linuxcompressed-devel — Linux Compressed Cache Development Mailing List

Re: [lc-devel] Help

[Rodrigo S. de C. <rc...@im...>]

On Wed, Mar 27, 2002 at 11:44:35AM +0800, Patrick wrote:
> Hi, I've downloaded the latest patch  Patch 2.4.18-0.23pre1 . 
> I've try to patch into Red Hat 7.1 and Mandrake 8.2 but failed. It
> seems like there're a lot of errors. Usually are saving rejects to
> file mm/*.
> What is wrong with my PC or the source code?

Probably nothing is wrong with your PC or the source code. It's likely
you've made some mistake when applying the patch. How've you done?

Regards,
-- 
Rodrigo S. de Castro <rc...@im...>

Re: [lc-devel] Page cache compression

[Rodrigo S. de C. <rc...@im...>]

Hi David et al,

On Fri, Mar 22, 2002 at 10:15:25PM +0800, David Chow wrote:
> Would anyone tell me what is page cache compression, on the change
> log for 0.22 Rodrigo says page cache is also compressed, can Rodrigo
> explain please. I would like to know what to aware when I implement
> my filesystem, because I also have compression in my page
> caches. Thanks.

The 0.22 version includes some new features like page cache
compression and clean pages support. The latter means that not only
the dirty pages, but also the clean pages, are added to compressed
cache. It something very sensible for compressed cache idea but wasn't
implemented so far.

The page cache support in compressed cache allows all file mapped
pages (ie, page cache pages) to be added to compressed cache. I
thought of implementing page cache support after noticing that some IO
intensive applications didn't have performance gains with the "usual"
compressed cache (the one which stores only anonymous pages). Some
tests had been run and out of them some statistics were made available
on our project page (statistics for version 0.21) where we could
notice performance drop when running IO intensive applications (for
example, dbench) on a kernel with compressed cache enabled.

At that moment I didn't think very much on the impact of an idea like
that and started implementing this support to make compressed cache
work better for these cases. Along with the clean page support, it
actually helped IO intensive applications, like dbench (you can have a
look at the statistics page for 0.22pre7 and later version).

Nevertheless, this page cache support is new and experimental, never
implemented in other compressed cache implementations and not even
mentioned in Scott Kaplan's thesis. We still have to perform tests and
some deep analyses to check if this support will be kept in the
future, mainly after implementing adaptivity.

You can notice that a compressed cache like that, with swap cache and
page cache support, will hold pages with very different
behaviours. Unlike swap cache pages, dirty file mapped pages need be
synced to the disk, which will follow OS-specific parameters (kupdated
parameters in Linux, for example) or user/applications parameters (for
instance, syscalls or mount with sync flag). That's why I don't know
exactly when it's worth to compress a file mapped page, since it might
be decompressed to be synced right after the compression compression
(thus we might be wasting compressions).

Besides that, in both cases storing a page from either swap cache or
page caches might save us a disk read (if we reference it again and we
still have it in compressed cache and it would have to be read from
disk otherwise). In swap cache page, it might even save us a disk
write if it happens to be compressed and never have to write out to
swap until program exits (in common case we have dirty swap
pages). With a page cache page, we are still not sure it might save us
a disk write, since it will need to synced soon if dirty.

Those reasons above make me believe we won't have a single parameter
for adaptivity that will work with swap and page cache at the same
time. I don't even think we would be able to find a single parameter
that would work well only with page cache. Therefore, I am not sure
this support will be kept when adaptivity gets implemented.

Anyway, I think it's been worth to think about this problem no matter
what happens to this page cache support. This page cache support
showed us that we were right about some assumptions (for example, that
dbench bad performance was, even partialy, due to a smaller page and
buffer caches) and also helped us see some similar problems in other
compressed cache implementations.

Best regards,
-- 
Rodrigo S. de Castro <rc...@im...>

[lc-devel] Help

[Patrick <pat...@my...>]

Hi, I've downloaded the latest patch  Patch 2.4.18-0.23pre1 . 
I've try to patch into Red Hat 7.1 and Mandrake 8.2 but failed. It seems like there're a lot of errors. Usually are saving rejects to file mm/*.
What is wrong with my PC or the source code?

BTW i've tried on Intel Pentium MMX and Celeron , both with 64MB SDRAM. 

