* [2.6 patch] fix arm26 THREAD_SIZE
@ 2006-01-09 22:22 Adrian Bunk
0 siblings, 0 replies; 3+ messages in thread
From: Adrian Bunk @ 2006-01-09 22:22 UTC (permalink / raw)
To: Andrew Morton; +Cc: spyro, linux-kernel
arm26 currently has a 256 kB THREAD_SIZE (sic).
Looking at the comment in the code, this seems to be based on a
misunderstanding.
The comment says:
this needs attention (see kernel/fork.c which gets a nice div by zero if
this is lower than 8*32768
kernel/fork.c does:
max_threads = mempages / (8 * THREAD_SIZE / PAGE_SIZE)
Therefore, a division by 0 is impossible for all reasonable cases with
THREAD_SIZE >= PAGE_SIZE.
Since the minimum PAGE_SIZE Linux uses on the arm26 architecture is 16k,
PAGE_SIZE should be sufficient for THREAD_SIZE.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Ian Molton <spyro@f2s.com>
---
This patch was already sent on:
- Fri, 6 Jan 2006
--- linux-2.6.15-mm1-full/include/asm-arm26/thread_info.h.old 2006-01-06 16:45:40.000000000 +0100
+++ linux-2.6.15-mm1-full/include/asm-arm26/thread_info.h 2006-01-06 16:46:07.000000000 +0100
@@ -80,8 +80,7 @@
return (struct thread_info *)(sp & ~0x1fff);
}
-/* FIXME - PAGE_SIZE < 32K */
-#define THREAD_SIZE (8*32768) // FIXME - this needs attention (see kernel/fork.c which gets a nice div by zero if this is lower than 8*32768
+#define THREAD_SIZE PAGE_SIZE
#define task_pt_regs(task) ((struct pt_regs *)(task_stack_page(task) + THREAD_SIZE - 8) - 1)
extern struct thread_info *alloc_thread_info(struct task_struct *task);
^ permalink raw reply [flat|nested] 3+ messages in thread
* [2.6 patch] fix arm26 THREAD_SIZE
@ 2006-01-06 16:16 Adrian Bunk
2006-01-07 0:37 ` Ian Molton
0 siblings, 1 reply; 3+ messages in thread
From: Adrian Bunk @ 2006-01-06 16:16 UTC (permalink / raw)
To: spyro; +Cc: linux-kernel
arm26 currently has a 256 kB THREAD_SIZE (sic).
Looking at the comment in the code, this seems to be based on a
misunderstanding.
The comment says:
this needs attention (see kernel/fork.c which gets a nice div by zero if
this is lower than 8*32768
kernel/fork.c does:
max_threads = mempages / (8 * THREAD_SIZE / PAGE_SIZE)
Therefore, a division by 0 is impossible for all reasonable cases with
THREAD_SIZE >= PAGE_SIZE.
Since the minimum PAGE_SIZE Linux uses on the arm26 architecture is 16k,
PAGE_SIZE should be sufficient for THREAD_SIZE.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
--- linux-2.6.15-mm1-full/include/asm-arm26/thread_info.h.old 2006-01-06 16:45:40.000000000 +0100
+++ linux-2.6.15-mm1-full/include/asm-arm26/thread_info.h 2006-01-06 16:46:07.000000000 +0100
@@ -80,8 +80,7 @@
return (struct thread_info *)(sp & ~0x1fff);
}
-/* FIXME - PAGE_SIZE < 32K */
-#define THREAD_SIZE (8*32768) // FIXME - this needs attention (see kernel/fork.c which gets a nice div by zero if this is lower than 8*32768
+#define THREAD_SIZE PAGE_SIZE
#define task_pt_regs(task) ((struct pt_regs *)(task_stack_page(task) + THREAD_SIZE - 8) - 1)
extern struct thread_info *alloc_thread_info(struct task_struct *task);
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [2.6 patch] fix arm26 THREAD_SIZE
2006-01-06 16:16 Adrian Bunk
@ 2006-01-07 0:37 ` Ian Molton
0 siblings, 0 replies; 3+ messages in thread
From: Ian Molton @ 2006-01-07 0:37 UTC (permalink / raw)
To: Adrian Bunk; +Cc: linux-kernel
Adrian Bunk wrote:
Looks good to me. cant recall why that got that way, but the div by zero
definately went away at the time. certainly its wrong as-is, the
implication is that an 8 MB machine would get a max_threads of about 4 :)
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Ian Molton <spyro@f2s.com>
--- linux-2.6.15-mm1-full/include/asm-arm26/thread_info.h.old 2006-01-06
16:45:40.000000000 +0100
+++ linux-2.6.15-mm1-full/include/asm-arm26/thread_info.h 2006-01-06
16:46:07.000000000 +0100
@@ -80,8 +80,7 @@
return (struct thread_info *)(sp & ~0x1fff);
}
-/* FIXME - PAGE_SIZE < 32K */
-#define THREAD_SIZE (8*32768) // FIXME - this needs attention (see
kernel/fork.c which gets a nice div by zero if this is lower than 8*32768
+#define THREAD_SIZE PAGE_SIZE
#define task_pt_regs(task) ((struct pt_regs *)(task_stack_page(task) +
THREAD_SIZE - 8) - 1)
extern struct thread_info *alloc_thread_info(struct task_struct *task);
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2006-01-09 22:22 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-01-09 22:22 [2.6 patch] fix arm26 THREAD_SIZE Adrian Bunk
-- strict thread matches above, loose matches on Subject: below --
2006-01-06 16:16 Adrian Bunk
2006-01-07 0:37 ` Ian Molton
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox