Zen-Sources: 2.6.26-rc3-zen0

[dusanc]

The patch sysresccd uses for reiser4 are exactly the same ones as we use, so there shouldn't be anything weird happening. As i said before, it works perfectly for me with cryptcompress and gzip1. I'm also on x86_64.

Hi
As one of reiser4 masochists I'd like to ask you if you could please send r4 errors you got with latest -mm or latest stable patch and ways to reproduce them to reiser4 m-l (reiserfs-devel@vger.kernel.org) as I couldn't make it fail with my setup using 2.6.26-rc2-mm1 so that developers could repair it .

Thanks in advance
Dushan

It only failed when using cryptcompress, the rest worked. With cryptcompress, while copying data to it, or using it, it would either panic for some people, or (for me) hang with any heavier io.

dodo

Well I'm using cryptocompress too on plain x86, and couldn't make it panic. I tested with NFS and loopback too.
Dunno I'll try making some heavier IO, and if it oopses for someone coluld they please post that oops.

Thanks
Dushan

[drescherjm]

Anyone have problems with suspend and hibernation? Since sometime in the 2.6.25 kernel (git-sources) I have not been able to do either. So I tried a few zen 2.6.26 (rc2 and rc3) kernels but they did not fix the problem. A few days ago I dropped back to tuxonice-sources-2.6.25-r4 and both are again working. My problems is that suspend reboots during the wake up process. And hibernation leaves me with a X session that is non functional. The mouse works for a few seconds on a black screen then it freezes and flashes a few times repeating this. This is on a dell d820 laptop with intel video. On all kernels I used the same .config except for the settings that were not available on the older kernel. The one setting I remember that was new was to activate the iwlwifi led however I doubt this is the cause.

Is this too off topic to post here since this happens with git-sources as well as zen? If so I am sorry and I can delete it from here.

[roadnottaken]

All this talk of reiser4 brought to mind a question that has been bugging me since I've tried to move to reiser4. Using reiser4progs 1.0.6, I have been able to display the list of plugins. However, I have no clue how to use most of them because I don't know the plugin types, except for those with the same types as a plugin in the default profile. For obvious reasons, I can't just look up plugin types on the Namesys website. Does anyone know what the various plugin types are?

[Mr. Tao]

roadnottaken: You can override which plugin to use for certain operations when you making filesystem. For example, if you want to make your FS using cryptocompress plugin do something like mkfs.reiser4 -o create=ccreg40 /dev/sda1.

Well I'm using cryptocompress too on plain x86, and couldn't make it panic. I tested with NFS and loopback too.
Dunno I'll try making some heavier IO, and if it oopses for someone coluld they please post that oops.

For me it only panicks when intensive operations are performed on loop file (portage sync, unmount). Main partition is R4CC with default settings with two loop files (portage and ccache) on it. Both are R4CC formatted as well. The first one (portage) is created as sparse file, second one (ccache) is ordinary prealocated loop file. Kernel panic occurs when I sync/unmount portage loop file. Only messages I got I've [post=5110009]already mentioned[/post]. I'd like to point out that this setup worked for few days without problems, but later, probably when sparse file started to grow, it started panicking.

[rmh3093]

roadnottaken: You can override which plugin to use for certain operations when you making filesystem. For example, if you want to make your FS using cryptocompress plugin do something like mkfs.reiser4 -o create=ccreg40 /dev/sda1.

Well I'm using cryptocompress too on plain x86, and couldn't make it panic. I tested with NFS and loopback too.
Dunno I'll try making some heavier IO, and if it oopses for someone coluld they please post that oops.

