__udivdi3

__udivdi3

[Gergely Gábor]

[-- Attachment #1.1: Type: text/plain, Size: 749 bytes --]

hello

i was about to try nilfs (testing-8) but altough it compiles, it doesnt work, as it can't find a symbol:

  Building modules, stage 2.
  MODPOST 1 modules
WARNING: "__udivdi3" [/home/ggergely/tmp/nilfs/nilfs-2.0.0-testing-8/fs/nilfs2.ko] undefined!
WARNING: "__umoddi3" [/home/ggergely/tmp/nilfs/nilfs-2.0.0-testing-8/fs/nilfs2.ko] undefined!
  CC      /home/ggergely/tmp/nilfs/nilfs-2.0.0-testing-8/fs/nilfs2.mod.o

insmod also complains about this. Found a latter on this list, but I couldn't find a solution. Serche the net but i couldn't solve the problem (I didnt even find where is a reference to this symbol....)

Regards: Gábor Gergely

-- 
Gergely Gábor <elentirmo.gilgalad-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>

[-- Attachment #1.2: Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 158 bytes --]

_______________________________________________
users mailing list
users-JrjvKiOkagjYtjvyW6yDsg@public.gmane.org
https://www.nilfs.org/mailman/listinfo/users

Re: __udivdi3

[Chris Samuel]

[-- Attachment #1.1: Type: text/plain, Size: 664 bytes --]

On Thu, 3 Jan 2008, Gergely Gábor wrote:

> Serche the net but i couldn't solve the problem (I didnt even find
> where is a reference to this symbol....)

A reference from the ARM list:

http://marc.info/?l=linux-arm-kernel&m=112125281504347&w=2

Andrew Morton back in February implied that it was still there for 
x86:

http://www.ussg.iu.edu/hypermail/linux/kernel/0702.3/1064.html

Dunno if it's been pulled in later kernels though..

cheers!
Chris
-- 
 Chris Samuel  :  http://www.csamuel.org/  :  Melbourne, VIC

This email may come with a PGP signature as a file. Do not panic.
For more info see: http://en.wikipedia.org/wiki/OpenPGP

[-- Attachment #1.2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 481 bytes --]

[-- Attachment #2: Type: text/plain, Size: 158 bytes --]

_______________________________________________
users mailing list
users-JrjvKiOkagjYtjvyW6yDsg@public.gmane.org
https://www.nilfs.org/mailman/listinfo/users

Re: __udivdi3

[Gergely Gábor]

[-- Attachment #1.1: Type: text/plain, Size: 1338 bytes --]

On Thu, 3 Jan 2008 22:38:42 +1100
Chris Samuel <chris-Ru3vu9m2XlBAfugRpC6u6w@public.gmane.org> wrote:

> On Thu, 3 Jan 2008, Gergely Gábor wrote:
> 
> > Serche the net but i couldn't solve the problem (I didnt even find
> > where is a reference to this symbol....)
> 
> A reference from the ARM list:
> 
> http://marc.info/?l=linux-arm-kernel&m=112125281504347&w=2
> 
> Andrew Morton back in February implied that it was still there for 
> x86:
> 
> http://www.ussg.iu.edu/hypermail/linux/kernel/0702.3/1064.html
> 
> Dunno if it's been pulled in later kernels though..
> 
> cheers!
> Chris

If I understand well, this is used by the standard '/' operator on x86 (that is my case, and btw. I have 2.6.23 kernel with gentoo patches, if that counts), so some division on the code references that macro. this case I cannot solve this myself, yet I can't get the code to work... Please help, I'm really eager to try nilfs!

Should the final 2.0 nilfs be released, will it be ready for use for example on a /home partition in real life, or will that still be considered risky/dangerous?

BTW: if I'll be able to get nilfs to work, I'll make an ebuild for gentoo, that can be put up to the nilfs site.

