* Re: 2.6.12-rc4-mm2
2005-05-16 9:13 2.6.12-rc4-mm2 Andrew Morton
@ 2005-05-16 9:25 ` Russell King
2005-05-16 10:50 ` 2.6.12-rc4-mm2 Danny ter Haar
` (9 subsequent siblings)
10 siblings, 0 replies; 32+ messages in thread
From: Russell King @ 2005-05-16 9:25 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
On Mon, May 16, 2005 at 02:13:02AM -0700, Andrew Morton wrote:
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc4/2.6.12-rc4-mm2/
If you wish, you could add the three git trees I have on m.k.o into
the mix - ~rmk/linux-2.6-*.git
--
Russell King
Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/
maintainer of: 2.6 Serial core
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 9:13 2.6.12-rc4-mm2 Andrew Morton
2005-05-16 9:25 ` 2.6.12-rc4-mm2 Russell King
@ 2005-05-16 10:50 ` Danny ter Haar
2005-05-16 11:17 ` 2.6.12-rc4-mm2 Alexey Dobriyan
2005-05-16 12:30 ` 2.6.12-rc4-mm2 Brice Goglin
` (8 subsequent siblings)
10 siblings, 1 reply; 32+ messages in thread
From: Danny ter Haar @ 2005-05-16 10:50 UTC (permalink / raw)
To: linux-kernel
Andrew Morton <akpm@osdl.org> wrote:
>ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc4/2.6.12-rc4-mm2/
include/acpi/achware.h:159: warning: `struct acpi_gpe_block_info' declared inside parameter list
include/acpi/achware.h:159: warning: `struct acpi_gpe_xrupt_info' declared inside parameter list
include/acpi/achware.h:159: warning: type defaults to `int' in declaration of `acpi_hw_enable_runtime_gpe_block'
include/acpi/achware.h:159: warning: data definition has no type or storage class
make[2]: *** [arch/x86_64/kernel/time.o] Error 1
make[1]: *** [arch/x86_64/kernel] Error 2
make[1]: Leaving directory `/usr/src/linux-2.6.12-rc4-mm2'
make: *** [stamp-build] Error 2
Debian-amd64
Kernel config: http://newsgate.newsserver.nl/kernel/dotconfig-2.6.12-rc4-mm2
Danny
--
The foundation of evil is made up of lies and marketing - UF2004
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 10:50 ` 2.6.12-rc4-mm2 Danny ter Haar
@ 2005-05-16 11:17 ` Alexey Dobriyan
2005-05-16 11:38 ` 2.6.12-rc4-mm2 Danny ter Haar
0 siblings, 1 reply; 32+ messages in thread
From: Alexey Dobriyan @ 2005-05-16 11:17 UTC (permalink / raw)
To: Danny ter Haar; +Cc: linux-kernel, Andrew Morton
On Monday 16 May 2005 14:50, Danny ter Haar wrote:
> Andrew Morton <akpm@osdl.org> wrote:
> >ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc4/2.6.12-rc4-mm2/
>
> include/acpi/achware.h:159: warning: `struct acpi_gpe_block_info' declared inside parameter list
> include/acpi/achware.h:159: warning: `struct acpi_gpe_xrupt_info' declared inside parameter list
> include/acpi/achware.h:159: warning: type defaults to `int' in declaration of `acpi_hw_enable_runtime_gpe_block'
> include/acpi/achware.h:159: warning: data definition has no type or storage class
> make[2]: *** [arch/x86_64/kernel/time.o] Error 1
Does this help?
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
--- linux-2.6.12-rc4-mm2/include/acpi/achware.h 2005-05-16 14:24:02.000000000 +0400
+++ linux-2.6.12-rc4-mm2-acpi/include/acpi/achware.h 2005-05-16 15:11:39.000000000 +0400
@@ -44,6 +44,14 @@
#ifndef __ACHWARE_H__
#define __ACHWARE_H__
+#include <linux/types.h>
+#include <acpi/actypes.h>
+
+struct acpi_bit_register_info;
+struct acpi_generic_address;
+struct acpi_gpe_event_info;
+struct acpi_gpe_xrupt_info;
+struct acpi_gpe_block_info;
/* PM Timer ticks per second (HZ) */
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 11:17 ` 2.6.12-rc4-mm2 Alexey Dobriyan
@ 2005-05-16 11:38 ` Danny ter Haar
2005-05-16 12:15 ` 2.6.12-rc4-mm2 Alexey Dobriyan
0 siblings, 1 reply; 32+ messages in thread
From: Danny ter Haar @ 2005-05-16 11:38 UTC (permalink / raw)
To: linux-kernel
Alexey Dobriyan <adobriyan@gmail.com> wrote:
>Does this help?
>--- linux-2.6.12-rc4-mm2/include/acpi/achware.h 2005-05-16 14:24:02.000000000 +0400
>+++ linux-2.6.12-rc4-mm2-acpi/include/acpi/achware.h 2005-05-16 15:11:39.000000000 +0400
[SNIP]
[PATCH APPLIED]
Nope, (unfortunatly)
Complete error is at:
http://newsgate.newsserver.nl/kernel/2.6.12-rc4-mm2-error-out.txt
Danny
--
The foundation of evil is made up of lies and marketing - UF2004
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 11:38 ` 2.6.12-rc4-mm2 Danny ter Haar
@ 2005-05-16 12:15 ` Alexey Dobriyan
2005-05-16 17:11 ` 2.6.12-rc4-mm2 Danny ter Haar
0 siblings, 1 reply; 32+ messages in thread
From: Alexey Dobriyan @ 2005-05-16 12:15 UTC (permalink / raw)
To: Danny ter Haar; +Cc: linux-kernel, Andrew Morton
On Monday 16 May 2005 15:38, Danny ter Haar wrote:
> Alexey Dobriyan <adobriyan@gmail.com> wrote:
> >Does this help?
> >--- linux-2.6.12-rc4-mm2/include/acpi/achware.h
> >+++ linux-2.6.12-rc4-mm2-acpi/include/acpi/achware.h
> Nope, (unfortunatly)
Please, try this.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
--- linux-2.6.12-rc4-mm2/include/acpi/achware.h 2005-05-16 14:24:02.000000000 +0400
+++ linux-2.6.12-rc4-mm2-acpi/include/acpi/achware.h 2005-05-16 16:05:41.000000000 +0400
@@ -44,6 +44,8 @@
#ifndef __ACHWARE_H__
#define __ACHWARE_H__
+struct acpi_gpe_xrupt_info;
+struct acpi_gpe_block_info;
/* PM Timer ticks per second (HZ) */
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 12:15 ` 2.6.12-rc4-mm2 Alexey Dobriyan
@ 2005-05-16 17:11 ` Danny ter Haar
2005-05-16 17:43 ` 2.6.12-rc4-mm2 Alexey Dobriyan
0 siblings, 1 reply; 32+ messages in thread
From: Danny ter Haar @ 2005-05-16 17:11 UTC (permalink / raw)
To: linux-kernel
Alexey Dobriyan <adobriyan@gmail.com> wrote:
>> Alexey Dobriyan <adobriyan@gmail.com> wrote:
>> >Does this help?
>> Nope, (unfortunatly)
>Please, try this.