For me it only panicks when intensive operations are performed on loop file (portage sync, unmount). Main partition is R4CC with default settings with two loop files (portage and ccache) on it. Both are R4CC formatted as well. The first one (portage) is created as sparse file, second one (ccache) is ordinary prealocated loop file. Kernel panic occurs when I sync/unmount portage loop file. Only messages I got I've [post=5110009]already mentioned[/post]. I'd like to point out that this setup worked for few days without problems, but later, probably when sparse file started to grow, it started panicking.

why would u use reiser4 as a loopback file, now everytime u write to that there are 2 write actions instead of one, that is in essence defeating the whole purpose of using r4

[Mr. Tao]

why would u use reiser4 as a loopback file, now everytime u write to that there are 2 write actions instead of one, that is in essence defeating the whole purpose of using r4

Because of delayed write feature in R4 there isn't much of a write induced slowdown. I want to keep data physicaly close and I prefer loop files over separate partitions because of flexibility.

[roadnottaken]

roadnottaken: You can override which plugin to use for certain operations when you making filesystem. For example, if you want to make your FS using cryptocompress plugin do something like mkfs.reiser4 -o create=ccreg40 /dev/sda1.

Perhaps I wasn't quite clear enough. I already knew about the -o type=plugin switch, and I can generate a list of all available plugins, but their types within this list are given only as hex codes, while I need the actual names of the types to be able to use any plugins. When a plugin type is listed in the default profile, I know the name for that plugin type, but there are another 8 plugin types for which I know only the hex code, not the name.

Code: Select all

$ /sbin/mkfs.reiser4
Known plugins:
"stat40"            (id:0x0 type:0x1) [StatData item plugin.]
"cde40"             (id:0x2 type:0x1) [Compound directory entry item plugin.]
"nodeptr40"         (id:0x3 type:0x1) [Node pointer item plugin.]
"extent40"          (id:0x5 type:0x1) [Extent file body item plugin.]
"plain40"           (id:0x6 type:0x1) [Plain file body item plugin.]
"ctail40"           (id:0x7 type:0x1) [Compressed file body item plugin.]
"bbox40"            (id:0x8 type:0x1) [Safe link item plugin.]
"node40"            (id:0x0 type:0x2) [Node plugin.]
"sdext_lw"          (id:0x0 type:0x6) [Light stat data extension plugin.]
"sdext_unix"        (id:0x1 type:0x6) [Unix stat data extension plugin.]
"sdext_lt"          (id:0x2 type:0x6) [Large times stat data extension plugin.]
"sdext_symlink"     (id:0x3 type:0x6) [Symlink stat data extension plugin.]
"sdext_plugin_set"  (id:0x4 type:0x6) [Plugin Set StatData extension plugin.]
"sdext_flags"       (id:0x5 type:0x6) [Inode flags stat data extension plugin.]
"sdext_crypto"      (id:0x7 type:0x6) [Crypto stat data extension plugin.]
"sdext_heir_set"    (id:0x8 type:0x6) [Heir Set StatData extension plugin.]
"format40"          (id:0x0 type:0x7) [Disk-format plugin.]
"oid40"             (id:0x0 type:0x8) [Inode number allocator plugin.]
"alloc40"           (id:0x0 type:0x9) [Space allocator plugin.]
"journal40"         (id:0x0 type:0xa) [Journal plugin.]

Does anyone know what types these plugins represent?

[kernelOfTruth]

All this talk of reiser4 brought to mind a question that has been bugging me since I've tried to move to reiser4. Using reiser4progs 1.0.6, I have been able to display the list of plugins. However, I have no clue how to use most of them because I don't know the plugin types, except for those with the same types as a plugin in the default profile. For obvious reasons, I can't just look up plugin types on the Namesys website. Does anyone know what the various plugin types are?

cryptcompress related plugin options:

Settings related to transparent compression in Reiser4

The following is a list of pset members 1 which are needed to create/manage cryptcompress files. Currently values of listed pset members (except @file) are inherited from parent objects, and for root directory they are installed by mkfs.reiser4 via appropriate options.

file

Points to a file plugin which is to service VFS entrypoints.

List of currently possible values (label/description):

*

"reg40", "unix-file plugin". Manages regular files.
*

"ccreg40", "cryptcompress file plugin". Manages regular files with encrypted and(or) compressed bodies.
*

"dir40", "directory file plugin". Manages directories.
*

"sym40", "symlink file plugin". Manages symlinks.