Best regards: Gábor Gergely
-- 
Gergely Gábor <elentirmo.gilgalad-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>

[-- Attachment #1.2: Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 158 bytes --]

_______________________________________________
users mailing list
users-JrjvKiOkagjYtjvyW6yDsg@public.gmane.org
https://www.nilfs.org/mailman/listinfo/users

Re: __udivdi3

[amagai-sG5X7nlA6pw]

Hi,

Could you try following steps?
================
cd nilfs-2.0.0-testing-8/fs
edit Makefile, uncomment this line to enable debug option.

#CONFIG_NILFS_DEBUG=y

make clean
make

egrep -i udivdi *.lst
================
I hope that the egrep shows filename which includes the
undefined symbols.

Should the final 2.0 nilfs be released, will it be ready for use for 
example on a /home partition in real life, or will that still be
considered risky

My colleagues (and include me) are used nilfs 2 as /home 
of developping PC's with onw risk :-)
Nilfs 2 cannot handle a disk full condition well, 
so, parameter tuning in /etc/nilfs_cleanerd.conf are very critical.

Thank you.

Amagai Yoshiji
amagai-sG5X7nlA6pw@public.gmane.org

Re: __udivdi3

[Gergely Gábor]

[-- Attachment #1.1: Type: text/plain, Size: 2301 bytes --]

On Fri,  4 Jan 2008 10:47:21 +0900 (JST)
amagai-sG5X7nlA6pw@public.gmane.org wrote:

> Hi,
> 
> Could you try following steps?
> ================
> cd nilfs-2.0.0-testing-8/fs
> edit Makefile, uncomment this line to enable debug option.
> 
> #CONFIG_NILFS_DEBUG=y
> 
> make clean
> make
> 
> egrep -i udivdi *.lst
> ================
> I hope that the egrep shows filename which includes the
> undefined symbols.
> 
To me it looks like the Makefile has some errors, or my make behaves strangely
/*
ggergely@vosztok ~ $ make -v
GNU Make 3.81
Copyright (C) 2006  Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.

This program built for i686-pc-linux-gnu
ggergely@vosztok ~ $ 
*/
I think the Makefile error is that make doesn't expand $@ in line 20
/*
  EXTRA_CFLAGS += -g -Wa,-adhlms=$@.lst -fno-omit-frame-pointer
*/
as it flooded me with info. i saved the stderr, and realised this:
/*
{standard input}: Assembler messages:
{standard input}:45564: Warning: can't open .lst: Permission denied
...

although, the uber dirty trick to change the line to 
/*
  CFLAGS += -g -Wa,-adhlms=$@.lst -fno-omit-frame-pointer
*/
did the trick. ;) I recommend examining this error as well, (I'm not a makefile master, it'd take ages for me to find the cause and the fix) but if I see the situation wee for the first glimpse this as well may be some linux kernel makefile related stuff.

ggergely@vosztok ~/tmp/nilfs/nilfs-2.0.0-testing-8/fs $ egrep -i udivdi *.lst
dat.o.lst: 974              	.globl __udivdi3
dat.o.lst: 1245 03cb E8FCFFFF 	call __udivdi3
dat.o.lst:__udivdi3
ggergely@vosztok ~/tmp/nilfs/nilfs-2.0.0-testing-8/fs $ 


> My colleagues (and include me) are used nilfs 2 as /home 
> of developping PC's with onw risk :-)
> Nilfs 2 cannot handle a disk full condition well, 
> so, parameter tuning in /etc/nilfs_cleanerd.conf are very critical.

Thank you, this is good news for me! I hope soon I'll be running nilfs2!

Sorry for the long letter, hope I could help, and I'm waiting for the news.
Best regards: Gábor Gergely
-- 
Gergely Gábor <elentirmo.gilgalad-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>

[-- Attachment #1.2: Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 158 bytes --]

_______________________________________________
users mailing list
users-JrjvKiOkagjYtjvyW6yDsg@public.gmane.org
https://www.nilfs.org/mailman/listinfo/users

Re: __udivdi3

[amagai-sG5X7nlA6pw]

--text follows this line--
Gergely, thank you your report!

> GNU Make 3.81

The version of gnu make is same as one that I used on OpenSuse
10.3.  What's wrong???

