colinux-devel — Cooperative Linux development mailing list

[coLinux-devel] "Console and daemons" related issue ...

[Shyokou O. <sh...@gm...>]

Just read the file, "NEWS.txt", and found a line under the section, "Console
and daemons", of "Version 0.7.2", stating, "Stop with error, if double
defined device (cobd,cofs,eth,exec)" ... but I think it would be better
if the so-called "double defined" can be tolerated as a feature in a config
file, dont you?

On most of the boxes running coLinux/Linux here in this lab, two or more
admins might modify the same config file; but they all keep the original one
intact as the default backup, while adding their individual sections as if
the parameter(s) "double defined" ... why ? at least, it is easy to read
both the default and the modified parameter(s), and even easier to port some
similar modifications with a single "diff -urN"; meanwhile, as a one-way
modification, in some circumstances, it is easy to automatically update
config file using patch algorithm.

I wonder if it is acceptable to just warn about "double defined" rather than
"Stop with error", and to be compatible with most of the script style config
file, latter definition overrides any former one.
Thanks

-- 
Shyokou Ouyou <sh...@gm...>

[coLinux-devel] [ colinux-Feature Requests-2114493 ] add iscsi kernel support

[SourceForge.net <no...@so...>]

Feature Requests item #2114493, was opened at 2008-09-16 18:34
Message generated for change (Comment added) made by henryn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622066&aid=2114493&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Priority: 5
Private: No
Submitted By: Ritesh Raj Sarraf (riteshsarraf)
Assigned to: Nobody/Anonymous (nobody)
Summary: add iscsi kernel support

Initial Comment:
Hi,

I noticed that coLinux is modular but the features enabled is very minimal.

I see room for many more features to be enabled, like iscsi et cetera.

Is there a particular reason why these features are disabled?
Or is it something a user can enable?


Ritesh

----------------------------------------------------------------------

>Comment By: Henry N. (henryn)
Date: 2008-11-26 01:34

Message:
New modules are not in coLinux, because we have ported from one of very old
config step by step between kernel versions without changing to many
things. So, simply not all modules are enabled. Some modules does not
working under coLinux. In this case iscsi seems me a module with software
layer only (no hardware device) and should work without any problems.

Please use the 20081123-Snapshot and update the modules from
http://www.henrynestler.com/colinux/testing/devel-0.8.0/20081123-Snapshot/iscsi/

It would be nice if you can give a positive result.

Henry


----------------------------------------------------------------------

Comment By: Ritesh Raj Sarraf (riteshsarraf)
Date: 2008-11-25 13:17

Message:
Hi,

Can you please just comment if this is a valid request ?

I just want to confirm if coLinux by default should be providing the
modules or is it that the user is expected to compile additional modules
are required.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622066&aid=2114493&group_id=98788

[coLinux-devel] [ colinux-Feature Requests-2114493 ] add iscsi kernel support

[SourceForge.net <no...@so...>]

Feature Requests item #2114493, was opened at 2008-09-16 22:04
Message generated for change (Comment added) made by riteshsarraf
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622066&aid=2114493&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Priority: 5
Private: No
Submitted By: Ritesh Raj Sarraf (riteshsarraf)
Assigned to: Nobody/Anonymous (nobody)
Summary: add iscsi kernel support

Initial Comment:
Hi,

I noticed that coLinux is modular but the features enabled is very minimal.

I see room for many more features to be enabled, like iscsi et cetera.

Is there a particular reason why these features are disabled?
Or is it something a user can enable?


Ritesh

----------------------------------------------------------------------

>Comment By: Ritesh Raj Sarraf (riteshsarraf)
Date: 2008-11-25 17:47

Message:
Hi,

Can you please just comment if this is a valid request ?

I just want to confirm if coLinux by default should be providing the
modules or is it that the user is expected to compile additional modules
are required.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622066&aid=2114493&group_id=98788

[coLinux-devel] autobuild org.colinux.devel 20081123 21:08:35

[coLinux a. <col...@he...>]

The autobuild system has detected a new revision in the source repository.
Review last changed from changelog.txt, also attached in mail.

Download the compiled version:
http://www.henrynestler.com/colinux/autobuild/devel-20081123/

 colinux-0.8.0-20081123.src.tgz  (694578 Bytes)
 daemons-0.8.0-20081123.dbg.zip  (595377 Bytes)
 daemons-0.8.0-20081123.zip  (480725 Bytes)

Note, the autobuild compilation does not include an installer.
Remember to reload the driver with these commands:

	colinux-daemon.exe --remove-driver
	colinux-daemon.exe --install-driver

The vmlinux and modules are up to date. Please use last version from
http://www.henrynestler.com/colinux/autobuild/devel-20081112/

The autobuild compilations are not official releases of Cooperative Linux
software.  There is no warranty that any autobuild version is stable.  If
use this autobuild version, please give us feedback of your experience.

Job runs on machine with 64 bit version of gcc 4.1.2.
A service from http://gcc.gnu.org/wiki/CompileFarm

-- 
Lots of fun with newest version,
Henry Nestler
------------------------------------------------------------------------
r1144 | henryn | 2008-11-23 21:08:35 +0000 (Sun, 23 Nov 2008) | 4 lines
Changed paths:
   M /branches/devel/src/colinux/os/winnt/kernel/fileio.c

* cofs: ZwQueryInformationFile/ZwSetInformationFile (used for set file time):
  FILE_READ_ATTRIBUTES/FILE_WRITE_ATTRIBUTES is enough.
  Remove flags FILE_READ_DATA|FILE_WRITE_DATA and the second open.

------------------------------------------------------------------------
r1143 | henryn | 2008-11-23 20:57:50 +0000 (Sun, 23 Nov 2008) | 4 lines
Changed paths:
   M /branches/devel/src/colinux/os/winnt/kernel/fileio.c
   M /branches/devel/src/colinux/os/winnt/kernel/filesystem.c

* cofs: Bugfix attribs for mount point. C:\ is not a directory on ntfs.
  Mode was badly forwarded from first file or directory in the drive.
  (Reported by Fr?\195?\169d?\195?\169ric L.W. Meunier.)