Currently there is no interface to set this pset member explicitly. For regular files @file is assigned by default as @create of parent directory at creation time (see below).

create

Points to a file plugin, which is to define values of "file" pset member of regular children at creation time, i.e. the following assignment takes place:

reiser4_inode * object; /* directory */
reiser4_inode * child; /* regular file */
...

child->pset->file = object->pset->create;

List of currently possible values (label/description):

*

"reg40", "unix-file plugin". Create a regular file managed by unix-file plugin.
*

"ccreg40", "cryptcompress file plugin". Create a regular file managed by cryptcompress file plugin.

Default value 2 is "reg40" (create a unix-file).

cluster

Points to a cluster plugin which defines size of logical cluster in a cryptcompress file.

List of currently possible values (labels):

*

"4K", "8K", "16K", "32K", "64K".

Default value is "64K".

Note: logical clusters smaller then page size are unsupported.

compression

Points to a compression transform plugin, which defines algorithms of compression, decompression and checksum for data transform. Checksums are appended per disk cluster to verify data integrity before decompression.

List of currently possible values (label/description):

*

"lzo1", wrappers for minimal compression library of Markus Oberhumer, adler32 checksum.
*

"gzip1", wrappers for the fastest compression algorithms of linux zlib library, adler32 checksum.

Default value is "lzo1".

compression_mode

Points to a compression mode plugin which defines some heuristic for data compression. Currently this is a set of handlers called by deflate manager in flush time. Properly assigned compression mode allows to reduce cpu and memory usage when handling incompressible data.

List of currently possible values (label/description):

*

"ultim", "Check ultimately". In this mode compression will be "turned off" forever as soon as deflate manager meet incompressible page cluster.
*

"latt", "Check on dynamic lattice". This is adaptive mode that manage special per-inode compression toggle. This toggle shows last compression experience: ->discard_hook() sets it to 0; ->accept_hook() sets it to 1. ->should_compress() first looks at this toggle. If it is 1, then ->should_compress() advises to compress current logical cluster. Otherwise ->should_compress() advises to compress it only in the case when its index belongs to the lattice of variable factor X (X = 2, 4, ..., 32) which get increased after each unsuccessful experience (successful one returns it back to minimal value).
*

"conv", "Convert to extent". This mode activates a special hook in reiser4 context of sys_write() (when writing from special offset in some write iteration) which tries to estimate whether a file is compressible by testing its first logical cluster (64K by default). Such evaluation is performed only once per file's life. If evaluation result is negative, then fragments will be converted to extents, and management will be passed to unix-file plugin. Back conversion does not take place. If evaluation result is positive, then file remains under cryptcompress plugin control, but compression mode plugin will be converted to "latt", i.e all future decisions about compressibility will be in the competence of the flush algorithm.
*

"force", "Compress everything". In this mode deflate manager is forced to compress every logical cluster (unless it is too small).
*

"none", "Do not compress". In this mode deflate manager does not invoke compression.

Default value is "conv".

examples:

mkfs.reiser4 -o formatting=tails /dev/sda6

mkfs.reiser4 -o formatting=extents /dev/sda6

mkfs.reiser4 -o create=ccreg40,compress=gzip1 /dev/sda6
mkfs.reiser4 -o create=ccreg40,compress=lzo1 /dev/sda6

mkfs.reiser4 -o create=ccreg40,compress=gzip1,compressMode=conv /dev/sda6

mkfs.reiser4 -o create=ccreg40,compress=gzip1,compressMode=conv /dev/sda6
mkfs.reiser4 -o create=ccreg40,compress=gzip1,compressMode=latt /dev/sda6 (default ?)
mkfs.reiser4 -o create=ccreg40,compress=gzip1,compressMode=force /dev/sda6
mkfs.reiser4 -o create=ccreg40,compress=gzip1,compressMode=ultim /dev/sda6

currently I'm too busy to look up for more, this should suffice for the beginning ...