[Patch #2 applied]
Still not succesful..
Error is at
http://newsgate.newsserver.nl/kernel/2.6.12-rc4-mm2-patch%232-error-out.txt
btw:
newsgate:/usr/src/linux-2.6.12-rc4-mm2# gcc -v
Reading specs from /usr/lib/gcc-lib/x86_64-linux/3.3.6/specs
Configured with: ../src/configure -v
--enable-languages=c,c++,java,f77,pascal,objc,ada,treelang
--prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info
--with-gxx-include-dir=/usr/include/c++/3.3 --enable-shared
--enable-__cxa_atexit --with-system-zlib --enable-nls
--without-included-gettext --enable-clocale=gnu --enable-debug
--enable-java-gc=boehm --enable-java-awt=xlib --enable-objc-gc
--disable-multilib x86_64-linux
Thread model: posix
gcc version 3.3.6 (Debian 1:3.3.6-4)
newsgate:/var/www/kernel# ls -lrt
-rw-r--r-- 1 root root 3638 May 16 13:36 2.6.12-rc4-mm2-error-out.txt
-rw-r--r-- 1 root root 5389 May 16 19:07 2.6.12-rc4-mm2-patch#2-error-out.txt
Thanks for your efforts to help me!
Danny
--
The foundation of evil is made up of lies and marketing - UF2004
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 17:11 ` 2.6.12-rc4-mm2 Danny ter Haar
@ 2005-05-16 17:43 ` Alexey Dobriyan
2005-05-16 19:30 ` 2.6.12-rc4-mm2 Danny ter Haar
0 siblings, 1 reply; 32+ messages in thread
From: Alexey Dobriyan @ 2005-05-16 17:43 UTC (permalink / raw)
To: Danny ter Haar; +Cc: linux-kernel
On Monday 16 May 2005 21:11, Danny ter Haar wrote:
> Alexey Dobriyan <adobriyan@gmail.com> wrote:
> >> Alexey Dobriyan <adobriyan@gmail.com> wrote:
> >> >Does this help?
> >> Nope, (unfortunatly)
> >Please, try this.
>
> [Patch #2 applied]
>
> Still not succesful..
>
> Error is at
> http://newsgate.newsserver.nl/kernel/2.6.12-rc4-mm2-patch%232-error-out.txt
Urgh... ACPI uses catch-all header file.
If this won't work, I'll get a cross-compiler.
--- linux-2.6.12-rc4-mm2/arch/x86_64/kernel/time.c 2005-05-16 21:38:04.000000000 +0400
+++ linux-2.6.12-rc4-mm2-acpi/arch/x86_64/kernel/time.c 2005-05-16 21:38:49.000000000 +0400
@@ -27,7 +27,7 @@
#include <linux/bcd.h>
#include <linux/kallsyms.h>
#include <linux/acpi.h>
-#include <acpi/achware.h> /* for PM timer frequency */
+#include <acpi/acpi.h>
#include <asm/8253pit.h>
#include <asm/pgtable.h>
#include <asm/vsyscall.h>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 17:43 ` 2.6.12-rc4-mm2 Alexey Dobriyan
@ 2005-05-16 19:30 ` Danny ter Haar
0 siblings, 0 replies; 32+ messages in thread
From: Danny ter Haar @ 2005-05-16 19:30 UTC (permalink / raw)
To: linux-kernel
Alexey Dobriyan <adobriyan@gmail.com> wrote:
>Urgh... ACPI uses catch-all header file.
>If this won't work, I'll get a cross-compiler.
take #3
still no luck.
Less output then previous trial.
Danny
PS: need an account so you can compile on the machine in question ?
--
The foundation of evil is made up of lies and marketing - UF2004
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 9:13 2.6.12-rc4-mm2 Andrew Morton
2005-05-16 9:25 ` 2.6.12-rc4-mm2 Russell King
2005-05-16 10:50 ` 2.6.12-rc4-mm2 Danny ter Haar
@ 2005-05-16 12:30 ` Brice Goglin
2005-05-16 17:46 ` 2.6.12-rc4-mm2, alpha and mips broke Jan Dittmer
` (7 subsequent siblings)
10 siblings, 0 replies; 32+ messages in thread
From: Brice Goglin @ 2005-05-16 12:30 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
[-- Attachment #1: Type: text/plain, Size: 1247 bytes --]
Andrew Morton a écrit :
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc4/2.6.12-rc4-mm2/
>
>
> - davem has set up a mm-commits mailing list so people can review things
> which are added to or removed from the -mm tree. Do
>
> echo subscribe mm-commits | mail majordomo@vger.kernel.org
>
> - x86_64 architecture update from Andi.
>
> - Everything up to and including `spurious-interrupt-fix.patch' is planned
> for 2.6.12 merging. Plus a few other things in there.
>
> - Another DVB subsystem update
Hi Andrew,
CONFIG_PPP_MPPE can be enabled without CONFIG_CRYPTO.
This results in this warning when running make modules_install:
if [ -r System.map -a -x /sbin/depmod ]; then /sbin/depmod -ae -F
System.map 2.6.12-rc4-mm2=LoulousMobile; fi
WARNING:
/lib/modules/2.6.12-rc4-mm2=LoulousMobile/kernel/drivers/net/ppp_mppe.ko
needs unknown symbol crypto_alloc_tfm
WARNING:
/lib/modules/2.6.12-rc4-mm2=LoulousMobile/kernel/drivers/net/ppp_mppe.ko
needs unknown symbol crypto_free_tfm
By the way, looking at drivers/net/ppp_mppe.c, it looks like sha1 and
arc4 are needed at runtime.
The attached patch selects all these when PPP_MPPE is selected.
Signed-off-by: Brice Goglin <Brice.Goglin@ens-lyon.org>
Brice
[-- Attachment #2: fix_ppp-mppe_dependencies.patch --]
[-- Type: text/x-patch, Size: 521 bytes --]
--- linux-mm/drivers/net/Kconfig.old 2005-05-16 14:15:29.000000000 +0200
+++ linux-mm/drivers/net/Kconfig 2005-05-16 14:24:44.000000000 +0200
@@ -2431,6 +2431,9 @@ config PPP_BSDCOMP
config PPP_MPPE
tristate "PPP MPPE compression (encryption) (EXPERIMENTAL)"
depends on PPP && EXPERIMENTAL
+ select CRYPTO
+ select CRYPTO_SHA1
+ select CRYPTO_ARC4
---help---
Support for the MPPE Encryption protocol, as employed by the
Microsoft Point-to-Point Tunneling Protocol.
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2, alpha and mips broke
2005-05-16 9:13 2.6.12-rc4-mm2 Andrew Morton
` (2 preceding siblings ...)