Please help.

Thanks.

[lc-devel] Page cache compression

[David C. <dav...@sh...>]

Would anyone tell me what is page cache compression, on the change log 
for 0.22 Rodrigo says page cache is also  compressed, can Rodrigo 
explain please. I would like to know what to aware when I implement my 
filesystem, because I also have compression in my page caches. Thanks.

regards,

David

Re: [lc-devel] Module Problems with 0.22 final AND vanilla 2.4.18

[Marc-Christian P. <m....@gm...>]

On Donnerstag 21 M=E4rz 2002 14:32 you wrote:

Hi Rodrigo,

> Have you saved the patch and removed my message and also the
> sourceforge message at the bottom of the email? I've just done that,
> and it applied cleaned on a vanilla + 0.22 tree.
you must be joking! SURE!! :-)

> 1. Save the email to a file, removing the crap (my message and sf
>    note)
exactly done so.

> [rcastro@vinci:~/mestrado/kernel/orig]$ patch -p1 -i /tmp/patch
> patching file mm/comp_cache/Makefile
> patching file mm/comp_cache/swapin.c
> [rcastro@vinci:~/mestrado/kernel/orig]$
very very curious :)

--=20
Kind regards
=09Marc-Christian Petersen

http://sourceforge.net/projects/wolk

PGP/GnuPG Key: 1024D/569DE2E3DB441A16
Fingerprint: 3469 0CF8 CA7E 0042 7824  080A 569D E2E3 DB44 1A16
Key available at wwwkeys.pgp.net. =A0 Encrypted e-mail preferred.

Re: [lc-devel] Module Problems with 0.22 final AND vanilla 2.4.18

[Rodrigo S. de C. <rc...@im...>]

On Thu, Mar 21, 2002 at 02:00:54PM +0100, Marc-Christian Petersen wrote:
> On Donnerstag 21 März 2002 12:57 Rodrigo Souza de Castro wrote:
> >That patch was for my pool, that already has some changes. Given that
> >you have 0.22, try the next one.
> 
> root@codeman:/usr/src/linux-2.4.18# patch -p1 --dry-run < 
> patch-comp-cache-2.4.18-0.22-symexpfix 
> patch: **** Only garbage was found in the patch input.

Have you saved the patch and removed my message and also the
sourceforge message at the bottom of the email? I've just done that,
and it applied cleaned on a vanilla + 0.22 tree. 

Anyway, since it only changes comp cache files, it should not have any
error, even if you have other patches applied (likely to be your
case).

1. Save the email to a file, removing the crap (my message and sf
   note)

2. Apply it to your tree.

[rcastro@vinci:~/mestrado/kernel/orig]$ patch -p1 -i /tmp/patch
patching file mm/comp_cache/Makefile
patching file mm/comp_cache/swapin.c
[rcastro@vinci:~/mestrado/kernel/orig]$

Regards,
-- 
Rodrigo S. de Castro <rc...@im...>

Re: [lc-devel] Module Problems with 0.22 final AND vanilla 2.4.18

[Marc-Christian P. <m....@gm...>]

On Donnerstag 21 M=E4rz 2002 12:57 Rodrigo Souza de Castro wrote:

Hi Rodrigo,

>That patch was for my pool, that already has some changes. Given that
>you have 0.22, try the next one.

root@codeman:/usr/src/linux-2.4.18# patch -p1 --dry-run <=20
patch-comp-cache-2.4.18-0.22-symexpfix=20
patch: **** Only garbage was found in the patch input.

?

--=20
Kind regards
=09Marc-Christian Petersen

http://sourceforge.net/projects/wolk

PGP/GnuPG Key: 1024D/569DE2E3DB441A16
Fingerprint: 3469 0CF8 CA7E 0042 7824  080A 569D E2E3 DB44 1A16
Key available at wwwkeys.pgp.net. =A0 Encrypted e-mail preferred.