[kernelOfTruth]

roadnottaken: You can override which plugin to use for certain operations when you making filesystem. For example, if you want to make your FS using cryptocompress plugin do something like mkfs.reiser4 -o create=ccreg40 /dev/sda1.

Well I'm using cryptocompress too on plain x86, and couldn't make it panic. I tested with NFS and loopback too.
Dunno I'll try making some heavier IO, and if it oopses for someone coluld they please post that oops.

For me it only panicks when intensive operations are performed on loop file (portage sync, unmount). Main partition is R4CC with default settings with two loop files (portage and ccache) on it. Both are R4CC formatted as well. The first one (portage) is created as sparse file, second one (ccache) is ordinary prealocated loop file. Kernel panic occurs when I sync/unmount portage loop file. Only messages I got I've [post=5110009]already mentioned[/post]. I'd like to point out that this setup worked for few days without problems, but later, probably when sparse file started to grow, it started panicking.

please report, which zen-sources version you're using exactly, if it panics with 2.6.25-zen1 try 2.6.25-zen2 instead since it got concurrent pagecache removed which might make it less fragile ...

have fun

that's all for now (my time allows) - over and out

update:

ah it's supposed to be 2.6.26-rc3-* ?

[Mr. Tao]

please report, which zen-sources version you're using exactly, if it panics with 2.6.25-zen1 try 2.6.25-zen2 instead since it got concurrent pagecache removed which might make it less fragile ...

I've tried 2.6.24-zen2, 2.6.24-zen4 and 2.6.26-rc4-zen1.2 with same results.

[rmh3093]

My problems with Reiser4 are limited to R4 ccreg40 loop files on R4 ccreg40 partition. In this configuration, it panics either when updating paludis cache or when unmounting loop file. (While both works in from SR CD chrooted 2.6.25 environment).

Could you please try vanilla kernel with the reiser4 patch from kernel.org (link coupla posts above)? Maybe it's something in zen that is causing it, although i doubt it.

Also, the reiser4 patch applies cleanly, so it's definitely not patching fault.

dodo

It could be something to do with the CK patches, I seem to recall that 2.6.25-dark1 had this problem as well, and it pretty much includes nothing but part of CK and some small patches.

we dont use any ck-patches any more

[rmh3093]

why would u use reiser4 as a loopback file, now everytime u write to that there are 2 write actions instead of one, that is in essence defeating the whole purpose of using r4

Because of delayed write feature in R4 there isn't much of a write induced slowdown. I want to keep data physicaly close and I prefer loop files over separate partitions because of flexibility.

well I dont believe your first argument, but your second one I would go for, using a loopfile for portage which help a bunch in reducing fragmentation, i use a ext3 loop file for that

[Mr. Tao]

well I dont believe your first argument, but your second one I would go for, using a loopfile for portage which help a bunch in reducing fragmentation, i use a ext3 loop file for that

You should consider compression, therefore less data to be written, consequently resulting in smaller amount of write actions. Now you should become believer.

[MostAwesomeDude]

well I dont believe your first argument, but your second one I would go for, using a loopfile for portage which help a bunch in reducing fragmentation, i use a ext3 loop file for that

You should consider compression, therefore less data to be written, consequently resulting in smaller amount of write actions. Now you should become believer.

Already use compression. Squashfs + aufs = 40MB /usr/portage.

[dusanc]

roadnottaken: You can override which plugin to use for certain operations when you making filesystem. For example, if you want to make your FS using cryptocompress plugin do something like mkfs.reiser4 -o create=ccreg40 /dev/sda1.

Well I'm using cryptocompress too on plain x86, and couldn't make it panic. I tested with NFS and loopback too.
Dunno I'll try making some heavier IO, and if it oopses for someone coluld they please post that oops.