2005-05-16 12:30 ` 2.6.12-rc4-mm2 Brice Goglin
@ 2005-05-16 17:46 ` Jan Dittmer
2005-05-16 20:09 ` Andrew Morton
2005-05-16 19:18 ` 2.6.12-rc4-mm2: proc-pid-smaps.patch broke nommu Adrian Bunk
` (6 subsequent siblings)
10 siblings, 1 reply; 32+ messages in thread
From: Jan Dittmer @ 2005-05-16 17:46 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
Comparing 2.6.12-rc4-mm1 to 2.6.12-rc4-mm2 (defconfig)
======================================================
- alpha: broke
AR arch/alpha/lib/lib.a
GEN .version
CHK include/linux/compile.h
UPD include/linux/compile.h
CC init/version.o
LD init/built-in.o
LD .tmp_vmlinux1
mm/built-in.o(.text+0xe79c):/usr/src/ctest/mm/kernel/mm/slab.c:339: undefined reference to `__bad_size'
mm/built-in.o(.text+0xe7a0):/usr/src/ctest/mm/kernel/mm/slab.c:339: undefined reference to `__bad_size'
make[1]: *** [.tmp_vmlinux1] Error 1
make: *** [_all] Error 2
Details: http://l4x.org/k/?d=3741
- mips: broke
CC mm/mempool.o
CC mm/oom_kill.o
CC mm/fadvise.o
CC mm/page_alloc.o
CC mm/page-writeback.o
CC mm/pdflush.o
CC mm/readahead.o
CC mm/slab.o
mm/slab.c:117:2: #error "Broken Configuration: CONFIG_NUMA not set but MAX_NUMNODES !=1 !!"
make[1]: *** [mm/slab.o] Error 1
make: *** [mm] Error 2
Details: http://l4x.org/k/?d=3753
--
Jan
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2, alpha and mips broke
2005-05-16 17:46 ` 2.6.12-rc4-mm2, alpha and mips broke Jan Dittmer
@ 2005-05-16 20:09 ` Andrew Morton
0 siblings, 0 replies; 32+ messages in thread
From: Andrew Morton @ 2005-05-16 20:09 UTC (permalink / raw)
To: Jan Dittmer; +Cc: linux-kernel
Jan Dittmer <jdittmer@ppp0.net> wrote:
>
> Comparing 2.6.12-rc4-mm1 to 2.6.12-rc4-mm2 (defconfig)
> ======================================================
>
> - alpha: broke
> AR arch/alpha/lib/lib.a
> GEN .version
> CHK include/linux/compile.h
> UPD include/linux/compile.h
> CC init/version.o
> LD init/built-in.o
> LD .tmp_vmlinux1
> mm/built-in.o(.text+0xe79c):/usr/src/ctest/mm/kernel/mm/slab.c:339: undefined reference to `__bad_size'
> mm/built-in.o(.text+0xe7a0):/usr/src/ctest/mm/kernel/mm/slab.c:339: undefined reference to `__bad_size'
> make[1]: *** [.tmp_vmlinux1] Error 1
> make: *** [_all] Error 2
>
argh, I forgot to add the patch to the series file, sorry.
From: Andrew Morton <akpm@osdl.org>
This doesn't work (on alpha, at least).
It's not inside __builtin_constant_p() so the compiler cannot be sure that all
the possible sizes have been checked for.
Cc: Christoph Lameter <clameter@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
---
mm/slab.c | 12 +++++++-----
1 files changed, 7 insertions(+), 5 deletions(-)
diff -puN mm/slab.c~numa-aware-slab-allocator-v3-__bad_size-fix mm/slab.c
--- 25-alpha/mm/slab.c~numa-aware-slab-allocator-v3-__bad_size-fix 2005-05-15 22:25:33.000000000 -0700
+++ 25-alpha-akpm/mm/slab.c 2005-05-15 22:30:10.000000000 -0700
@@ -325,7 +325,8 @@ struct kmem_list3 __initdata initkmem_li
*/
static inline int index_of(const size_t size)
{
- int i = 0;
+ if (__builtin_constant_p(size)) {
+ int i = 0;
#define CACHE(x) \
if (size <=x) \
@@ -334,11 +335,12 @@ static inline int index_of(const size_t
i++;
#include "linux/kmalloc_sizes.h"
#undef CACHE
- {
- extern void __bad_size(void);
- __bad_size();
- return 0;
+ {
+ extern void __bad_size(void);
+ __bad_size();
+ }
}
+ return 0;
}
#define INDEX_AC index_of(sizeof(struct array_cache))
_
>
> - mips: broke
> CC mm/mempool.o
> CC mm/oom_kill.o
> CC mm/fadvise.o
> CC mm/page_alloc.o
> CC mm/page-writeback.o
> CC mm/pdflush.o
> CC mm/readahead.o
> CC mm/slab.o
> mm/slab.c:117:2: #error "Broken Configuration: CONFIG_NUMA not set but MAX_NUMNODES !=1 !!"
> make[1]: *** [mm/slab.o] Error 1
> make: *** [mm] Error 2
>
OK, Christoph is scratching his head over that one.
^ permalink raw reply [flat|nested] 32+ messages in thread
* 2.6.12-rc4-mm2: proc-pid-smaps.patch broke nommu
2005-05-16 9:13 2.6.12-rc4-mm2 Andrew Morton
` (3 preceding siblings ...)
2005-05-16 17:46 ` 2.6.12-rc4-mm2, alpha and mips broke Jan Dittmer
@ 2005-05-16 19:18 ` Adrian Bunk
2005-05-21 2:19 ` Mauricio Lin
2005-05-17 9:06 ` 2.6.12-rc4-mm2 Brice Goglin
` (5 subsequent siblings)
10 siblings, 1 reply; 32+ messages in thread
From: Adrian Bunk @ 2005-05-16 19:18 UTC (permalink / raw)
To: Andrew Morton, Mauricio Lin; +Cc: linux-kernel
It seems proc-pid-smaps.patch is guilty for this nommu breakage in -mm:
<-- snip -->
...
LD vmlinux
fs/built-in.o(.text+0x32b08): In function `smaps_open':
/usr/src/ctest/mm/kernel/fs/proc/base.c:560: undefined reference to `_proc_pid_smaps_op'
make[1]: *** [vmlinux] Error 1
<-- snip -->
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2: proc-pid-smaps.patch broke nommu
2005-05-16 19:18 ` 2.6.12-rc4-mm2: proc-pid-smaps.patch broke nommu Adrian Bunk
@ 2005-05-21 2:19 ` Mauricio Lin
2005-05-21 2:39 ` Mauricio Lin
2005-07-21 15:04 ` Adrian Bunk
0 siblings, 2 replies; 32+ messages in thread
From: Mauricio Lin @ 2005-05-21 2:19 UTC (permalink / raw)
To: Adrian Bunk; +Cc: Andrew Morton, linux-kernel
Hi Adrian,
How did you get this error? What is your configuration?
I would like to replicate it.
BR,
Mauricio Lin.
On 5/16/05, Adrian Bunk <bunk@stusta.de> wrote:
> It seems proc-pid-smaps.patch is guilty for this nommu breakage in -mm:
>
> <-- snip -->
>
> ...
> LD vmlinux
> fs/built-in.o(.text+0x32b08): In function `smaps_open':
> /usr/src/ctest/mm/kernel/fs/proc/base.c:560: undefined reference to `_proc_pid_smaps_op'
> make[1]: *** [vmlinux] Error 1
>
> <-- snip -->
>
> cu
> Adrian
>
> --
>
> "Is there not promise of rain?" Ling Tan asked suddenly out
> of the darkness. There had been need of rain for many days.
> "Only a promise," Lao Er said.