------------------------------------------------------------------------
r1142 | henryn | 2008-11-23 20:11:47 +0000 (Sun, 23 Nov 2008) | 1 line
Changed paths:
   M /branches/devel/src/colinux/os/winnt/kernel/fileio.c

* co_rc_t co_os_change_file_information, co_rc_t co_os_set_file_information: Functions as static.
------------------------------------------------------------------------
r1141 | henryn | 2008-11-23 20:09:32 +0000 (Sun, 23 Nov 2008) | 1 line
Changed paths:
   M /branches/devel/NEWS
   M /branches/devel/RUNNING
   M /branches/devel/src/colinux/os/winnt/user/install/iDl.ini

* Wiki moved back to colinux.wikia.com, reverts SVN r1098.
------------------------------------------------------------------------

[coLinux-devel] [ colinux-Feature Requests-1898496 ] prepackaged GUI image

[SourceForge.net <no...@so...>]

Feature Requests item #1898496, was opened at 2008-02-21 02:05
Message generated for change (Comment added) made by gboutwel
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622066&aid=1898496&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: prepackaged GUI image

Initial Comment:
The slashdot discussion (http://developers.slashdot.org/developers/04/04/12/211236.shtml?tid=106) shows quite well that many users would find it easier if there was a prepackaged version that comes with all the necessary programs and configs to easily run a UI (i.e. X11/KDE) version on top of coLinux.

Something like this would preferably come with the necessary X server for the windows side of things and be sufficiently preconfigured

----------------------------------------------------------------------

>Comment By: George P Boutwell (gboutwel)
Date: 2008-11-21 13:59

Message:
The original author and I had a few discussions about this back when we
first started the port to 2.6 kernels...  The current developers may have
different opinions, but I tend to agree with what was decided then.  To
summarize:

coLinux project's main purpose is to provide the patched kernel and host
machine drivers necessary to accomplish fast (near native performance)
virtualized Linux.  With the few developers working on the project is is
completely unrealistic to expect or extend this purpose beyond that. 
Focusing on that, it was decided that coLinux as a project would not go out
of it's way to build and support an installer which included everything for
graphical images, this in turn allowed several projects to be built on
coLinux that servered that purpose similar to that of andLinux (used to be
one called Colonize or something like that too but couldn't find it on
Source Forge today when I looked).

Henry, you think this choice to continue to focus on our main goal of
providing the kernel and drivers for a fast virtualized linux is the right
one, then you can probably consider this request resolved.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-08-31 15:33

Message:
Logged In: NO 

this is a rather academic discussion, those who tackle this idea are going
to decide what tools and languages they use

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-08-31 12:29

Message:
Logged In: NO 

"
Most
mainstream scripting languages are well-supported on a rather wide number
of platforms, unlike C# development tools, which still require numerous
non-standard dependencies to be properly installed and set up for each
platform individually.
"

Sorry, but this isin't true! To develope using C# you can use mono to
compile the stuff you write on all supported platform, windows too. You can
write it with any ide/rad you like, naturally there are tools to do like
monodevelop on linux and sharpdevelop/visual studio (express) on windows.
This is the same for any scripting language: you need the runtime to
execute the script and any editor to write the stuff

I didn't see any kinda of non-standard dependencies:
- On windows, you have the runtime simply doing updates
- On linux, many major distros has it ((open)suse, ubuntu, fedora,
gentoo)

Another example: if you use GTK you need binding ... like any other
programming language that doesn't resides directly on the libraries (note:
technically you can compile in the bindings using a couple of mono tools)



----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-08-31 00:55

Message:
Logged In: NO 

I think, this frontend would be Win32-specific in any way, so it really
doesn't matter what language is used for implementation, however it may be
easier to favor an implementation language that is commonly and widely
known within the community of potential contributors, so that future
maintenance of such a frontend would not be limited to an unnecessarily
small group of contributor who are sufficiently familiar with the
corresponding language. Using a scripting language such as Python or Ruby
on the other hand, brings the added advantage that potential
contributors/maintainers of such a frontend would not necessarily have to
install any complex dependencies (think IDE, compilers, runtime libs). Most
mainstream scripting languages are well-supported on a rather wide number
of platforms, unlike C# development tools, which still require numerous
non-standard dependencies to be properly installed and set up for each
platform individually.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-08-29 14:16

Message:
Logged In: NO 

Just a note to the previous comment ...

if necessary you can use python too for development using ironpython!

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-08-29 14:14

Message:
Logged In: NO 

Python is awesome, i know, but why not C#?

Mono is distribuited with major linux systems and on windows you're fully
integrated with .NET! I think that a lot of users are windows based so a
fully native interface and a fully integrated application can do better
than a python application

Another really important stuff is that with .NET you can easely build two
interfaces and load the preferred at startup time simply checking the
operative system ... with the interfaces that relies on a set of
functionalities implemented in a backend


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-08-29 13:50

Message:
Logged In: YES 
user_id=579204
Originator: NO

If somebody has knowledges about Python, then here are some of GUI scripts
for coLinux configuration, that has to be finish:

http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/src/colinux/user/configurator/

It's very old. First needs to remove the outdated XML from that.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-08-29 11:26

Message:
Logged In: NO 

The following site contains a listing of available GUI toolkit bindings
for python, including a section about UI-building toolkits (e.g. dialog
editors):

http://wiki.python.org/moin/GuiProgramming

In addition, here's  a site providing a full tutorial about PyQT
programming:

http://www.commandprompt.com/community/pyqt/book1
http://wiki.python.org/moin/PyQt

PythonStudio includes a full GUI editor, as well:
http://www.codeplex.com/IronPythonStudio

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-08-29 01:57

Message:
Logged In: NO 

thinking about it, the most promising approach would probably be to simply
implement this totally separate, using a scripting language and a GUI
toolkit, i.e. something like python or ruby with GTK/QT bindings. This way,
it would not unnecessarily affect any of the existing code, would be very
much straight forward and also flexible and easily extendsible for future
changes

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-08-29 01:24

Message:
Logged In: NO 

something like this would probably be best implemented as a separate,
standaline "profile manager", so that creating/editing and deleting
profiles can be done via this app, and starting a profile would only entail
passing the stored parameters to the corresponding binaries, this would
keep the overhead to a minimum.