For me it only panicks when intensive operations are performed on loop file (portage sync, unmount). Main partition is R4CC with default settings with two loop files (portage and ccache) on it. Both are R4CC formatted as well. The first one (portage) is created as sparse file, second one (ccache) is ordinary prealocated loop file. Kernel panic occurs when I sync/unmount portage loop file. Only messages I got I've [post=5110009]already mentioned[/post]. I'd like to point out that this setup worked for few days without problems, but later, probably when sparse file started to grow, it started panicking.

Ask and you shall recieve

Here's the answer:

Loop devices over cryptocompress files don't work.
I disabled this functionality, but it seems the stub has
been broken at some point.

I'll try to fix the loop functionality for cryptcompress files.
For now in order to create loop devices use files managed
by (classic) unix-file plugin, Ok?

On a ccreg40 partition a file powered by unix-file plugin
can be prepared, for instance, by the following way:

dd if=incompressible_file of=loopfile bs=65537 count=1

then rewrite the loopfile with any data you want: it won't
be converted back to cryptcompress format anymore.
As "incompressible_file" have any .gz or .bz2 file of size > 64K

I know, it looks like ugly workaround, but I believe we'll invent
more user-friendly interface to assign desired plugins in future..

In short: it worked, it broke, it's known, it'll get fixed soon (as there no other outstanding bugs AFAIK it shouldn't be long), there's a workaround (which doesn't suits you) .

Moral of the story: Use the mailinglist Luke

Have fun
Dushan

[dusanc]

As reiser4 is pretty much plugin based, everithing it creates is based on some plugiun, so a lot of those plugins are for internal fs use.
If you're interested in their use please ask on m-l and you should get an answer in a day or two

[Mr. Tao]

dd if=incompressible_file of=loopfile bs=65537 count=1

then rewrite the loopfile with any data you want: it won't
be converted back to cryptcompress format anymore.
As "incompressible_file" have any .gz or .bz2 file of size > 64K

I know, it looks like ugly workaround, but I believe we'll invent
more user-friendly interface to assign desired plugins in future.

Is there a way to check which plugin manages given file? Until now I created these files dd if=/dev/random of=loopfile bs=16M count=32.

[whitesouls]

Just a teeny weeny question..hope you ppl dont mind.. I cant find the option of IBM Thinkpad in ACPI? or did i missed it?

[dodo1122]

Just a teeny weeny question..hope you ppl dont mind.. I cant find the option of IBM Thinkpad in ACPI? or did i missed it?

You missed it It's definitely there, as i have a thinkpad and use it.


dodo

[hirakendu]

Just a teeny weeny question..hope you ppl dont mind.. I cant find the option of IBM Thinkpad in ACPI? or did i missed it?

Device Drivers ---> Misc devices ---> Thinkpad ACPI Laptop Extras

[rmh3093]

Just a teeny weeny question..hope you ppl dont mind.. I cant find the option of IBM Thinkpad in ACPI? or did i missed it?

Device Drivers ---> Misc devices ---> Thinkpad ACPI Laptop Extras

FYI: you can press '/' in menuconfig and search for any config option

[hirakendu]

Just a teeny weeny question..hope you ppl dont mind.. I cant find the option of IBM Thinkpad in ACPI? or did i missed it?

Device Drivers ---> Misc devices ---> Thinkpad ACPI Laptop Extras

FYI: you can press '/' in menuconfig and search for any config option

neat

[whitesouls]

Just a teeny weeny question..hope you ppl dont mind.. I cant find the option of IBM Thinkpad in ACPI? or did i missed it?

Device Drivers ---> Misc devices ---> Thinkpad ACPI Laptop Extras

FYI: you can press '/' in menuconfig and search for any config option

Simply Superb!!! Thanks!!

[dodo1122]

Does anyone else experience fn keys on thinkpads not working correctly? Or is it just me? With .26 brightness down button no longer works, and all the volume buttons don't work as well.

dodo

[dodo1122]

2.6.26-rc5-zen1.0 tagged.

Updated to -rc5.
Tuxonice updated.
Wireless-testing updated.
drm updated.

Enjoy