> Pearl S. Buck - Dragon Seed
>
>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2: proc-pid-smaps.patch broke nommu
2005-05-21 2:19 ` Mauricio Lin
@ 2005-05-21 2:39 ` Mauricio Lin
2005-07-21 15:04 ` Adrian Bunk
1 sibling, 0 replies; 32+ messages in thread
From: Mauricio Lin @ 2005-05-21 2:39 UTC (permalink / raw)
To: Adrian Bunk; +Cc: Andrew Morton, linux-kernel
Hi Adrian,
I guess that the code below you have to include in the
fs/proc/task_nommu.c. But I would like to replicate the error you
found and apply the code below to verify if it is correct. So if you
can explain how you got this error I can test the code below. If you
want to test the code below by yourself, let me know about the result
please.
static int show_smap(struct seq_file *m, void *v)
{
return 0;
}
struct seq_operations proc_pid_smaps_op = {
.start = m_start,
.next = m_next,
.stop = m_stop,
.show = show_smap
};
BR,
Mauricio Lin.
On 5/20/05, Mauricio Lin <mauriciolin@gmail.com> wrote:
> Hi Adrian,
>
> How did you get this error? What is your configuration?
>
> I would like to replicate it.
>
> BR,
>
> Mauricio Lin.
>
> On 5/16/05, Adrian Bunk <bunk@stusta.de> wrote:
> > It seems proc-pid-smaps.patch is guilty for this nommu breakage in -mm:
> >
> > <-- snip -->
> >
> > ...
> > LD vmlinux
> > fs/built-in.o(.text+0x32b08): In function `smaps_open':
> > /usr/src/ctest/mm/kernel/fs/proc/base.c:560: undefined reference to `_proc_pid_smaps_op'
> > make[1]: *** [vmlinux] Error 1
> >
> > <-- snip -->
> >
> > cu
> > Adrian
> >
> > --
> >
> > "Is there not promise of rain?" Ling Tan asked suddenly out
> > of the darkness. There had been need of rain for many days.
> > "Only a promise," Lao Er said.
> > Pearl S. Buck - Dragon Seed
> >
> >
>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2: proc-pid-smaps.patch broke nommu
2005-05-21 2:19 ` Mauricio Lin
2005-05-21 2:39 ` Mauricio Lin
@ 2005-07-21 15:04 ` Adrian Bunk
1 sibling, 0 replies; 32+ messages in thread
From: Adrian Bunk @ 2005-07-21 15:04 UTC (permalink / raw)
To: Mauricio Lin; +Cc: Andrew Morton, linux-kernel
On Fri, May 20, 2005 at 10:19:11PM -0400, Mauricio Lin wrote:
> Hi Adrian,
Hi Mauricio,
first of all sorry for the late answer.
> How did you get this error? What is your configuration?
>
> I would like to replicate it.
I've seen then at the logs of the h8300 failures listed at [1].
> BR,
>
> Mauricio Lin.
cu
Adrian
[1] http://l4x.org/k/
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 9:13 2.6.12-rc4-mm2 Andrew Morton
` (4 preceding siblings ...)
2005-05-16 19:18 ` 2.6.12-rc4-mm2: proc-pid-smaps.patch broke nommu Adrian Bunk
@ 2005-05-17 9:06 ` Brice Goglin
2005-05-17 16:38 ` 2.6.12-rc4-mm2 Richard Purdie
2005-05-18 7:14 ` 2.6.12-rc4-mm2 Coywolf Qi Hunt
` (4 subsequent siblings)
10 siblings, 1 reply; 32+ messages in thread
From: Brice Goglin @ 2005-05-17 9:06 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
Andrew Morton a écrit :
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc4/2.6.12-rc4-mm2/
Hi Andrew,
Cardmgr does not automatically start my pcmcia wireless card anymore.
orinoco modules are not loaded at all.
I still can modprobe orinoco_cs to get my wireless to work.
Cardmgr says this when starting:
cardmgr[27367]: no pcmcia driver in /proc/devices
Is this a feature related to the upcoming deprecation of cardctl ?
Am I supposed to use pcmcia-utils ?
Thanks,
Brice
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-17 9:06 ` 2.6.12-rc4-mm2 Brice Goglin
@ 2005-05-17 16:38 ` Richard Purdie
2005-05-18 22:45 ` 2.6.12-rc4-mm2 Richard Purdie
0 siblings, 1 reply; 32+ messages in thread
From: Richard Purdie @ 2005-05-17 16:38 UTC (permalink / raw)
To: Brice Goglin, Andrew Morton; +Cc: linux-kernel
Brice Goglin:
> Cardmgr does not automatically start my pcmcia wireless card anymore.
> orinoco modules are not loaded at all.
> I still can modprobe orinoco_cs to get my wireless to work.
>
> Cardmgr says this when starting:
> cardmgr[27367]: no pcmcia driver in /proc/devices
>
> Is this a feature related to the upcoming deprecation of cardctl ?
> Am I supposed to use pcmcia-utils ?
I also see the above message on the arm pxa zaurus with -mm2. I'm still
using pcmcia-cs rather than pcmcia-utils. pcmcia+cardmgr works fine in -mm1.
I'm also not sure if this is by design or not...
Regards,
Richard
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-17 16:38 ` 2.6.12-rc4-mm2 Richard Purdie
@ 2005-05-18 22:45 ` Richard Purdie
0 siblings, 0 replies; 32+ messages in thread
From: Richard Purdie @ 2005-05-18 22:45 UTC (permalink / raw)
To: Brice Goglin, Andrew Morton; +Cc: linux-kernel
On Tuesday 17 May 2005 5:38 pm, Richard Purdie wrote:
> Brice Goglin:
> > Cardmgr does not automatically start my pcmcia wireless card anymore.
> > orinoco modules are not loaded at all.
> > I still can modprobe orinoco_cs to get my wireless to work.
> >
> > Cardmgr says this when starting:
> > cardmgr[27367]: no pcmcia driver in /proc/devices
> >
> > Is this a feature related to the upcoming deprecation of cardctl ?
> > Am I supposed to use pcmcia-utils ?
>
> I also see the above message on the arm pxa zaurus with -mm2. I'm still
> using pcmcia-cs rather than pcmcia-utils. pcmcia+cardmgr works fine in
> -mm1. I'm also not sure if this is by design or not...
I found the problem. The pcmcia-move-pcmcia-ioctl-to-a-separate-file patch was
corrupted in -mm2 causing this problem. The fix is below.
Richard
--- linux-2.6.11.orig/drivers/pcmcia/ds.c 2005-05-18 23:27:43.000000000 +0100
+++ linux-2.6.11/drivers/pcmcia/ds.c 2005-05-17 17:13:50.000000000 +0100
@@ -1199,6 +1199,9 @@
bus_register(&pcmcia_bus_type);
class_interface_register(&pcmcia_bus_interface);
+
+ pcmcia_setup_ioctl();
+
return 0;
}
fs_initcall(init_pcmcia_bus); /* one level after subsys_initcall so that
@@ -1212,7 +1215,6 @@
class_interface_unregister(&pcmcia_bus_interface);
bus_unregister(&pcmcia_bus_type);
- pcmcia_setup_ioctl();
}
module_exit(exit_pcmcia_bus);
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 9:13 2.6.12-rc4-mm2 Andrew Morton
` (5 preceding siblings ...)