----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-08-29 01:03

Message:
Logged In: NO 

see "andLinux"

however, usability could still be somewhat improved for users with a
non-nix background, in particular the whole setup process should either be
more intuitive or simply provide some more GUI frontends, i.e. by allowing
the most common actions to be directly done from the executables via UI
means, rather than requiring users to pass lots of cryptic options via
command line parameters. In this context, it would already be a significant
improvement if the most basic and relevant parameters could be configured
via some sort of GUI frontend, be it embedded or not, so that running a
process without REQUIRED parameters automatically brings up a dialog for
users to configure the most important parameters, and optionally save them
as startup profiles, so that they may next time select a startup profile or
create a new one.

Similar to how firefox starts actually: provide a way for customization,
preferably using lots of nifty tooltips and explanations and once a user
has a working config, enable them to save and reuse configs

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622066&aid=1898496&group_id=98788

[coLinux-devel] autobuild org.colinux.devel 20081120 23:07:13

[coLinux a. <col...@he...>]

The autobuild system has detected a new revision in the source repository.
Review last changed from changelog.txt, also attached in mail.

Download the compiled version:
http://www.henrynestler.com/colinux/autobuild/devel-20081120/

 colinux-0.8.0-20081120.src.tgz  (694532 Bytes)
 daemons-0.8.0-20081120.dbg.zip  (595262 Bytes)
 daemons-0.8.0-20081120.zip  (480608 Bytes)

Note, the autobuild compilation does not include an installer.
Remember to reload the driver with these commands:

	colinux-daemon.exe --remove-driver
	colinux-daemon.exe --install-driver

The vmlinux and modules are up to date. Please use last version from
http://www.henrynestler.com/colinux/autobuild/devel-20081112/

The autobuild compilations are not official releases of Cooperative Linux
software.  There is no warranty that any autobuild version is stable.  If
use this autobuild version, please give us feedback of your experience.

Job runs on machine with 64 bit version of gcc 4.1.2.
A service from http://gcc.gnu.org/wiki/CompileFarm

-- 
Lots of fun with newest version,
Henry Nestler
------------------------------------------------------------------------
r1140 | henryn | 2008-11-20 23:07:13 +0000 (Thu, 20 Nov 2008) | 1 line
Changed paths:
   M /branches/devel/src/colinux/kernel/filesystem.c
   M /branches/devel/src/colinux/os/kernel/filesystem.h
   M /branches/devel/src/colinux/os/linux/kernel/module/fs.c
   M /branches/devel/src/colinux/os/winnt/kernel/fileio.c

* cofs: Create new files with desired mode from "creat(...,mode)"
------------------------------------------------------------------------
r1139 | henryn | 2008-11-20 22:40:48 +0000 (Thu, 20 Nov 2008) | 1 line
Changed paths:
   M /branches/devel/src/colinux/os/winnt/kernel/fileio.c

* cofs: Allows rename on readonly files.
------------------------------------------------------------------------

[coLinux-devel] [ colinux-Feature Requests-2081864 ] properly register coLinux as a service

[SourceForge.net <no...@so...>]

Feature Requests item #2081864, was opened at 2008-08-29 06:29
Message generated for change (Comment added) made by nobody
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622066&aid=2081864&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: properly register coLinux as a service 

Initial Comment:
if coLinux is fully registered as a service, it can be easily managed from the windows control panel, without necessarily requiring console commands to be executed manually by the user

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-11-16 14:13

Message:
having this would definitively improve the level of integration for colinux
under Windows

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-19 21:09

Message:
still installing it by default as a service that can be managed from the
control panel would be pretty intuitive

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-14 14:59

Message:
colinux already runs just fine as a service, at least it did the last time
I checked (before I moved to Vista x64).  There are a ton of different
programs to register services, I always used cygwin cygrunsrv to do it but
others like srvany from Microsoft should work too.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-09-21 17:07

Message:
this might actually be a good idea?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622066&aid=2081864&group_id=98788

Re: [coLinux-devel] autobuild org.colinux.devel 20081112 20:13:26

[Henry N. <hen...@ar...>]

Harald Weidner wrote:
> coLinux autobuild<col...@he...>:
>
>> Download the compiled version:
>> http://www.henrynestler.com/colinux/autobuild/devel-20081112/
>
> Sorry, but I can not access this directory.
>
> 404 Not Found

I know. The transfer from build bot to web space was broken.

You can use the yesterday snapshot, it's the same build:
http://www.colinux.org/snapshots/
http://www.henrynestler.com/colinux/testing/devel-0.8.0/20081112-Snapshot/

-- 
Henry N.

[coLinux-devel] autobuild org.colinux.devel 20081112 20:13:26

[coLinux a. <col...@he...>]

The autobuild system has detected a new revision in the source repository.
Review last changed from changelog.txt, also attached in mail.

Download the compiled version:
http://www.henrynestler.com/colinux/autobuild/devel-20081112/

 colinux-0.8.0-20081112.src.tgz  (694429 Bytes)
 daemons-0.8.0-20081112.dbg.zip  (595204 Bytes)
 daemons-0.8.0-20081112.zip  (480555 Bytes)
 modules-2.6.22.18-co-0.8.0-20081112.tgz  (2602819 Bytes)
 vmlinux-2.6.22.18-co-0.8.0-20081112.zip  (1762529 Bytes)

Note, the autobuild compilation does not include an installer.
Remember to reload the driver with these commands:

	colinux-daemon.exe --remove-driver
	colinux-daemon.exe --install-driver

Inside coLinux please update modules as follow:

	rm -rf /lib/modules/*-co-*
	tar -xzf modules-*-co-*-20081112.tgz -C /

The autobuild compilations are not official releases of Cooperative Linux
software.  There is no warranty that any autobuild version is stable.  If
use this autobuild version, please give us feedback of your experience.