Re: [lc-devel] Module Problems with 0.22 final AND vanilla 2.4.18

[Rodrigo S. de C. <rc...@im...>]

On Thu, Mar 21, 2002 at 08:44:01AM -0300, Rodrigo Souza de Castro wrote:
> On Thu, Mar 21, 2002 at 10:01:36AM +0100, Marc-Christian Petersen wrote:
> > root@codeman:/# depmod -ae
> > depmod: *** Unresolved symbols in 
> > /lib/modules/2.4.18/kernel/fs/ncpfs/ncpfs.o
> > depmod:         flush_comp_cache
> > depmod: *** Unresolved symbols in 
> > /lib/modules/2.4.18/kernel/fs/smbfs/smbfs.o
> > depmod:         flush_comp_cache
> > 
> > How can i get rid of this??
> 
> Thanks for the report. I forgot those filesystems could be built as
> modules and didn't export that symbol. Check if the following patch
> solves your problem. The fix will be in 0.23pre1 too.

That patch was for my pool, that already has some changes. Given that
you have 0.22, try the next one.

Regards,
-- 
Rodrigo S. de Castro <rc...@im...>

--- orig/mm/comp_cache/Makefile.orig	Thu Mar 21 08:48:10 2002
+++ orig/mm/comp_cache/Makefile	Thu Mar 21 08:48:31 2002
@@ -4,6 +4,8 @@
 
 O_TARGET := comp_cache.o
 
+export-objs := swapin.o
+
 obj-y    := main.o vswap.o free.o swapout.o swapin.o adaptivity.o aux.o proc.o WK4x4.o WKdm.o
 
 include $(TOPDIR)/Rules.make
--- orig/mm/comp_cache/swapin.c.orig	Thu Mar 21 08:48:06 2002
+++ orig/mm/comp_cache/swapin.c	Thu Mar 21 08:49:15 2002
@@ -1,7 +1,7 @@
 /*
  * linux/mm/comp_cache/swapin.c
  *
- * Time-stamp: <2002-03-14 16:41:47 rcastro>
+ * Time-stamp: <2002-03-21 08:49:15 rcastro>
  *  
  * Linux Virtual Memory Compressed Cache
  * - Swap In and Decompression Routines
@@ -13,6 +13,7 @@
 #include <linux/comp_cache.h>
 #include <linux/mm.h>
 #include <linux/pagemap.h>
+#include <linux/module.h>
 
 #include <asm/uaccess.h>
 #include <net/checksum.h>
@@ -28,6 +29,8 @@
 
 	return err;
 }
+
+EXPORT_SYMBOL(flush_comp_cache);
 
 int
 flush_comp_cache(struct address_space *mapping, unsigned long offset, struct page * page)

Re: [lc-devel] Module Problems with 0.22 final AND vanilla 2.4.18

[Rodrigo S. de C. <rc...@im...>]

Hi Marc-Christian,

On Thu, Mar 21, 2002 at 10:01:36AM +0100, Marc-Christian Petersen wrote:
> root@codeman:/# depmod -ae
> depmod: *** Unresolved symbols in 
> /lib/modules/2.4.18/kernel/fs/ncpfs/ncpfs.o
> depmod:         flush_comp_cache
> depmod: *** Unresolved symbols in 
> /lib/modules/2.4.18/kernel/fs/smbfs/smbfs.o
> depmod:         flush_comp_cache
> 
> How can i get rid of this??

Thanks for the report. I forgot those filesystems could be built as
modules and didn't export that symbol. Check if the following patch
solves your problem. The fix will be in 0.23pre1 too.

> Hmm, 0.22pre6 has no problems but some oops'ses at runtime :-(((

Yes, one of the problems in 0.22pre6 was the lack of this
flush_comp_cache() to keep the compressed cache uptodate :-)

Regards,
-- 
Rodrigo S. de Castro <rc...@im...>

--- linuxcompressed/mm/comp_cache/Makefile.orig	Wed Mar 20 11:45:03 2002
+++ linuxcompressed/mm/comp_cache/Makefile	Thu Mar 21 08:30:26 2002
@@ -4,6 +4,8 @@
 
 O_TARGET := comp_cache.o
 
+export-objs := swapin.o
+
 obj-y    := main.o vswap.o free.o swapout.o swapin.o adaptivity.o aux.o proc.o WK4x4.o WKdm.o
 
 include $(TOPDIR)/Rules.make
--- linuxcompressed/mm/comp_cache/swapin.c.orig	Wed Mar 20 15:50:04 2002
+++ linuxcompressed/mm/comp_cache/swapin.c	Thu Mar 21 08:37:01 2002
@@ -1,7 +1,7 @@
 /*
  * linux/mm/comp_cache/swapin.c
  *
- * Time-stamp: <2002-03-20 15:50:04 rcastro>
+ * Time-stamp: <2002-03-21 08:37:01 rcastro>
  *  
  * Linux Virtual Memory Compressed Cache
  * - Swap In and Decompression Routines
@@ -13,6 +13,7 @@
 #include <linux/comp_cache.h>
 #include <linux/mm.h>
 #include <linux/pagemap.h>
+#include <linux/module.h>
 