2005-05-17 9:06 ` 2.6.12-rc4-mm2 Brice Goglin
@ 2005-05-18 7:14 ` Coywolf Qi Hunt
2005-05-18 20:26 ` 2.6.12-rc4-mm2 Alexander Nyberg
` (3 subsequent siblings)
10 siblings, 0 replies; 32+ messages in thread
From: Coywolf Qi Hunt @ 2005-05-18 7:14 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
On 5/16/05, Andrew Morton <akpm@osdl.org> wrote:
>
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc4/2.6.12-rc4-mm2/
>
> - davem has set up a mm-commits mailing list so people can review things
> which are added to or removed from the -mm tree. Do
>
> echo subscribe mm-commits | mail majordomo@vger.kernel.org
I see patches being added to and removed from -mm tree. Is it
possible to know the reason why they are removed, whether they are
merged upstream or dropped?
--
Coywolf Qi Hunt
http://sosdg.org/~coywolf/
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 9:13 2.6.12-rc4-mm2 Andrew Morton
` (6 preceding siblings ...)
2005-05-18 7:14 ` 2.6.12-rc4-mm2 Coywolf Qi Hunt
@ 2005-05-18 20:26 ` Alexander Nyberg
2005-05-19 14:59 ` 2.6.12-rc4-mm2 Brice Goglin
` (2 subsequent siblings)
10 siblings, 0 replies; 32+ messages in thread
From: Alexander Nyberg @ 2005-05-18 20:26 UTC (permalink / raw)
To: Andrew Morton; +Cc: acpi-devel, linux-kernel
mån 2005-05-16 klockan 02:13 -0700 skrev Andrew Morton:
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc4/2.6.12-rc4-mm2/
>
>
> - davem has set up a mm-commits mailing list so people can review things
> which are added to or removed from the -mm tree. Do
>
> echo subscribe mm-commits | mail majordomo@vger.kernel.org
>
> - x86_64 architecture update from Andi.
>
> - Everything up to and including `spurious-interrupt-fix.patch' is planned
> for 2.6.12 merging. Plus a few other things in there.
>
> - Another DVB subsystem update
>
I need this to be able to compile on x64 without CONFIG_ACPI (yeah it's
probably wrong all the way, but can someone that knows acpi fix it
please?). Tested on i386 & x64 w & w/o CONFIG_ACPI.
Index: kexec/include/linux/acpi.h
===================================================================
--- kexec.orig/include/linux/acpi.h 2005-05-18 22:04:11.000000000 +0200
+++ kexec/include/linux/acpi.h 2005-05-18 22:13:13.000000000 +0200
@@ -25,8 +25,6 @@
#ifndef _LINUX_ACPI_H
#define _LINUX_ACPI_H
-#ifdef CONFIG_ACPI
-
#ifndef _LINUX
#define _LINUX
#endif
@@ -419,6 +417,8 @@
#else /*!CONFIG_ACPI_BOOT*/
+static inline int acpi_boot_init (void) { return 0; }
+static inline int acpi_boot_table_init (void) { return 0; }
#define acpi_mp_config 0
#endif /*!CONFIG_ACPI_BOOT*/
@@ -531,18 +531,4 @@
extern int pnpacpi_disabled;
-#else /* CONFIG_ACPI */
-
-static inline int acpi_boot_init(void)
-{
- return 0;
-}
-
-static inline int acpi_boot_table_init(void)
-{
- return 0;
-}
-
-#endif /* CONFIG_ACPI */
-
#endif /* _LINUX_ACPI_H */
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: 2.6.12-rc4-mm2
2005-05-16 9:13 2.6.12-rc4-mm2 Andrew Morton
` (7 preceding siblings ...)
2005-05-18 20:26 ` 2.6.12-rc4-mm2 Alexander Nyberg
@ 2005-05-19 14:59 ` Brice Goglin
2005-05-22 21:27 ` [bugfix] try_to_unmap_cluster() passes out-of-bounds pte to pte_unmap() William Lee Irwin III
2007-06-27 0:35 ` Problems with fb console [was Re: 2.6.12-rc4-mm2] J.A. Magallón
10 siblings, 0 replies; 32+ messages in thread
From: Brice Goglin @ 2005-05-19 14:59 UTC (permalink / raw)
To: Andrew Morton, Dominik Brodowski; +Cc: linux-kernel
Andrew Morton a écrit :
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc4/2.6.12-rc4-mm2/
>
>
> - davem has set up a mm-commits mailing list so people can review things
> which are added to or removed from the -mm tree. Do
>
> echo subscribe mm-commits | mail majordomo@vger.kernel.org
>
> - x86_64 architecture update from Andi.
>
> - Everything up to and including `spurious-interrupt-fix.patch' is planned
> for 2.6.12 merging. Plus a few other things in there.
>
> - Another DVB subsystem update
Hi Andrew and Dominik,
Since mm2, udev cannot rename my pcmcia wireless interface.
rc4 and rc4-mm1 successfully rename it from eth0 to wifi.
rc4-mm2 only renames the internal interface, not the pcmcia wifi one.
When I insert the card, This line appears in syslog:
udev[9500]: configured rule in '/etc/udev/rules.d/brice.rules[4]'
applied, 'eth0' becomes 'wifi'
But the interface is still called eth0.
Any idea ?
Thanks,
Brice
PS: Richard Purdie's patch (http://lkml.org/lkml/2005/5/18/303) fixed my
previous cardctl breakage, but it doesn't change anything regarding this
bug.
^ permalink raw reply [flat|nested] 32+ messages in thread
* [bugfix] try_to_unmap_cluster() passes out-of-bounds pte to pte_unmap()
2005-05-16 9:13 2.6.12-rc4-mm2 Andrew Morton
` (8 preceding siblings ...)
2005-05-19 14:59 ` 2.6.12-rc4-mm2 Brice Goglin
@ 2005-05-22 21:27 ` William Lee Irwin III
2005-05-22 22:00 ` Andrew Morton
2005-05-24 0:14 ` Andrew Morton
2007-06-27 0:35 ` Problems with fb console [was Re: 2.6.12-rc4-mm2] J.A. Magallón
10 siblings, 2 replies; 32+ messages in thread
From: William Lee Irwin III @ 2005-05-22 21:27 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel, gregkh
[-- Attachment #1: Type: text/plain, Size: 1397 bytes --]
On Mon, May 16, 2005 at 02:13:02AM -0700, Andrew Morton wrote:
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc4/2.6.12-rc4-mm2/
> - davem has set up a mm-commits mailing list so people can review things
> which are added to or removed from the -mm tree. Do
> echo subscribe mm-commits | mail majordomo@vger.kernel.org
> - x86_64 architecture update from Andi.
> - Everything up to and including `spurious-interrupt-fix.patch' is planned
> for 2.6.12 merging. Plus a few other things in there.
> - Another DVB subsystem update
try_to_unmap_cluster() does:
for (pte = pte_offset_map(pmd, address);
address < end; pte++, address += PAGE_SIZE) {
...