Job runs on machine with 64 bit version of gcc 4.1.2.
A service from http://gcc.gnu.org/wiki/CompileFarm

-- 
Lots of fun with newest version,
Henry Nestler
------------------------------------------------------------------------
r1137 | henryn | 2008-11-12 20:13:26 +0000 (Wed, 12 Nov 2008) | 1 line
Changed paths:
   M /branches/devel/src/colinux/user/daemon.c

* daemon.c: Functions as static. Small source face lifts.
------------------------------------------------------------------------
r1136 | henryn | 2008-11-12 20:03:20 +0000 (Wed, 12 Nov 2008) | 1 line
Changed paths:
   M /branches/devel/src/colinux/common/config.h
   M /branches/devel/src/colinux/kernel/monitor.c
   M /branches/devel/src/colinux/user/daemon.c

* Add size check to struct co_config. Remove unused var block_root_device_index.
------------------------------------------------------------------------
r1135 | henryn | 2008-11-12 19:57:45 +0000 (Wed, 12 Nov 2008) | 1 line
Changed paths:
   M /branches/devel/src/colinux/user/config.c

* Wrong const used for checking parameter "videdoX=" (MAX_COVIDEO).
------------------------------------------------------------------------
r1134 | henryn | 2008-11-12 19:23:31 +0000 (Wed, 12 Nov 2008) | 3 lines
Changed paths:
   M /branches/devel/NEWS
   M /branches/devel/conf/example.conf
   M /branches/devel/doc/colinux-daemon
   M /branches/devel/src/colinux/common/common_base.h
   M /branches/devel/src/colinux/common/config.h
   M /branches/devel/src/colinux/common/console.c
   M /branches/devel/src/colinux/kernel/monitor.c
   M /branches/devel/src/colinux/user/config.c

* Added "cocon=COLSxROWS" to define console size, other than 80x25.
* Incrase periphery API version to 22, to reflect the change.

------------------------------------------------------------------------
r1133 | henryn | 2008-11-12 18:52:02 +0000 (Wed, 12 Nov 2008) | 1 line
Changed paths:
   M /branches/devel/NEWS
   M /branches/devel/patch/cofs-core.diff

* Bugfix #2176188: File sync wrong on cofs, command after mv/rename fails.
------------------------------------------------------------------------

[coLinux-devel] [ colinux-Bugs-2176188 ] file sync wrong on cofs, command after mv/rename fails

[SourceForge.net <no...@so...>]

Bugs item #2176188, was opened at 2008-10-18 09:20
Message generated for change (Comment added) made by henryn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
>Resolution: Fixed
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
>Assigned to: Henry N. (henryn)
Summary: file sync wrong on cofs, command after mv/rename fails

Initial Comment:
I found some problem in cofs filesystem.
ex:
mv t1 t2
stat t2

many times I got "t2 not found".
it see like the second command exec before the first finished.


----------------------------------------------------------------------

>Comment By: Henry N. (henryn)
Date: 2008-11-12 19:59

Message:
Your patch was committed to SVN as revision r1133.
Thanks!


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-11-12 16:28

Message:
chenm001 (chenm003 AT 163 DOT com)
if this patch has any bug, please tell me, I will tract again.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-11-12 01:26

Message:
Yes, that solved this bug. NFS does it too. So it it very good. Thanks.
What name I can honor in SVN? :-)

Henry


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-11-11 10:19

Message:
Here is my patch, I do not known it is good way, but it work fine in my
computer.

static int fuse_rename(struct inode *olddir, struct dentry *oldent,
		       struct inode *newdir, struct dentry *newent)
{
	struct fuse_conn *fc = INO_FC(olddir);
	struct fuse_in in = FUSE_IN_INIT;
	struct fuse_out out = FUSE_OUT_INIT;
	struct fuse_rename_in inarg;

	memset(&inarg, 0, sizeof(inarg));
	inarg.newdir = newdir->i_ino;

	in.h.opcode = FUSE_RENAME;
	in.h.ino = olddir->i_ino;
	in.numargs = 3;
	in.args[0].size = sizeof(inarg);
	in.args[0].value = &inarg;
	in.args[1].size = oldent->d_name.len + 1;
	in.args[1].value = oldent->d_name.name;
	in.args[2].size = newent->d_name.len + 1;
	in.args[2].value = newent->d_name.name;
	request_send(fc, &in, &out);

	if (!out.h.error) {
		uncache_dir(olddir);
+		d_move(oldent, newent);
		if (olddir != newdir)
			uncache_dir(newdir);
	}

	return out.h.error;
}


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-11-06 22:17

Message:
Yes, SVN r1127 corrected only the swapped atime and mtime, and not the
cache problem on rename.

You have a patch? Nice. Please let's see.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-29 04:04

Message:
Unfortunately, SVN revision r1127 do not fix this bug.
When I remove my "d_move" patch, problem was recurrence.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-28 00:24

Message:
With these script the "touch -d 1970-1-1 b" after the "mv" fails all
times:
  rm -f a b
  touch a
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

Same problem with overwrite on move:
  touch a b
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

On cofs the touch after "mv" fails:
touch: setting times of `b': No such file or directory
-rw-r--r-- 1 root root 0 2008-10-27 23:29:50.000000000 +0000 b

On cobd or with 'nocache' it works:
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 b


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-23 03:18

Message:
Close inode cache possible loss performance, it is a little bug in rename
only.
I fix this bug yesterday. and I will use the my new kernel some days,
if that has other problem, I can fix it.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-22 13:41