 #include <asm/uaccess.h>
 #include <net/checksum.h>
@@ -29,6 +30,8 @@
 	return err;
 }
 
+EXPORT_SYMBOL(flush_comp_cache);
+
 int FASTCALL(flush_comp_cache(struct page *));
 int flush_comp_cache(struct page * page)
 {	
@@ -56,7 +59,6 @@
  out:	
 	return err;
 }
-
 
 extern void comp_cache_update_decomp_stats(unsigned short, stats_page_t *);
 

[lc-devel] Module Problems with 0.22 final AND vanilla 2.4.18

[Marc-Christian P. <m....@gm...>]

Hi Rodrigo,

if i use vanilla 2.4.18 with final 0.22 compressed cache and i compile my=
=20
modules and install them the following occurs:


root@codeman:/# depmod -ae
depmod: *** Unresolved symbols in=20
/lib/modules/2.4.18/kernel/fs/ncpfs/ncpfs.o
depmod:         flush_comp_cache
depmod: *** Unresolved symbols in=20
/lib/modules/2.4.18/kernel/fs/smbfs/smbfs.o
depmod:         flush_comp_cache

How can i get rid of this??

Hmm, 0.22pre6 has no problems but some oops'ses at runtime :-(((


--=20
Kind regards
=09Marc-Christian Petersen

http://sourceforge.net/projects/wolk

PGP/GnuPG Key: 1024D/569DE2E3DB441A16
Fingerprint: 3469 0CF8 CA7E 0042 7824  080A 569D E2E3 DB44 1A16
Key available at wwwkeys.pgp.net. =A0 Encrypted e-mail preferred.

Re: [lc-devel] compressed cache final 22

[Rodrigo S. de C. <rc...@im...>]

On Mon, Mar 18, 2002 at 08:40:24PM +0100, Marc-Christian Petersen wrote:
> gcc -D__KERNEL__ -I/usr/src/linux-2.4.18/include  -Wall -Wstrict-prototypes 
> -Wno-trigrap
> hs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pipe 
> -mpreferred-stack-boundary=2 
> -march=i686  
> -I/usr/src/linux-2.4.18/tools/badMEM/badmem/badmemlib/include  -DKBUILD_BAS
> ENAME=filemap  -DEXPORT_SYMTAB -c filemap.c
> filemap.c: In function `filemap_fdatawait':
> filemap.c:634: parse error before `int'
> filemap.c:646: `_lock_break_count' undeclared (first use in this function)
> filemap.c:646: (Each undeclared identifier is reported only once
> filemap.c:646: for each function it appears in.)
> make[2]: *** [filemap.o] Error 1
> make[2]: Leaving directory `/usr/src/linux-2.4.18/mm'
> make[1]: *** [first_rule] Error 2
> make[1]: Leaving directory `/usr/src/linux-2.4.18/mm'
> make: *** [_dir_mm] Error 2
> 
> .22pre6 worked fine. 
> I want to release a new WOLK Kernel, v3.1, i want to include final 22 
> compressed cache, WOLK 3.0 has 22pre6 included, got some oops.

I just recompiled a clean vanilla tree patched with 0.22 final and had
no compile errors whatsoever. By the way, this _lock_break_count is
something not in the vanilla (at least not in filemap_fdatawait())
code nor in compressed cache code. Don't you have another patch
applied to this tree that may be the cause of these errors?