}
pte_unmap(pte);
It may take a little staring to notice, but pte can actually fall off
the end of the pte page in this iteration, which makes life difficult
for kmap_atomic() and the users not expecting it to BUG(). Of course,
we're somewhat lucky in that arithmetic elsewhere in the function
guarantees that at least one iteration is made, lest this force larger
rearrangements to be made. This issue and patch also apply to non-mm
mainline and with trivial adjustments, at least two related kernels.
Discovered during internal testing at Oracle. Sample BUG() message
included along with patch as a MIME attachment.
Signed-off-by: William Irwin <wli@holomorphy.com>
-- wli
[-- Attachment #2: vlm-kunmap-atomic-fix.patch --]
[-- Type: text/plain, Size: 244 bytes --]
--- ./mm/rmap.c.orig 2005-05-20 01:29:14.066467151 -0700
+++ ./mm/rmap.c 2005-05-20 01:30:06.620649901 -0700
@@ -694,7 +694,7 @@
(*mapcount)--;
}
- pte_unmap(pte);
+ pte_unmap(pte-1);
out_unlock:
spin_unlock(&mm->page_table_lock);
}
[-- Attachment #3: vlm.oops --]
[-- Type: text/plain, Size: 1890 bytes --]
May 18 23:50:30 palnx1 kernel: ------------[ cut here ]------------
May 18 23:50:30 palnx1 kernel: kernel BUG at arch/i386/mm/highmem.c:96!
May 18 23:50:30 palnx1 kernel: invalid operand: 0000 [#1]
May 18 23:50:30 palnx1 kernel: SMP
May 18 23:50:30 palnx1 kernel: Modules linked in: nfsd exportfs md5 ipv6 parport_pc lp parport autofs4 i2c_dev i2c_core nfs lockd sunrpc dm_mod button battery ac uhci_hcd e1000 e100 mii floppy ext3 jbd qla2300 qla2xxx scsi_transport_fc aic79xx sd_mod scsi_mod
May 18 23:50:30 palnx1 kernel: CPU: 2
May 18 23:50:30 palnx1 kernel: EIP: 0060:[<c011bfc4>] Not tainted VLI
May 18 23:50:30 palnx1 kernel: EFLAGS: 00010206 (2.6.9-9.ELsmp)
May 18 23:50:30 palnx1 kernel: EIP is at kunmap_atomic+0x2e/0x58
May 18 23:50:30 palnx1 kernel: eax: 00074000 ebx: 00000001 ecx: fff8b000 edx: 0000005e
May 18 23:50:30 palnx1 kernel: esi: fff8c000 edi: d5dac000 ebp: fff8c000 esp: d5daccd0
May 18 23:50:30 palnx1 kernel: ds: 007b es: 007b ss: 0068
May 18 23:50:30 palnx1 kernel: Process oracle (pid: 23771, threadinfo=d5dac000 task=f4df4bb0)
May 18 23:50:30 palnx1 kernel: Stack: 00000001 97800000 c402e480 c014e1fa 81724007 00000001 97800000 d6d59c80
May 18 23:50:30 palnx1 kernel: e3a41c24 d5dacd18 00100000 e3a41c24 f73012fc c1198580 c014e3e4 20000000
May 18 23:50:30 palnx1 kernel: 00100000 00000001 ffffffe1 00000000 00000000 00000000 00000000 f7301318
May 18 23:50:30 palnx1 kernel: Call Trace:
May 18 23:50:30 palnx1 kernel: [<c014e1fa>] try_to_unmap_cluster+0x1b1/0x1c4
May 18 23:50:30 palnx1 kernel: [<c014e3e4>] try_to_unmap_file+0x16d/0x21c
May 18 23:50:30 palnx1 kernel: [<c014e4c9>] try_to_unmap+0x36/0x49
May 18 23:50:30 palnx1 kernel: [<c01451d6>] shrink_list+0x1ba/0x3ed
May 18 23:50:30 palnx1 kernel: [<c02c7d6a>] invalidate_interrupt+0x1a/0x20
May 18 23:50:30 palnx1 kernel: [<c01455e6>] shrink_cache+0x1dd/0x34d
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [bugfix] try_to_unmap_cluster() passes out-of-bounds pte to pte_unmap()
2005-05-22 21:27 ` [bugfix] try_to_unmap_cluster() passes out-of-bounds pte to pte_unmap() William Lee Irwin III
@ 2005-05-22 22:00 ` Andrew Morton
2005-05-24 0:14 ` Andrew Morton
1 sibling, 0 replies; 32+ messages in thread
From: Andrew Morton @ 2005-05-22 22:00 UTC (permalink / raw)
To: William Lee Irwin III; +Cc: linux-kernel, gregkh
William Lee Irwin III <wli@holomorphy.com> wrote:
>
> try_to_unmap_cluster() does:
> for (pte = pte_offset_map(pmd, address);
> address < end; pte++, address += PAGE_SIZE) {
> ...
> }
>
> pte_unmap(pte);
>
> It may take a little staring to notice, but pte can actually fall off
> the end of the pte page in this iteration,
That's about the third place we've had this bug. Whoever keeps adding it
really should stop.
Thanks.
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [bugfix] try_to_unmap_cluster() passes out-of-bounds pte to pte_unmap()
2005-05-22 21:27 ` [bugfix] try_to_unmap_cluster() passes out-of-bounds pte to pte_unmap() William Lee Irwin III
2005-05-22 22:00 ` Andrew Morton
@ 2005-05-24 0:14 ` Andrew Morton
2005-05-24 2:48 ` William Lee Irwin III
1 sibling, 1 reply; 32+ messages in thread
From: Andrew Morton @ 2005-05-24 0:14 UTC (permalink / raw)
To: William Lee Irwin III; +Cc: linux-kernel, gregkh, Hugh Dickins
William Lee Irwin III <wli@holomorphy.com> wrote:
>
> --- ./mm/rmap.c.orig 2005-05-20 01:29:14.066467151 -0700
> +++ ./mm/rmap.c 2005-05-20 01:30:06.620649901 -0700
> @@ -694,7 +694,7 @@
> (*mapcount)--;
> }
>
> - pte_unmap(pte);
> + pte_unmap(pte-1);
> out_unlock:
> spin_unlock(&mm->page_table_lock);
> }
I must say that I continue to find this approach a bit queazifying.
After some reading of the code I'd agree that yes, it's not possible for us
to get here with `pte' pointing at the first slot of the pte page, but it's
not 100% obvious and it's possible that someone will come along later and
will change things in try_to_unmap_cluster() which cause this unmap to
suddenly do the wrong thing in rare circumstances.
IOW: I'd sleep better at night if we took a temporary and actually unmapped
the thing which we we got back from pte_offset_map().. Am I being silly?
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [bugfix] try_to_unmap_cluster() passes out-of-bounds pte to pte_unmap()
2005-05-24 0:14 ` Andrew Morton
@ 2005-05-24 2:48 ` William Lee Irwin III
2005-05-24 4:38 ` Hugh Dickins
0 siblings, 1 reply; 32+ messages in thread
From: William Lee Irwin III @ 2005-05-24 2:48 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel, gregkh, Hugh Dickins
William Lee Irwin III <wli@holomorphy.com> wrote:
>> --- ./mm/rmap.c.orig 2005-05-20 01:29:14.066467151 -0700
>> +++ ./mm/rmap.c 2005-05-20 01:30:06.620649901 -0700
[...]