Message:
The inode cache is default enabled for 1 Second. You can disable it by
adding "nocache" to mount options, for example:

  mount -o nocache -t cofs 0 /mnt/windows

More cosf options, see cofs.txt in your coLinux installation or here:
 
http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/doc/cofs

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-21 13:48

Message:
I trace the code, it see like rename bug, the filename change, but inode
not.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-18 09:24

Message:
Excuse me, I think this bug same as #2152550

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

[coLinux-devel] [ colinux-Bugs-2176188 ] file sync wrong on cofs, command after mv/rename fails

[SourceForge.net <no...@so...>]

Bugs item #2176188, was opened at 2008-10-18 07:20
Message generated for change (Comment added) made by nobody
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: file sync wrong on cofs, command after mv/rename fails

Initial Comment:
I found some problem in cofs filesystem.
ex:
mv t1 t2
stat t2

many times I got "t2 not found".
it see like the second command exec before the first finished.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-11-12 15:28

Message:
chenm001 (chenm003 AT 163 DOT com)
if this patch has any bug, please tell me, I will tract again.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-11-12 00:26

Message:
Yes, that solved this bug. NFS does it too. So it it very good. Thanks.
What name I can honor in SVN? :-)

Henry


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-11-11 09:19

Message:
Here is my patch, I do not known it is good way, but it work fine in my
computer.

static int fuse_rename(struct inode *olddir, struct dentry *oldent,
		       struct inode *newdir, struct dentry *newent)
{
	struct fuse_conn *fc = INO_FC(olddir);
	struct fuse_in in = FUSE_IN_INIT;
	struct fuse_out out = FUSE_OUT_INIT;
	struct fuse_rename_in inarg;

	memset(&inarg, 0, sizeof(inarg));
	inarg.newdir = newdir->i_ino;

	in.h.opcode = FUSE_RENAME;
	in.h.ino = olddir->i_ino;
	in.numargs = 3;
	in.args[0].size = sizeof(inarg);
	in.args[0].value = &inarg;
	in.args[1].size = oldent->d_name.len + 1;
	in.args[1].value = oldent->d_name.name;
	in.args[2].size = newent->d_name.len + 1;
	in.args[2].value = newent->d_name.name;
	request_send(fc, &in, &out);

	if (!out.h.error) {
		uncache_dir(olddir);
+		d_move(oldent, newent);
		if (olddir != newdir)
			uncache_dir(newdir);
	}

	return out.h.error;
}


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-11-06 21:17

Message:
Yes, SVN r1127 corrected only the swapped atime and mtime, and not the
cache problem on rename.

You have a patch? Nice. Please let's see.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-29 03:04

Message:
Unfortunately, SVN revision r1127 do not fix this bug.
When I remove my "d_move" patch, problem was recurrence.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-27 23:24

Message:
With these script the "touch -d 1970-1-1 b" after the "mv" fails all
times:
  rm -f a b
  touch a
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

Same problem with overwrite on move:
  touch a b
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

On cofs the touch after "mv" fails:
touch: setting times of `b': No such file or directory
-rw-r--r-- 1 root root 0 2008-10-27 23:29:50.000000000 +0000 b

On cobd or with 'nocache' it works:
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 b


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-23 01:18

Message:
Close inode cache possible loss performance, it is a little bug in rename
only.
I fix this bug yesterday. and I will use the my new kernel some days,
if that has other problem, I can fix it.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-22 11:41

Message:
The inode cache is default enabled for 1 Second. You can disable it by
adding "nocache" to mount options, for example:

  mount -o nocache -t cofs 0 /mnt/windows

More cosf options, see cofs.txt in your coLinux installation or here:
 
http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/doc/cofs

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-21 11:48

Message:
I trace the code, it see like rename bug, the filename change, but inode
not.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-18 07:24

Message:
Excuse me, I think this bug same as #2152550

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

[coLinux-devel] [ colinux-Bugs-2176188 ] file sync wrong on cofs, command after mv/rename fails

[SourceForge.net <no...@so...>]

Bugs item #2176188, was opened at 2008-10-18 09:20
Message generated for change (Comment added) made by henryn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: file sync wrong on cofs, command after mv/rename fails

Initial Comment:
I found some problem in cofs filesystem.
ex:
mv t1 t2
stat t2

many times I got "t2 not found".
it see like the second command exec before the first finished.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-11-12 01:26

Message:
Yes, that solved this bug. NFS does it too. So it it very good. Thanks.
What name I can honor in SVN? :-)

Henry


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-11-11 10:19

Message:
Here is my patch, I do not known it is good way, but it work fine in my
computer.

static int fuse_rename(struct inode *olddir, struct dentry *oldent,
		       struct inode *newdir, struct dentry *newent)
{
	struct fuse_conn *fc = INO_FC(olddir);
	struct fuse_in in = FUSE_IN_INIT;
	struct fuse_out out = FUSE_OUT_INIT;
	struct fuse_rename_in inarg;

	memset(&inarg, 0, sizeof(inarg));
	inarg.newdir = newdir->i_ino;

	in.h.opcode = FUSE_RENAME;
	in.h.ino = olddir->i_ino;
	in.numargs = 3;
	in.args[0].size = sizeof(inarg);
	in.args[0].value = &inarg;
	in.args[1].size = oldent->d_name.len + 1;
	in.args[1].value = oldent->d_name.name;
	in.args[2].size = newent->d_name.len + 1;
	in.args[2].value = newent->d_name.name;
	request_send(fc, &in, &out);

	if (!out.h.error) {
		uncache_dir(olddir);
+		d_move(oldent, newent);
		if (olddir != newdir)
			uncache_dir(newdir);
	}

	return out.h.error;
}


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-11-06 22:17

Message:
Yes, SVN r1127 corrected only the swapped atime and mtime, and not the
cache problem on rename.