Regards,
-- 
Rodrigo S. de Castro <rc...@im...>

[lc-devel] compressed cache final 22

[Marc-Christian P. <m....@gm...>]

Hi there,

gcc -D__KERNEL__ -I/usr/src/linux-2.4.18/include  -Wall -Wstrict-prototyp=
es=20
-Wno-trigrap
hs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pipe=20
-mpreferred-stack-boundary=3D2=20
-march=3Di686 =20
-I/usr/src/linux-2.4.18/tools/badMEM/badmem/badmemlib/include  -DKBUILD_B=
AS
ENAME=3Dfilemap  -DEXPORT_SYMTAB -c filemap.c
filemap.c: In function `filemap_fdatawait':
filemap.c:634: parse error before `int'
filemap.c:646: `_lock_break_count' undeclared (first use in this function=
)
filemap.c:646: (Each undeclared identifier is reported only once
filemap.c:646: for each function it appears in.)
make[2]: *** [filemap.o] Error 1
make[2]: Leaving directory `/usr/src/linux-2.4.18/mm'
make[1]: *** [first_rule] Error 2
make[1]: Leaving directory `/usr/src/linux-2.4.18/mm'
make: *** [_dir_mm] Error 2

=2E22pre6 worked fine.=20
I want to release a new WOLK Kernel, v3.1, i want to include final 22=20
compressed cache, WOLK 3.0 has 22pre6 included, got some oops.

--=20
Kind regards
=09Marc-Christian Petersen

http://sourceforge.net/projects/wolk

PGP/GnuPG Key: 1024D/569DE2E3DB441A16
Fingerprint: 3469 0CF8 CA7E 0042 7824  080A 569D E2E3 DB44 1A16
Key available at wwwkeys.pgp.net. =A0 Encrypted e-mail preferred.

[lc-devel] compressed cache final 22

[Marc-Christian P. <m....@gm...>]

Hi there,

gcc -D__KERNEL__ -I/usr/src/linux-2.4.18/include  -Wall -Wstrict-prototypes 
-Wno-trigrap
hs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pipe 
-mpreferred-stack-boundary=2 
-march=i686  
-I/usr/src/linux-2.4.18/tools/badMEM/badmem/badmemlib/include  -DKBUILD_BAS
ENAME=filemap  -DEXPORT_SYMTAB -c filemap.c
filemap.c: In function `filemap_fdatawait':
filemap.c:634: parse error before `int'
filemap.c:646: `_lock_break_count' undeclared (first use in this function)
filemap.c:646: (Each undeclared identifier is reported only once
filemap.c:646: for each function it appears in.)
make[2]: *** [filemap.o] Error 1
make[2]: Leaving directory `/usr/src/linux-2.4.18/mm'
make[1]: *** [first_rule] Error 2
make[1]: Leaving directory `/usr/src/linux-2.4.18/mm'
make: *** [_dir_mm] Error 2

.22pre6 patch worked fine. 
I want to release a new WOLK Kernel, v3.1, i want to include final 22 
compressed cache cause WOLK 3.0 has 22pre6 included which causes some oops.

-- 
Kind regards
        Marc-Christian Petersen

http://sourceforge.net/projects/wolk

PGP/GnuPG Key: 1024D/569DE2E3DB441A16
Fingerprint: 3469 0CF8 CA7E 0042 7824  080A 569D E2E3 DB44 1A16
Key available at wwwkeys.pgp.net.   Encrypted e-mail preferred.

-- 
GMX - Die Kommunikationsplattform im Internet.
http://www.gmx.net

[lc-devel] [PATCH] 2.4.18-0.22

[Rodrigo S. de C. <rc...@im...>]

Hi all,

The new 0.22 version is out for 2.4.18 version of Linux Kernel. This
version has two main features: the whole page cache is supported, so
from now on all pages from page cache (not only swap cache pages) are
compressed; and manual adaptivity is fully implemented.

- Page Cache Support

These changes are not complex but took quite a while to accomplish a
good stability. Some prelimary tests with dbench and 0.22pre7 are
available at:

http://linuxcompressed.sourceforge.net/statistics/0.22pre7_dbench/

This code may still need some changes for a better performance (which
will be done as soon as kernprof patch is available for 2.4.18), but
the results from these tests are pretty better than the ones performed
with 0.21.

- Clean Pages Support + Pages Not Removed in Page Fault

All pages are compressed now, no matter they are clean or not. That
means that only dirty fragments (= compressed pages) are going to be
written out to the backing storage.

Another interesting feature introduced in this version is the one
which only removes the pages when it gets unmapped (the process dies,
the file is removed, etc) or when we are out of space in compressed
cache. Beforehand we always removed the fragment from compressed cache
when servicing a page fault.

