0;
}
foo(unsigned long *p)
{
while (constant_test_bit(0, p))
;
}
.file "a.c"
.text
.align 2
.p2align 2,,3
.globl foo
.type foo,@function
foo:
pushl %ebp
movl %esp, %ebp
movl 8(%ebp), %eax
testb $1, (%eax)
.p2align 2,,3
.L2:
jne .L2
leave
ret
.Lfe1:
.size foo,.Lfe1-foo
.ident "GCC: (GNU) 3.2 20020903 (Red Hat Linux 8.0 3.2-7)"
See, the little busywait loop doesn't reevaluate the test each time around:
it locks up instead.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH][2.6] constant_test_bit doesn't like my gcc
2003-10-16 4:10 ` Andrew Morton
@ 2003-10-16 4:12 ` Zwane Mwaikambo
2003-10-16 4:21 ` Andrew Morton
0 siblings, 1 reply; 8+ messages in thread
From: Zwane Mwaikambo @ 2003-10-16 4:12 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
On Wed, 15 Oct 2003, Andrew Morton wrote:
> Zwane Mwaikambo <zwane@arm.linux.org.uk> wrote:
> >
> > cpu_has_foo and friends don't work at all with my gcc 3.2.2-5 and i'm
> > branching off into all sorts of tests (which is another story...). i also
> > took the liberty of removing the const volatile...
>
> The volatile is rather important.
Good point, how about;
Index: linux-2.6.0-test7-mm1/include/asm-i386/bitops.h
===================================================================
RCS file: /build/cvsroot/linux-2.6.0-test7-mm1/include/asm-i386/bitops.h,v
retrieving revision 1.1.1.1
diff -u -p -B -r1.1.1.1 bitops.h
--- linux-2.6.0-test7-mm1/include/asm-i386/bitops.h 15 Oct 2003 09:02:10 -0000 1.1.1.1
+++ linux-2.6.0-test7-mm1/include/asm-i386/bitops.h 16 Oct 2003 04:10:37 -0000
@@ -241,7 +241,7 @@ static int test_bit(int nr, const volati
static __inline__ int constant_test_bit(int nr, const volatile unsigned long * addr)
{
- return ((1UL << (nr & 31)) & (((const volatile unsigned int *) addr)[nr >> 5])) != 0;
+ return ((1UL << (nr & 31)) & (addr[nr >> 5])) != 0;
}
static __inline__ int variable_test_bit(int nr, const volatile unsigned long * addr)
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH][2.6] constant_test_bit doesn't like my gcc
2003-10-16 4:12 ` Zwane Mwaikambo
@ 2003-10-16 4:21 ` Andrew Morton
2003-10-16 4:22 ` Zwane Mwaikambo
0 siblings, 1 reply; 8+ messages in thread
From: Andrew Morton @ 2003-10-16 4:21 UTC (permalink / raw)
To: Zwane Mwaikambo; +Cc: linux-kernel
Zwane Mwaikambo <zwane@arm.linux.org.uk> wrote:
>
> > The volatile is rather important.
>
> Good point, how about;
>
> Index: linux-2.6.0-test7-mm1/include/asm-i386/bitops.h
> ===================================================================
> RCS file: /build/cvsroot/linux-2.6.0-test7-mm1/include/asm-i386/bitops.h,v
> retrieving revision 1.1.1.1
> diff -u -p -B -r1.1.1.1 bitops.h
> --- linux-2.6.0-test7-mm1/include/asm-i386/bitops.h 15 Oct 2003 09:02:10 -0000 1.1.1.1
> +++ linux-2.6.0-test7-mm1/include/asm-i386/bitops.h 16 Oct 2003 04:10:37 -0000
> @@ -241,7 +241,7 @@ static int test_bit(int nr, const volati
>
> static __inline__ int constant_test_bit(int nr, const volatile unsigned long * addr)
> {
> - return ((1UL << (nr & 31)) & (((const volatile unsigned int *) addr)[nr >> 5])) != 0;
> + return ((1UL << (nr & 31)) & (addr[nr >> 5])) != 0;
> }
>
Looks fine. Does your compiler get this right?