You have a patch? Nice. Please let's see.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-29 04:04

Message:
Unfortunately, SVN revision r1127 do not fix this bug.
When I remove my "d_move" patch, problem was recurrence.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-28 00:24

Message:
With these script the "touch -d 1970-1-1 b" after the "mv" fails all
times:
  rm -f a b
  touch a
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

Same problem with overwrite on move:
  touch a b
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

On cofs the touch after "mv" fails:
touch: setting times of `b': No such file or directory
-rw-r--r-- 1 root root 0 2008-10-27 23:29:50.000000000 +0000 b

On cobd or with 'nocache' it works:
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 b


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-23 03:18

Message:
Close inode cache possible loss performance, it is a little bug in rename
only.
I fix this bug yesterday. and I will use the my new kernel some days,
if that has other problem, I can fix it.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-22 13:41

Message:
The inode cache is default enabled for 1 Second. You can disable it by
adding "nocache" to mount options, for example:

  mount -o nocache -t cofs 0 /mnt/windows

More cosf options, see cofs.txt in your coLinux installation or here:
 
http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/doc/cofs

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-21 13:48

Message:
I trace the code, it see like rename bug, the filename change, but inode
not.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-18 09:24

Message:
Excuse me, I think this bug same as #2152550

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

[coLinux-devel] [ colinux-Bugs-2176188 ] file sync wrong on cofs, command after mv/rename fails

[SourceForge.net <no...@so...>]

Bugs item #2176188, was opened at 2008-10-18 07:20
Message generated for change (Comment added) made by nobody
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: file sync wrong on cofs, command after mv/rename fails

Initial Comment:
I found some problem in cofs filesystem.
ex:
mv t1 t2
stat t2

many times I got "t2 not found".
it see like the second command exec before the first finished.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-11-11 09:19

Message:
Here is my patch, I do not known it is good way, but it work fine in my
computer.

static int fuse_rename(struct inode *olddir, struct dentry *oldent,
		       struct inode *newdir, struct dentry *newent)
{
	struct fuse_conn *fc = INO_FC(olddir);
	struct fuse_in in = FUSE_IN_INIT;
	struct fuse_out out = FUSE_OUT_INIT;
	struct fuse_rename_in inarg;

	memset(&inarg, 0, sizeof(inarg));
	inarg.newdir = newdir->i_ino;

	in.h.opcode = FUSE_RENAME;
	in.h.ino = olddir->i_ino;
	in.numargs = 3;
	in.args[0].size = sizeof(inarg);
	in.args[0].value = &inarg;
	in.args[1].size = oldent->d_name.len + 1;
	in.args[1].value = oldent->d_name.name;
	in.args[2].size = newent->d_name.len + 1;
	in.args[2].value = newent->d_name.name;
	request_send(fc, &in, &out);

	if (!out.h.error) {
		uncache_dir(olddir);
+		d_move(oldent, newent);
		if (olddir != newdir)
			uncache_dir(newdir);
	}

	return out.h.error;
}


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-11-06 21:17

Message:
Yes, SVN r1127 corrected only the swapped atime and mtime, and not the
cache problem on rename.

You have a patch? Nice. Please let's see.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-29 03:04

Message:
Unfortunately, SVN revision r1127 do not fix this bug.
When I remove my "d_move" patch, problem was recurrence.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-27 23:24

Message:
With these script the "touch -d 1970-1-1 b" after the "mv" fails all
times:
  rm -f a b
  touch a
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

Same problem with overwrite on move:
  touch a b
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

On cofs the touch after "mv" fails:
touch: setting times of `b': No such file or directory
-rw-r--r-- 1 root root 0 2008-10-27 23:29:50.000000000 +0000 b

On cobd or with 'nocache' it works:
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 b


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-23 01:18

Message:
Close inode cache possible loss performance, it is a little bug in rename
only.
I fix this bug yesterday. and I will use the my new kernel some days,
if that has other problem, I can fix it.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-22 11:41

Message:
The inode cache is default enabled for 1 Second. You can disable it by
adding "nocache" to mount options, for example:

  mount -o nocache -t cofs 0 /mnt/windows

More cosf options, see cofs.txt in your coLinux installation or here:
 
http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/doc/cofs

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-21 11:48

Message:
I trace the code, it see like rename bug, the filename change, but inode
not.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-18 07:24

Message:
Excuse me, I think this bug same as #2152550

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

[coLinux-devel] [ colinux-Bugs-2176188 ] file sync wrong on cofs, command after mv/rename fails

[SourceForge.net <no...@so...>]

Bugs item #2176188, was opened at 2008-10-18 09:20
Message generated for change (Comment added) made by henryn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: file sync wrong on cofs, command after mv/rename fails

Initial Comment:
I found some problem in cofs filesystem.
ex:
mv t1 t2
stat t2

many times I got "t2 not found".
it see like the second command exec before the first finished.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-11-06 22:17

Message:
Yes, SVN r1127 corrected only the swapped atime and mtime, and not the
cache problem on rename.

You have a patch? Nice. Please let's see.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-29 04:04

Message:
Unfortunately, SVN revision r1127 do not fix this bug.
When I remove my "d_move" patch, problem was recurrence.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-28 00:24

Message:
With these script the "touch -d 1970-1-1 b" after the "mv" fails all
times:
  rm -f a b
  touch a
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

Same problem with overwrite on move:
  touch a b
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

On cofs the touch after "mv" fails:
touch: setting times of `b': No such file or directory
-rw-r--r-- 1 root root 0 2008-10-27 23:29:50.000000000 +0000 b

On cobd or with 'nocache' it works:
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 b


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-23 03:18

Message:
Close inode cache possible loss performance, it is a little bug in rename
only.
I fix this bug yesterday. and I will use the my new kernel some days,
if that has other problem, I can fix it.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-22 13:41

Message:
The inode cache is default enabled for 1 Second. You can disable it by
adding "nocache" to mount options, for example:

  mount -o nocache -t cofs 0 /mnt/windows

More cosf options, see cofs.txt in your coLinux installation or here:
 
http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/doc/cofs

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-21 13:48

Message:
I trace the code, it see like rename bug, the filename change, but inode
not.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-18 09:24