- Manual Adaptivity

Now all main data structures (fragment hash table and vswap) are
correctly resized to support the new compressed cache size. You can do
that echoing the new number of pages to be used by compressed cache to
the following file: /proc/sys/vm/comp_cache/size. A cat in this file
will show the current number of pages reserved for compressed cache.

Kernel parameter and Config option are going to set the _initial_ (not
maximum) number of pages used by compressed pages, you will be able to
change this number whenever you want. Note: The compressed cache
cannot be disable yet, so a very small number of pages for compressed
cache may become a bottleneck depending on the scenario. An option to
disable the compressed cache on the fly is one of our high-priority
itens on our todo list.

- Other Features

This version also have shared memory support back, many cleanups and
bug fixes, simpler and much more documented code (mainly vswap) and
also a code more stable (with new allocation failure handling, for
instance).

- Stability 

I tested a lot this patch, running concurrently mmap001, fillmem and
changing compressed cache size. I also compiled several times the
Linux Kernel and had a hard time to make dbench stable, but finally
did. The last tests I ran were those first tests (mmap001, fillmem and
comp cache resize) for a period of over 12 hours trying to make the
system really short of memory (usual scenario to trigger annoying
bugs). The code passed it. Anyway, I'd like to hear reports about its
stability.

Detailed Change Log:

http://sourceforge.net/project/shownotes.php?release_id=79628

Download:

http://prdownloads.sourceforge.net/linuxcompressed/patch-comp-cache-2.4.18-0.22.bz2

Regards,
-- 
Rodrigo S. de Castro <rc...@im...>

Re: [lc-devel] Further correction fo flush_comp_cache()

[Rodrigo S. de C. <rc...@im...>]

On Sat, Mar 16, 2002 at 09:19:19AM +0800, David Chow wrote:
> If we made the call in other filesystems to flush_comp_cache(),
> include comp_cache.h in fs/smbfs/dir.c and fs/nspfs/dir.c otherwise
> will end up with unresolved symbols, but I am not sure are you going
> to export all the symbols in comp_cache.h. Please make the
> correction.

Thanks, your fixes are in my pool. The 0.22 version will have them.

Regards,
-- 
Rodrigo S. de Castro <rc...@im...>

[lc-devel] Further correction fo flush_comp_cache()

[David C. <dav...@sh...>]

If we made the call in other filesystems to flush_comp_cache(), include 
comp_cache.h in fs/smbfs/dir.c and fs/nspfs/dir.c otherwise will end up 
with unresolved symbols, but I am not sure are you going to export all 
the symbols in comp_cache.h. Please make the correction.

regards,

David

[lc-devel] 0.22pre8 correction

[David C. <dav...@sh...>]

Dear all,

I think there is a typo error in the 2.4.18-0.22pre8 for 2 files 
fs/nspfs/dir.c fs/smbfs/dir.c  . It will result in a compilation error, 
here is the correction.

regards,

David


diff -Naur -X exclude-files orig/fs/ncpfs/dir.c 
linuxcompressed/fs/ncpfs/dir.c
--- orig/fs/ncpfs/dir.c    Wed Mar  7 21:53:48 2001
+++ linuxcompressed/fs/ncpfs/dir.c    Mon Mar  4 08:56:46 2002
@@ -450,6 +450,7 @@
             if (!ctl.page)
                 goto invalid_cache;
             ctl.cache = kmap(ctl.page);
+            flush_comp_cache(&inode->i_data, ctl.ofs, ctl.page);
             if (!Page_Uptodate(ctl.page))
                 goto invalid_cache;
         }
diff -Naur -X exclude-files orig/fs/ncpfs/dir.c 
linuxcompressed/fs/ncpfs/dir.c
--- orig/fs/ncpfs/dir.c    Wed Mar  7 21:53:48 2001
+++ linuxcompressed/fs/ncpfs/dir.c    Mon Mar  4 08:56:46 2002
@@ -450,6 +450,7 @@
             if (!ctl.page)
                 goto invalid_cache;
             ctl.cache = kmap(ctl.page);
+            flush_comp_cache(&inode->i_data, ctl.ofs, ctl.page);
             if (!Page_Uptodate(ctl.page))
                 goto invalid_cache;
         }

Re: [lc-devel] Bug at comp_swap.c line 170

[Rodrigo S. de C. <rc...@im...>]

On Fri, Mar 15, 2002 at 12:37:03AM +0800, David Chow wrote:
> I am running a Pentium 4 machine with 128MB ram having 8192 pages of 
> compressed cache. I use KDB to keep track of bugs. From my log file says 
> "BUG at comp_swap.c line 170", the oops triggers kdb and BT shows the 
> oops at swap_compressed() . 