On Mon, May 23, 2005 at 05:14:06PM -0700, Andrew Morton wrote:
> I must say that I continue to find this approach a bit queazifying.
> After some reading of the code I'd agree that yes, it's not possible for us
> to get here with `pte' pointing at the first slot of the pte page, but it's
> not 100% obvious and it's possible that someone will come along later and
> will change things in try_to_unmap_cluster() which cause this unmap to
> suddenly do the wrong thing in rare circumstances.
> IOW: I'd sleep better at night if we took a temporary and actually unmapped
> the thing which we we got back from pte_offset_map().. Am I being silly?
Not at all. I merely attempt to minimize diffsize by default. An
alternative implementation follows (changelog etc. to be taken
from the prior patch) in case it saves the time (however short) needed
to write it yourself.
-- wli
Index: mm2-2.6.12-rc4/mm/rmap.c
===================================================================
--- mm2-2.6.12-rc4.orig/mm/rmap.c 2005-05-20 01:44:18.000000000 -0700
+++ mm2-2.6.12-rc4/mm/rmap.c 2005-05-23 19:13:29.000000000 -0700
@@ -626,7 +626,7 @@
pgd_t *pgd;
pud_t *pud;
pmd_t *pmd;
- pte_t *pte;
+ pte_t *pte, *original_pte;
pte_t pteval;
struct page *page;
unsigned long address;
@@ -658,7 +658,7 @@
if (!pmd_present(*pmd))
goto out_unlock;
- for (pte = pte_offset_map(pmd, address);
+ for (original_pte = pte = pte_offset_map(pmd, address);
address < end; pte++, address += PAGE_SIZE) {
if (!pte_present(*pte))
@@ -694,7 +694,7 @@
(*mapcount)--;
}
- pte_unmap(pte);
+ pte_unmap(original_pte);
out_unlock:
spin_unlock(&mm->page_table_lock);
}
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [bugfix] try_to_unmap_cluster() passes out-of-bounds pte to pte_unmap()
2005-05-24 2:48 ` William Lee Irwin III
@ 2005-05-24 4:38 ` Hugh Dickins
2005-05-24 8:02 ` Nick Piggin
0 siblings, 1 reply; 32+ messages in thread
From: Hugh Dickins @ 2005-05-24 4:38 UTC (permalink / raw)
To: William Lee Irwin III; +Cc: Andrew Morton, linux-kernel, gregkh
On Mon, 23 May 2005, William Lee Irwin III wrote:
> On Mon, May 23, 2005 at 05:14:06PM -0700, Andrew Morton wrote:
> > I must say that I continue to find this approach a bit queazifying.
> > After some reading of the code I'd agree that yes, it's not possible for us
> > to get here with `pte' pointing at the first slot of the pte page, but it's
> > not 100% obvious and it's possible that someone will come along later and
> > will change things in try_to_unmap_cluster() which cause this unmap to
> > suddenly do the wrong thing in rare circumstances.
> > IOW: I'd sleep better at night if we took a temporary and actually unmapped
> > the thing which we we got back from pte_offset_map().. Am I being silly?
There's a similar argument for queasiness in all the other (8 or more)
instances of the idiom. I think we originally adopted (and I furthered)
this pte_unmap(pte - 1) idiom because in the majority of architecture's
configurations pte_unmap does nothing at all, so we resented assigning
a pointless variable in some critical loops.
> Not at all. I merely attempt to minimize diffsize by default. An
> alternative implementation follows (changelog etc. to be taken
> from the prior patch) in case it saves the time (however short) needed
> to write it yourself.
Either of wli's patches is fine with me. There are several levels on
which try_to_unmap_cluster is harder to understand than the others,
and no good reason to resist the variable assignment.
We could rewrite pte_unmap to avoid the issue completely, since its
job is to unmap (or pretend to unmap) KM_PTE0's pte if the address
is in the fixmap area: but changing it to tolerate an off-by-one
address gives a queasy feeling too.
Thanks,
Hugh
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [bugfix] try_to_unmap_cluster() passes out-of-bounds pte to pte_unmap()
2005-05-24 4:38 ` Hugh Dickins
@ 2005-05-24 8:02 ` Nick Piggin
0 siblings, 0 replies; 32+ messages in thread
From: Nick Piggin @ 2005-05-24 8:02 UTC (permalink / raw)
To: Hugh Dickins; +Cc: William Lee Irwin III, Andrew Morton, linux-kernel, gregkh
Hugh Dickins wrote:
> On Mon, 23 May 2005, William Lee Irwin III wrote:
>
>>On Mon, May 23, 2005 at 05:14:06PM -0700, Andrew Morton wrote:
>>
>>>I must say that I continue to find this approach a bit queazifying.
>>>After some reading of the code I'd agree that yes, it's not possible for us
>>>to get here with `pte' pointing at the first slot of the pte page, but it's
>>>not 100% obvious and it's possible that someone will come along later and
>>>will change things in try_to_unmap_cluster() which cause this unmap to
>>>suddenly do the wrong thing in rare circumstances.
>>>IOW: I'd sleep better at night if we took a temporary and actually unmapped
>>>the thing which we we got back from pte_offset_map().. Am I being silly?
>
>
> There's a similar argument for queasiness in all the other (8 or more)
> instances of the idiom. I think we originally adopted (and I furthered)
> this pte_unmap(pte - 1) idiom because in the majority of architecture's
> configurations pte_unmap does nothing at all, so we resented assigning
> a pointless variable in some critical loops.
>
Still, the compiler should be able to eliminate that extra register
as well as it can eliminate the intermediate (pte - 1) result (that
is to say, I hope perfectly in this day and age).
It may be more of an issue with architectures that actually *do* do
something in pte_unmap, in which case perhaps you increase the
register pressure over the critical loop? I guess we can just laugh
at them.
>
>>Not at all. I merely attempt to minimize diffsize by default. An
>>alternative implementation follows (changelog etc. to be taken
>>from the prior patch) in case it saves the time (however short) needed
>>to write it yourself.
>
>
> Either of wli's patches is fine with me. There are several levels on
> which try_to_unmap_cluster is harder to understand than the others,
> and no good reason to resist the variable assignment.
>
> We could rewrite pte_unmap to avoid the issue completely, since its
> job is to unmap (or pretend to unmap) KM_PTE0's pte if the address
> is in the fixmap area: but changing it to tolerate an off-by-one
> address gives a queasy feeling too.
>
Looks like no architecture (other than maybe frv?) even uses the
kvaddr argument to kunmap_atomic unless HIGHMEM_DEBUG/DEBUG_HIGHMEM
is enabled. If you stored that info elsewhere, you wouldn't even
need to pass the argument in.
But hmm... I don't see anyone getting motivated enough to rewrite the
debug code over this issue :)
--
SUSE Labs, Novell Inc.
^ permalink raw reply [flat|nested] 32+ messages in thread
* Problems with fb console [was Re: 2.6.12-rc4-mm2]
2005-05-16 9:13 2.6.12-rc4-mm2 Andrew Morton
` (9 preceding siblings ...)