Message:
Excuse me, I think this bug same as #2152550

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

[coLinux-devel] [ colinux-Bugs-2176188 ] file sync wrong on cofs, command after mv/rename fails

[SourceForge.net <no...@so...>]

Bugs item #2176188, was opened at 2008-10-18 07:20
Message generated for change (Comment added) made by nobody
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: file sync wrong on cofs, command after mv/rename fails

Initial Comment:
I found some problem in cofs filesystem.
ex:
mv t1 t2
stat t2

many times I got "t2 not found".
it see like the second command exec before the first finished.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-29 03:04

Message:
Unfortunately, SVN revision r1127 do not fix this bug.
When I remove my "d_move" patch, problem was recurrence.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-27 23:24

Message:
With these script the "touch -d 1970-1-1 b" after the "mv" fails all
times:
  rm -f a b
  touch a
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

Same problem with overwrite on move:
  touch a b
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

On cofs the touch after "mv" fails:
touch: setting times of `b': No such file or directory
-rw-r--r-- 1 root root 0 2008-10-27 23:29:50.000000000 +0000 b

On cobd or with 'nocache' it works:
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 b


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-23 01:18

Message:
Close inode cache possible loss performance, it is a little bug in rename
only.
I fix this bug yesterday. and I will use the my new kernel some days,
if that has other problem, I can fix it.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-22 11:41

Message:
The inode cache is default enabled for 1 Second. You can disable it by
adding "nocache" to mount options, for example:

  mount -o nocache -t cofs 0 /mnt/windows

More cosf options, see cofs.txt in your coLinux installation or here:
 
http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/doc/cofs

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-21 11:48

Message:
I trace the code, it see like rename bug, the filename change, but inode
not.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-18 07:24

Message:
Excuse me, I think this bug same as #2152550

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

[coLinux-devel] [ colinux-Bugs-2176188 ] file sync wrong on cofs, command after mv/rename fails

[SourceForge.net <no...@so...>]

Bugs item #2176188, was opened at 2008-10-18 09:20
Message generated for change (Comment added) made by henryn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
>Summary: file sync wrong on cofs, command after mv/rename fails

Initial Comment:
I found some problem in cofs filesystem.
ex:
mv t1 t2
stat t2

many times I got "t2 not found".
it see like the second command exec before the first finished.


----------------------------------------------------------------------

>Comment By: Henry N. (henryn)
Date: 2008-10-28 00:24

Message:
With these script the "touch -d 1970-1-1 b" after the "mv" fails all
times:
  rm -f a b
  touch a
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

Same problem with overwrite on move:
  touch a b
  mv a b
  touch -d 1970-1-1 b
  ls --full-time b

On cofs the touch after "mv" fails:
touch: setting times of `b': No such file or directory
-rw-r--r-- 1 root root 0 2008-10-27 23:29:50.000000000 +0000 b

On cobd or with 'nocache' it works:
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 b


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-23 03:18

Message:
Close inode cache possible loss performance, it is a little bug in rename
only.
I fix this bug yesterday. and I will use the my new kernel some days,
if that has other problem, I can fix it.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-22 13:41

Message:
The inode cache is default enabled for 1 Second. You can disable it by
adding "nocache" to mount options, for example:

  mount -o nocache -t cofs 0 /mnt/windows

More cosf options, see cofs.txt in your coLinux installation or here:
 
http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/doc/cofs

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-21 13:48

Message:
I trace the code, it see like rename bug, the filename change, but inode
not.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-18 09:24

Message:
Excuse me, I think this bug same as #2152550

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

[coLinux-devel] [ colinux-Bugs-2152550 ] rsync cannot set mtime correctly on cofs

[SourceForge.net <no...@so...>]

Bugs item #2152550, was opened at 2008-10-08 05:29
Message generated for change (Settings changed) made by henryn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2152550&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
>Resolution: Fixed
Priority: 5
Private: No
Submitted By: Xuefer (xuefer)
>Assigned to: Henry N. (henryn)
Summary: rsync cannot set mtime correctly on cofs

Initial Comment:
$ cd ~
(native mount, non-cofs, good result)
$ rm -f a b; touch -d 1970-1-1 a ; rsync -a a b;  ls --full-time a b ; rm -f a b-rw-r--r-- 1 xuefer xuefer 0 1970-01-01 00:00:00.000000000 +0800 a
-rw-r--r-- 1 xuefer xuefer 0 1970-01-01 00:00:00.000000000 +0800 b

$ cd /c/tmp
(cofs mount, wrong result)
$ rm -f a b; touch -d 1970-1-1 a ; rsync -a a b;  ls --full-time a b ; rm -f a b-rwxr-xr-x 1 xuefer users 0 1970-01-01 00:00:00.000000000 +0800 a*
-rwxr-xr-x 1 xuefer users 0 2008-10-08 11:29:22.000000000 +0800 b*


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-27 23:55

Message:
This bug is fixed by SVN revision r1127 and has nothing to do with nocache
and Bug #2176188.
The parameter for set atime and mtime are swapped by wrong variable names.
The current snapshot http://www.colinux.org/snapshots/ includes this
change.

For rsync test have added "stat a b", this shows the swapped times. The
test script:
  cd /mnt/temp
  rm -f a b
  sleep 1
  touch -d 1970-1-1 a
  rsync -a a b
  ls --full-time a b
  stat a b

Before  r1127 (non interesting lines removed):
-rwxr-xr-x 1 root root 0 1970-01-01 00:00:00.000000000 +0000 a
-rwxr-xr-x 1 root root 0 2008-10-27 22:31:25.000000000 +0000 b
  File: `a'
Access: 1970-01-01 00:00:00.000000000 +0000
Modify: 1970-01-01 00:00:00.000000000 +0000
Change: 2008-10-27 22:31:27.000000000 +0000
  File: `b'
Access: 1970-01-01 00:00:00.000000000 +0000
Modify: 2008-10-27 22:31:25.000000000 +0000
Change: 2008-10-27 22:31:27.000000000 +0000