Due to some surgeries in the code, I removed compressed swap support
when added page cache support.

> vswap_address(entry) returns non-zero value and says invalid swap
> file .

It seems some invalid swap file entry. 

> Would you please tell me why? Thanks.

I'd like to, but I have no clue.

Thanks for the report,
-- 
Rodrigo S. de Castro <rc...@im...>

[lc-devel] Bug at comp_swap.c line 170

[David C. <dav...@sh...>]

I am running a Pentium 4 machine with 128MB ram having 8192 pages of 
compressed cache. I use KDB to keep track of bugs. From my log file says 
"BUG at comp_swap.c line 170", the oops triggers kdb and BT shows the 
oops at swap_compressed() . vswap_address(entry) returns non-zero value 
and says invalid swap file . Would you please tell me why? Thanks.

Linux 2.4.17 with 0.22pre4

David

Re: [lc-devel] 2.4.18 + comp-cache-2.4.18-0.22pre6"

[Rodrigo S. de C. <rc...@im...>]

On Tue, Mar 12, 2002 at 08:20:31AM -0300, Rodrigo Souza de Castro wrote:
> On Tue, Mar 12, 2002 at 11:30:02AM +0100, Marc-Christian Petersen wrote:
> > Kernel "2.4.18" with patch "patch-comp-cache-2.4.18-0.22pre6"
> 
> The 0.22pre6 is very very unstable and should be used carefully. The
> next version (0.22pre7) fixed many memory corruption bugs in the page
> cache compression support which may be the cause of this oops. I will
> release a patch today and could you give it a go to see if the bug is
> still there? I really wouldn't like to have this bug in the 0.22 final
> version I intend to release these next days.

The 0.22pre8 patch for 2.4.18 version of Linux Kernel is already
available at:

http://prdownloads.sourceforge.net/linuxcompressed/patch-comp-cache-2.4.18-0.22pre8.bz2

And details about it can be found at:

http://sourceforge.net/forum/forum.php?forum_id=160297

This is (hopefuly) the last version before 0.22 final. I will be
checking its stability for the next hours before releasing the new
major version. Once stability is our focus now, I'd like to have a
feedback from whoever has spare time to test it.

Thanks,
-- 
Rodrigo S. de Castro <rc...@im...>

Re: [lc-devel] 2.4.18 + comp-cache-2.4.18-0.22pre6"

[Rodrigo S. de C. <rc...@im...>]

Hi Marc,

On Tue, Mar 12, 2002 at 11:30:02AM +0100, Marc-Christian Petersen wrote:
> Kernel "2.4.18" with patch "patch-comp-cache-2.4.18-0.22pre6"

The 0.22pre6 is very very unstable and should be used carefully. The
next version (0.22pre7) fixed many memory corruption bugs in the page
cache compression support which may be the cause of this oops. I will
release a patch today and could you give it a go to see if the bug is
still there? I really wouldn't like to have this bug in the 0.22 final
version I intend to release these next days.

> I used pre6 cause pre7 does not compile.

Fix already in my pool.

> The System is still up now, the oops is some minutes ago, but last night, from
> 4am to 6am every 10-15 minutes the oops occurs with ending in a kernel panic at
> 6am.

That happens all the time with bugs like that. It oopses and stops
only the current code path, but the kernel keeps running. Since it
seems a general memory corruption, the other code paths will
eventually oops too, ending up in a code with a panic, which stops all
the kernel :-)

> Bug? Feature? ;)

Surely a bug :-) Thank you for the report.