2005-05-22 21:27 ` [bugfix] try_to_unmap_cluster() passes out-of-bounds pte to pte_unmap() William Lee Irwin III
@ 2007-06-27 0:35 ` J.A. Magallón
2007-06-27 0:54 ` Andrew Morton
2007-06-27 7:20 ` DervishD
10 siblings, 2 replies; 32+ messages in thread
From: J.A. Magallón @ 2007-06-27 0:35 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
On Mon, 16 May 2005 02:13:02 -0700, Andrew Morton <akpm@osdl.org> wrote:
>
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc4/2.6.12-rc4-mm2/
>
>
Hi...
I have a (stupid, I suppose) problem with framebuffer console.
I have builtin VESAFB in this kernel, so:
werewolf:/boot# grep _FB config-2.6.21-jam09 | grep =y
CONFIG_FB=y
CONFIG_FB_CFB_FILLRECT=y
CONFIG_FB_CFB_COPYAREA=y
CONFIG_FB_CFB_IMAGEBLIT=y
CONFIG_FB_DEFERRED_IO=y
CONFIG_FB_MODE_HELPERS=y
CONFIG_FB_VESA=y
werewolf:/boot# grep CONSO config-2.6.21-jam09
# CONFIG_NETCONSOLE is not set
CONFIG_VT_CONSOLE=y
CONFIG_HW_CONSOLE=y
# CONFIG_VT_HW_CONSOLE_BINDING is not set
CONFIG_VGA_CONSOLE=y
CONFIG_DUMMY_CONSOLE=y
CONFIG_FRAMEBUFFER_CONSOLE=y
# CONFIG_FRAMEBUFFER_CONSOLE_DETECT_PRIMARY is not set
# CONFIG_FRAMEBUFFER_CONSOLE_ROTATION is not set
I put this line in grub's menu.lst:
kernel /boot/vmlinuz video=vesafb:mtrr,ywrap vga=0x31A ro root=/dev/sdc1
(tried both with hex and decimal).
but grub keeps telling me it can't set that video mode, and I have no
/dev/fb0 device to try with fbset. I have a '29 fb' line in /proc/devices.
Any ideas about why the device is missing ? udev is 113...
I have followed al the info I could get (linux/Documentation/fb/, Google ;) )
and all say that what I'm doing should work. What am I doing wrong ?
TIA
--
J.A. Magallon <jamagallon()ono!com> \ Software is like sex:
\ It's better when it's free
Mandriva Linux release 2008.0 (Cooker) for i586
Linux 2.6.21-jam09 (gcc 4.1.2 20070302 (4.1.2-1mdv2007.1)) SMP PREEMPT
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: Problems with fb console [was Re: 2.6.12-rc4-mm2]
2007-06-27 0:35 ` Problems with fb console [was Re: 2.6.12-rc4-mm2] J.A. Magallón
@ 2007-06-27 0:54 ` Andrew Morton
2007-06-27 14:21 ` H. Peter Anvin
2007-06-27 7:20 ` DervishD
1 sibling, 1 reply; 32+ messages in thread
From: Andrew Morton @ 2007-06-27 0:54 UTC (permalink / raw)
To: "J.A. Magallón"
Cc: linux-kernel, Antonino A. Daplas, H. Peter Anvin
On Wed, 27 Jun 2007 02:35:27 +0200 "J.A. Magallón" <jamagallon@ono.com> wrote:
> On Mon, 16 May 2005 02:13:02 -0700, Andrew Morton <akpm@osdl.org> wrote:
>
> >
> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc4/2.6.12-rc4-mm2/
> >
> >
>
> Hi...
>
> I have a (stupid, I suppose) problem with framebuffer console.
> I have builtin VESAFB in this kernel, so:
>
> werewolf:/boot# grep _FB config-2.6.21-jam09 | grep =y
> CONFIG_FB=y
> CONFIG_FB_CFB_FILLRECT=y
> CONFIG_FB_CFB_COPYAREA=y
> CONFIG_FB_CFB_IMAGEBLIT=y
> CONFIG_FB_DEFERRED_IO=y
> CONFIG_FB_MODE_HELPERS=y
> CONFIG_FB_VESA=y
> werewolf:/boot# grep CONSO config-2.6.21-jam09
> # CONFIG_NETCONSOLE is not set
> CONFIG_VT_CONSOLE=y
> CONFIG_HW_CONSOLE=y
> # CONFIG_VT_HW_CONSOLE_BINDING is not set
> CONFIG_VGA_CONSOLE=y
> CONFIG_DUMMY_CONSOLE=y
> CONFIG_FRAMEBUFFER_CONSOLE=y
> # CONFIG_FRAMEBUFFER_CONSOLE_DETECT_PRIMARY is not set
> # CONFIG_FRAMEBUFFER_CONSOLE_ROTATION is not set
>
> I put this line in grub's menu.lst:
>
> kernel /boot/vmlinuz video=vesafb:mtrr,ywrap vga=0x31A ro root=/dev/sdc1
>
> (tried both with hex and decimal).
>
> but grub keeps telling me it can't set that video mode, and I have no
> /dev/fb0 device to try with fbset. I have a '29 fb' line in /proc/devices.
>
> Any ideas about why the device is missing ? udev is 113...
> I have followed al the info I could get (linux/Documentation/fb/, Google ;) )
> and all say that what I'm doing should work. What am I doing wrong ?
>
Methinks that'll be git-newsetup changes?
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: Problems with fb console [was Re: 2.6.12-rc4-mm2]
2007-06-27 0:54 ` Andrew Morton
@ 2007-06-27 14:21 ` H. Peter Anvin
0 siblings, 0 replies; 32+ messages in thread
From: H. Peter Anvin @ 2007-06-27 14:21 UTC (permalink / raw)
To: Andrew Morton
Cc: "J.A. Magallón", linux-kernel, Antonino A. Daplas
Andrew Morton wrote:
>>
>> I have a (stupid, I suppose) problem with framebuffer console.
>> I have builtin VESAFB in this kernel, so:
>>
>> werewolf:/boot# grep _FB config-2.6.21-jam09 | grep =y
^^^^^^
Is this actually the -mm kernel? And if so, does it work in earlier
kernels?
-hpa
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: Problems with fb console [was Re: 2.6.12-rc4-mm2]
2007-06-27 0:35 ` Problems with fb console [was Re: 2.6.12-rc4-mm2] J.A. Magallón
2007-06-27 0:54 ` Andrew Morton
@ 2007-06-27 7:20 ` DervishD
1 sibling, 0 replies; 32+ messages in thread
From: DervishD @ 2007-06-27 7:20 UTC (permalink / raw)
To: J.A. Magallón; +Cc: Andrew Morton, linux-kernel
Hi J.A. :)
* J.A. Magallón <jamagallon@ono.com> dixit:
> kernel /boot/vmlinuz video=vesafb:mtrr,ywrap vga=0x31A ro root=/dev/sdc1
>
> (tried both with hex and decimal).
Try "vga=0x031A". I vaguely remember I had a similar problem and
IIRC that's how I solved it. Anyway, you can always try vga=ask and try
to enter the number by hand, just to see which modes work.
Raúl Núñez de Arenas Coronado
--
Linux Registered User 88736 | http://www.dervishd.net
It's my PC and I'll cry if I want to... RAmen!
^ permalink raw reply [flat|nested] 32+ messages in thread