> dat.o.lst: 974              	.globl __udivdi3
> dat.o.lst: 1245 03cb E8FCFFFF 	call __udivdi3
> dat.o.lst:__udivdi3

Could you send me ( amagai-sG5X7nlA6pw@public.gmane.org , not to ML) the whole
dat.o.lst ?  It's very useful to solve issue.

Thanks

amagai-sG5X7nlA6pw@public.gmane.org
NTT CyberSpace Laboratories
NILFS team

Re: __udivdi3

[amagai-sG5X7nlA6pw]

Gergely, thank you for your report!

> GNU Make 3.81

The version of gnu make is same as one that I used on OpenSuse
10.3.  What's wrong???

> dat.o.lst: 974              	.globl __udivdi3
> dat.o.lst: 1245 03cb E8FCFFFF 	call __udivdi3
> dat.o.lst:__udivdi3

Could you send me ( amagai-sG5X7nlA6pw@public.gmane.org , not to ML) the whole
dat.o.lst ?  It's very useful to solve issue.

amagai-sG5X7nlA6pw@public.gmane.org
NTT Cyber Space Laboratories
NILFS team

Re: __udivdi3

[amagai-sG5X7nlA6pw]

Hi, 
please try following patch, 

  cd nilfs-2.0.0-testing-8
  patch -p0 < patch_file
  cd fs
  make clean
  make
  make install

Thanks to Gergely, your dat.o.lst is very helpful!

Amagai Yoshiji
NTT
NILFS team.
================================================================

diff -u fs.save/dat.c fs/dat.c
--- fs.save/dat.c	2008-01-04 22:43:58.000000000 +0900
+++ fs/dat.c	2008-01-04 22:48:46.000000000 +0900
@@ -74,7 +74,7 @@
 inline static unsigned long
 nilfs_dat_group_offset(const struct inode *dat, nilfs_sector_t vblocknr)
 {
-	return sector_div(vblocknr, nilfs_dat_entries_per_group(dat));
+	return do_div(vblocknr, nilfs_dat_entries_per_group(dat));
 }
 
 inline static nilfs_gdno_t
@@ -1161,7 +1161,7 @@
 		kaddr = kmap_atomic(entry_bh->b_page, KM_USER0);
 		/* last virtual block number in this block */
 		first = vinfo[i].vi_vblocknr;
-		sector_div(first, entries_per_block);
+		do_div(first, entries_per_block);
 		first *= entries_per_block;
 		last = first + entries_per_block - 1;
 		for (j = i, n = 0;

Re: __udivdi3

[Pierre Etchemaite]

Le vendredi 04 janvier 2008 à 23:02 +0900, amagai-sG5X7nlA6pw@public.gmane.org a écrit :
> Hi, 
> please try following patch, 
> 
>   cd nilfs-2.0.0-testing-8
>   patch -p0 < patch_file
>   cd fs
>   make clean
>   make
>   make install

I could compile the kernel module, load it, mkfs a partition and mount
the filesystem...

So far, so good with this patch. Thanks!

Best regards,
Pierre.

Re: __udivdi3

[amagai-sG5X7nlA6pw]

> I could compile the kernel module, load it, mkfs a partition and mount
> the filesystem...

It's a very good news!
Thanks.

Amagai Yoshiji
NTT
NILFS team

Re: __udivdi3

[Gergely Gábor]

[-- Attachment #1.1: Type: text/plain, Size: 352 bytes --]

On Fri,  4 Jan 2008 23:02:46 +0900 (JST)
<amagai-sG5X7nlA6pw@public.gmane.org> wrote:

> Hi, 
> please try following patch, 

Thanks! Now it works like a charm (at least, for the few minutes in a loop device)

Great work! I'll make the ebuild in the weekend.

-- 
Gergely Gábor <elentirmo.gilgalad-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>

[-- Attachment #1.2: Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 158 bytes --]

_______________________________________________
users mailing list
users-JrjvKiOkagjYtjvyW6yDsg@public.gmane.org
https://www.nilfs.org/mailman/listinfo/users