Skip to contentSkip to conversation

Linux 2.6.6-rc3

DISCUSSION Vincent C Jones 5 participants 7 messages Apr 28, 2004 – May 1, 2004 5 messages not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

In article <[email protected]> you write:

Quoted 9 lines

s390, cifs, ntfs, ppc, ppc64, cpufreq upates. Oh, and DVB and USB.

I'm hoping to do a final 2.6.6 later this week, so I'm hoping as many people as possible will test this.

Thanks,

Linus

loop-AES 2.0g has refused to compile since 2.6.6-rc1. No problems up through 2.6.5. The make script cd's to the kernel source root, then can't find any of the kernel include files, and it's all downhill from there...

SuSE 9.0 on a ThinkPad X23. I'd be glad to provide more info if it would help.

Vince

Signature · 6 lines
Dr. Vincent C. Jones, PE              Expert advice and a helping hand
Computer Network Consultant           for those who want to manage and
Networking Unlimited, Inc.            control their networking destiny
14 Dogwood Lane, Tenafly, NJ 07670
http://www.networkingunlimited.com
[email protected]  +1 201 568-7810  Fax: +1 201 568-7269 
#2 lore raw
Quoted 7 lines · Vincent C Jones

In article <[email protected]> you write:

Quoted 6 lines

s390, cifs, ntfs, ppc, ppc64, cpufreq upates. Oh, and DVB and USB.

I'm hoping to do a final 2.6.6 later this week, so I'm hoping as many people as possible will test this.

Quoted 8 lines
Quoted 3 lines

Thanks,

Linus

loop-AES 2.0g has refused to compile since 2.6.6-rc1. No problems up through 2.6.5. The make script cd's to the kernel source root, then can't find any of the kernel include files, and it's all downhill from there...

This is an external module - right?
Could you mail me privately the Makefile used, or even better the full package (or link to it).

Sam

#3 lore raw
Quoted 12 lines · Sam Ravnborg
Quoted 11 lines · Vincent C Jones

In article <[email protected]> you write:

Quoted 9 lines

s390, cifs, ntfs, ppc, ppc64, cpufreq upates. Oh, and DVB and USB.

I'm hoping to do a final 2.6.6 later this week, so I'm hoping as many people as possible will test this.

Thanks,

Linus

Quoted 8 lines
Quoted 4 lines

loop-AES 2.0g has refused to compile since 2.6.6-rc1. No problems up through 2.6.5. The make script cd's to the kernel source root, then can't find any of the kernel include files, and it's all downhill from there...

This is an external module - right?
Could you mail me privately the Makefile used, or even better the full package (or link to it).

Full package:
http://loop-aes.sourceforge.net/loop-AES/loop-AES-v2.0g.tar.bz2

Patch to fix build problem:
http://loop-aes.sourceforge.net/updates/loop-AES-v2.0g-20040412.diff.bz2

New version of loop-AES will be released shortly after final 2.6.6 kernel is released.

Signature · 1 line
Jari Ruusu  1024R/3A220F51 5B 4B F9 BB D3 3F 52 E9  DB 1D EB E3 24 0E A9 DD
#4 lore raw
Quoted 3 lines

Full package:
http://loop-aes.sourceforge.net/loop-AES/loop-AES-v2.0g.tar.bz2

Hi Jari.
I took a look at how you use the build system in the 2.6 kernel. Inherited from the 2.4 days you assemble the commands yourself, which is plain wrong in a 2.6 kernel.
The only sane way to build external modules with a 2.6 kernel is to utilise the kbuild infrastructure.

For your reference here is a Makefile that I used to compile your module (made a symling for loop-patched.c file).

To compile the kernel I used:
make -C /home/sam/bk/v2.6/ M=$PWD
[Assumes latest linus kernel - 2.6.6-rc3]

For older kernels append the modules target

This has the added benefit that Module versioning is also supported.

#########################
# Minimal kbuild Makefile for loop-AES

EXTRA_CFLAGS := $(LOOP_AES_CFLAGS)

obj-m := loop.o

i586-$(CONFIG_M586) := -i586
i586-$(CONFIG_M686) := -i586

loop-y := loop-patched.o aes$(i586-y).o glue.o md5$(i586-y).o

##########################

I see in the Makefile that you do a lot of tricks to support various kernel versions. But they all end up in a few defines for the C compiler, which you just needs to supply in the variable LOOP_AES_CFLAGS. [And some file massaging whaich is done before starting the build]

What I would recommend you to do is to move all your backward compatibility stuff and general rules to a file named 'makefile'.
Then provide individual Makefiles for each kernel version:
Makefile.2.4, Makefile.2.6
Then symlink Makfile to the right one and build the module.