Best Regards,
-- 
Rodrigo S. de Castro <rc...@im...>

[lc-devel] 2.4.18 + comp-cache-2.4.18-0.22pre6"

[Marc-Christian P. <m....@gm...>]

Hi there,

Kernel "2.4.18" with patch "patch-comp-cache-2.4.18-0.22pre6"
I used pre6 cause pre7 does not compile.

kernel: Unable to handle kernel paging request at virtual address 7275736=
e
kernel: printing eip:
kernel: c0134d7d
kernel: *pde =3D 00000000
kernel: Oops: 0000
kernel: CPU: 0
kernel: EIP: 1010:[comp_cache_free+17/52] Tainted: P
kernel: EFLAGS: 00210246
kernel: eax: 7275736e ebx: 74636172 ecx: caeb6360 edx: 00000000
kernel: esi: 00000000 edi: cb10fdac ebp: ca85be60 esp: ca85be60
kernel: ds: 1018 es: 1018 ss: 1018
kernel: Process nmbd (pid: 11259, stackpage=3Dca85b000)
kernel: Stack: ca85be78 c01356af caeb6360 cb10fd94 00000000 00000000 ca85=
be98 c01356de
kernel: cb10fdac 00000000 00000000 00000000 00000000 cb10fd94 ca85bec0 c0=
128c7b
kernel: cb10fd94 00000000 00000000 00000000 00000000 cb10fd94 ca85bebc 00=
000000
kernel: Call Trace: [truncate_list_comp_pages+43/64] [truncate_comp_pages=
+26/48] [truncate_inode_pages+111/124] [vmtruncate+143/280] [inode_setatt=
r+43/192]
kernel: [ext3_setattr+428/500] [notify_change+81/332] [do_truncate+69/92]=
 [sys_ftruncate64+263/288] [system_call+58/68]
kernel:=20
kernel: Code: 8b 00 0f ab 50 18 19 d2 85 d2 74 02 0f 0b 51 e8 0f fb ff ff

The System is still up now, the oops is some minutes ago, but last night,=
 from
4am to 6am every 10-15 minutes the oops occurs with ending in a kernel pa=
nic at
6am.

Bug? Feature? ;)

--=20
Kind regards
=09Marc-Christian Petersen


PGP/GnuPG Key: 1024D/408B2D54947750EC
Fingerprint: 8602 69E0 A9C2 A509 8661  2B0B 408B 2D54 9477 50EC
Key available at wwwkeys.pgp.net. =A0 Encrypted e-mail preferred.

[lc-devel] I want to work on this

[vijender y. <vij...@ho...>]

Sir,


I am third year computer engg. student from INDIA.I am working with a 
company which works on systems basically in unix.We are a group of three 
students who have to do a project with the company.So we are searching for 
latest developements in the field of system programming.
We found this project on compressed cache very interesting.

So could u please help us to know more about ur workings.
We will be highly grateful to u.

Regards
vijender yadav





_________________________________________________________________
Send and receive Hotmail on your mobile device: http://mobile.msn.com

Re: [lc-devel] Linux - compressed caching

[Rodrigo S. de C. <rc...@im...>]

On Sun, Mar 03, 2002 at 10:02:53AM -0800, sangeetha wrote:
> we r supposed to do a mini-project work in Linux this sem,so we(a
> team of two) have chosen the topic Linux-compressed caching.........

Brave men. Good luck.

> plz can anyone tell me any useful websites that wud help in the
> development of our project work,

There is a list of useful web sites and references for our development
available on our web page (http://linuxcompressed.sf.net). User Mode
Linux will make your development quicker, so this tool is
essential. Oh, and gdb is your friend, so learn it if you don't know
it yet.

> and also tell us how to start off........thanx in advance.

First of all, I assume you two know C language very well. Then the
first step would be understanding reasonably well the Linux memory
management code (linux/mm directory), mainly the swapout path. Later
you devise your design ideas and start coding. Only coding you start
noticing minor details that you didn't pay attention at first.

You may download old versions of our code to have a look at the first
ideas (current code is a little bit more complex to start
with). Anyway, be aware that Linux VM has changed a lot since then.

Regards,
-- 
Rodrigo S. de Castro <rc...@im...>

[lc-devel] Linux - compressed caching

[sangeetha <san...@ya...>]

HI,
   we r supposed to do a mini-project work in Linux
this sem,so we(a team of two) have chosen the topic
Linux-compressed caching.........plz can anyone tell
me any useful websites that wud help in the
development of our project work, and also tell us how
to start off........thanx in advance.


sangeetha

__________________________________________________
Do You Yahoo!?
Yahoo! Sports - sign up for Fantasy Baseball
http://sports.yahoo.com