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH][2.6] constant_test_bit doesn't like my gcc
2003-10-16 4:21 ` Andrew Morton
@ 2003-10-16 4:22 ` Zwane Mwaikambo
2003-10-16 10:55 ` Ingo Oeser
0 siblings, 1 reply; 8+ messages in thread
From: Zwane Mwaikambo @ 2003-10-16 4:22 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
On Wed, 15 Oct 2003, Andrew Morton wrote:
> > static __inline__ int constant_test_bit(int nr, const volatile unsigned long * addr)
> > {
> > - return ((1UL << (nr & 31)) & (((const volatile unsigned int *) addr)[nr >> 5])) != 0;
> > + return ((1UL << (nr & 31)) & (addr[nr >> 5])) != 0;
> > }
> >
>
> Looks fine. Does your compiler get this right?
Yep, thanks.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH][2.6] constant_test_bit doesn't like my gcc
2003-10-16 4:22 ` Zwane Mwaikambo
@ 2003-10-16 10:55 ` Ingo Oeser
2003-10-16 14:55 ` Andrew Morton
2003-10-16 21:58 ` bill davidsen
0 siblings, 2 replies; 8+ messages in thread
From: Ingo Oeser @ 2003-10-16 10:55 UTC (permalink / raw)
To: Zwane Mwaikambo, Andrew Morton; +Cc: linux-kernel
On Thursday 16 October 2003 06:22, Zwane Mwaikambo wrote:
> On Wed, 15 Oct 2003, Andrew Morton wrote:
> > > static __inline__ int constant_test_bit(int nr, const volatile
> > > unsigned long * addr) {
> > > - return ((1UL << (nr & 31)) & (((const volatile unsigned int *)
> > > addr)[nr >> 5])) != 0; + return ((1UL << (nr & 31)) & (addr[nr >> 5]))
> > > != 0;
> > > }
> >
> > Looks fine. Does your compiler get this right?
>
> Yep, thanks.
Sorry, but I still don't get, what a "const volatile" is supposed to mean.
I would be thankful for an explanation (s. Reply-To).
Regards
Ingo Oeser
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH][2.6] constant_test_bit doesn't like my gcc
2003-10-16 10:55 ` Ingo Oeser
@ 2003-10-16 14:55 ` Andrew Morton
2003-10-16 21:58 ` bill davidsen
1 sibling, 0 replies; 8+ messages in thread
From: Andrew Morton @ 2003-10-16 14:55 UTC (permalink / raw)
To: Ingo Oeser; +Cc: zwane, linux-kernel
Ingo Oeser <ioe-lkml@rameria.de> wrote:
>
> On Thursday 16 October 2003 06:22, Zwane Mwaikambo wrote:
> > On Wed, 15 Oct 2003, Andrew Morton wrote:
> > > > static __inline__ int constant_test_bit(int nr, const volatile
> > > > unsigned long * addr) {
> > > > - return ((1UL << (nr & 31)) & (((const volatile unsigned int *)
> > > > addr)[nr >> 5])) != 0; + return ((1UL << (nr & 31)) & (addr[nr >> 5]))
> > > > != 0;
> > > > }
> > >
> > > Looks fine. Does your compiler get this right?
> >
> > Yep, thanks.
>
> Sorry, but I still don't get, what a "const volatile" is supposed to mean.
const: this function doesn't alter it
volatile: someone else does modify it, so the compiler needs to avoid
caching it in a register.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH][2.6] constant_test_bit doesn't like my gcc
2003-10-16 10:55 ` Ingo Oeser
2003-10-16 14:55 ` Andrew Morton
@ 2003-10-16 21:58 ` bill davidsen
1 sibling, 0 replies; 8+ messages in thread
From: bill davidsen @ 2003-10-16 21:58 UTC (permalink / raw)
To: linux-kernel
In article <200310161255.36380.ioe-lkml@rameria.de>,
Ingo Oeser <ioe-lkml@rameria.de> asked:
| Sorry, but I still don't get, what a "const volatile" is supposed to mean.
It's a volatile which we're not allowed to change.
The ones which I always have to look at multiple times are the "const
pointer to volatile" and "volatile pointer to const." As in unchanging
pointer to a volatile datum, or a volatile pointer to unchanging data.
--
bill davidsen <davidsen@tmr.com>
CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2003-10-16 22:08 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-10-16 3:13 [PATCH][2.6] constant_test_bit doesn't like my gcc Zwane Mwaikambo
2003-10-16 4:10 ` Andrew Morton
2003-10-16 4:12 ` Zwane Mwaikambo
2003-10-16 4:21 ` Andrew Morton
2003-10-16 4:22 ` Zwane Mwaikambo
2003-10-16 10:55 ` Ingo Oeser
2003-10-16 14:55 ` Andrew Morton
2003-10-16 21:58 ` bill davidsen
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.