This would clean up your Makefile and give you correct usage in 2.6

Sam

#5 lore raw
In reply to Jari Ruusu
Quoted 12 lines · Jari Ruusu
Quoted 10 lines · Sam Ravnborg
Quoted 5 lines · Vincent C Jones

loop-AES 2.0g has refused to compile since 2.6.6-rc1. No problems up through 2.6.5. The make script cd's to the kernel source root, then can't find any of the kernel include files, and it's all downhill from there...

This is an external module - right?
Could you mail me privately the Makefile used, or even better the full package (or link to it).

Quoted 11 lines

Full package:
http://loop-aes.sourceforge.net/loop-AES/loop-AES-v2.0g.tar.bz2

Patch to fix build problem:
http://loop-aes.sourceforge.net/updates/loop-AES-v2.0g-20040412.diff.bz2

New version of loop-AES will be released shortly after final 2.6.6 kernel is released.

Signature · 1 line
Jari Ruusu  1024R/3A220F51 5B 4B F9 BB D3 3F 52 E9  DB 1D EB E3 24 0E A9 DD

Thanks much for the quick response. With the loop-AES makefile patched, loop-AES and 2.6.6-rc3 are happy together.

One last question: Is there a reason there is no mention of the compile problem or presence of a patch on the loop-AES sourceforge home page (http://sourceforge.net/projects/loop-aes/)? Or is this common knowledge that everyone is supposed to know about sourceforge projects? At least with this thread, Google searches will no longer come up empty.

Now all I need is an APM implementation which suspends while on A/C power and I can stop playing with kernel releases and concentrate on work.

Thanks again!

Vince

Signature · 6 lines
Dr. Vincent C. Jones, PE              Expert advice and a helping hand
Computer Network Consultant           for those who want to manage and
Networking Unlimited, Inc.            control their networking destiny
14 Dogwood Lane, Tenafly, NJ 07670
http://www.networkingunlimited.com
[email protected]  +1 201 568-7810  Fax: +1 201 568-7269 

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

#10 lore raw

Re: 2.6.6-rc3: modular DVB tda1004x broken

Quoted 9 lines · Andrew Morton

Maybe we should change __syscall_return() to return the -ve errno rather than -1?

diff -puN include/asm-i386/unistd.h~a include/asm-i386/unistd.h

include/asm-i386/unistd.h line 299

Quoted 9 lines
include/asm-i386/unistd.h renamed from include/asm-i386/unistd.h~a +0 -4
@@ -295,10 +295,6 @@
 
 #define __syscall_return(type, res) \
 do { \
-	if ((unsigned long)(res) >= (unsigned long)(-125)) { \
-		errno = -(res); \
-		res = -1; \
-	} \
 	return (type) (res); \
 } while (0)

Don't do that please. That will break all the user space programs who use asm/unistd.h to define own system calls (it is quite common).

Make it conditional on __KERNEL__

-Andi

#11 lore raw

Re: 2.6.6-rc3: modular DVB tda1004x broken

Quoted 19 lines · Andi Kleen
Quoted 5 lines

diff -puN include/asm-i386/unistd.h~a include/asm-i386/unistd.h

include/asm-i386/unistd.h line 299

Quoted 9 lines
include/asm-i386/unistd.h renamed from include/asm-i386/unistd.h~a +0 -4
@@ -295,10 +295,6 @@
 
 #define __syscall_return(type, res) \
 do { \
-	if ((unsigned long)(res) >= (unsigned long)(-125)) { \
-		errno = -(res); \
-		res = -1; \
-	} \
 	return (type) (res); \
 } while (0)

Don't do that please. That will break all the user space programs who use asm/unistd.h to define own system calls (it is quite common).

Quoted 1 line

Make it conditional on __KERNEL__

err, that was just a "technology demonstration". Obviously there's a lot more involved. Such as fixing up all the other architectures and then killing off all the `errno' users.