After  r1127 (non interesting lines removed):
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 a
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 b
  File: `a'
Access: 1970-01-01 00:00:00.000000000 +0000
Modify: 1970-01-01 00:00:00.000000000 +0000
Change: 2008-10-27 22:37:22.000000000 +0000
  File: `b'
Access: 2008-10-27 22:37:22.000000000 +0000
Modify: 1970-01-01 00:00:00.000000000 +0000
Change: 2008-10-27 22:37:22.000000000 +0000

The same script on /tmp (cobd), have same times:
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 a
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 b
  File: `a'
Access: 1970-01-01 00:00:00.000000000 +0000
Modify: 1970-01-01 00:00:00.000000000 +0000
Change: 2008-10-27 22:56:18.000000000 +0000
  File: `b'
Access: 2008-10-27 22:56:18.000000000 +0000
Modify: 1970-01-01 00:00:00.000000000 +0000
Change: 2008-10-27 22:56:18.000000000 +0000


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-26 21:55

Message:
This is a duplicate of bug# 2176188 "file sync wrong on cofs".

Rsync creates temporally files like ".b..b8jlLdx". Than rsync tries to
change attributes after rename.
The rename is not complete from async file close, so the file can't reopen
shortly for time modifications.

As workaround you can add "nocache" to the mount options, then it works.


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2152550&group_id=98788

[coLinux-devel] [ colinux-Bugs-2152550 ] rsync cannot set mtime correctly on cofs

[SourceForge.net <no...@so...>]

Bugs item #2152550, was opened at 2008-10-08 05:29
Message generated for change (Comment added) made by henryn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2152550&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Xuefer (xuefer)
Assigned to: Nobody/Anonymous (nobody)
Summary: rsync cannot set mtime correctly on cofs

Initial Comment:
$ cd ~
(native mount, non-cofs, good result)
$ rm -f a b; touch -d 1970-1-1 a ; rsync -a a b;  ls --full-time a b ; rm -f a b-rw-r--r-- 1 xuefer xuefer 0 1970-01-01 00:00:00.000000000 +0800 a
-rw-r--r-- 1 xuefer xuefer 0 1970-01-01 00:00:00.000000000 +0800 b

$ cd /c/tmp
(cofs mount, wrong result)
$ rm -f a b; touch -d 1970-1-1 a ; rsync -a a b;  ls --full-time a b ; rm -f a b-rwxr-xr-x 1 xuefer users 0 1970-01-01 00:00:00.000000000 +0800 a*
-rwxr-xr-x 1 xuefer users 0 2008-10-08 11:29:22.000000000 +0800 b*


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-27 23:55

Message:
This bug is fixed by SVN revision r1127 and has nothing to do with nocache
and Bug #2176188.
The parameter for set atime and mtime are swapped by wrong variable names.
The current snapshot http://www.colinux.org/snapshots/ includes this
change.

For rsync test have added "stat a b", this shows the swapped times. The
test script:
  cd /mnt/temp
  rm -f a b
  sleep 1
  touch -d 1970-1-1 a
  rsync -a a b
  ls --full-time a b
  stat a b

Before  r1127 (non interesting lines removed):
-rwxr-xr-x 1 root root 0 1970-01-01 00:00:00.000000000 +0000 a
-rwxr-xr-x 1 root root 0 2008-10-27 22:31:25.000000000 +0000 b
  File: `a'
Access: 1970-01-01 00:00:00.000000000 +0000
Modify: 1970-01-01 00:00:00.000000000 +0000
Change: 2008-10-27 22:31:27.000000000 +0000
  File: `b'
Access: 1970-01-01 00:00:00.000000000 +0000
Modify: 2008-10-27 22:31:25.000000000 +0000
Change: 2008-10-27 22:31:27.000000000 +0000

After  r1127 (non interesting lines removed):
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 a
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 b
  File: `a'
Access: 1970-01-01 00:00:00.000000000 +0000
Modify: 1970-01-01 00:00:00.000000000 +0000
Change: 2008-10-27 22:37:22.000000000 +0000
  File: `b'
Access: 2008-10-27 22:37:22.000000000 +0000
Modify: 1970-01-01 00:00:00.000000000 +0000
Change: 2008-10-27 22:37:22.000000000 +0000

The same script on /tmp (cobd), have same times:
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 a
-rw-r--r-- 1 root root 0 1970-01-01 00:00:00.000000000 +0000 b
  File: `a'
Access: 1970-01-01 00:00:00.000000000 +0000
Modify: 1970-01-01 00:00:00.000000000 +0000
Change: 2008-10-27 22:56:18.000000000 +0000
  File: `b'
Access: 2008-10-27 22:56:18.000000000 +0000
Modify: 1970-01-01 00:00:00.000000000 +0000
Change: 2008-10-27 22:56:18.000000000 +0000


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-26 21:55

Message:
This is a duplicate of bug# 2176188 "file sync wrong on cofs".

Rsync creates temporally files like ".b..b8jlLdx". Than rsync tries to
change attributes after rename.
The rename is not complete from async file close, so the file can't reopen
shortly for time modifications.

As workaround you can add "nocache" to the mount options, then it works.


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2152550&group_id=98788

[coLinux-devel] autobuild org.colinux.devel 20081026 23:21:59

[coLinux a. <col...@he...>]

The autobuild system has detected a new revision in the source repository.
Review last changed from changelog.txt, also attached in mail.

Download the compiled version:
http://www.henrynestler.com/colinux/autobuild/devel-20081026/

 colinux-0.8.0-20081026.src.tgz  (693982 Bytes)
 daemons-0.8.0-20081026.dbg.zip  (595049 Bytes)
 daemons-0.8.0-20081026.zip  (480338 Bytes)
 modules-2.6.22.18-co-0.8.0-20081026.tgz  (2602717 Bytes)
 vmlinux-2.6.22.18-co-0.8.0-20081026.zip  (1762566 Bytes)

Note, the autobuild compilation does not include an installer.
Remember to reload the driver with these commands:

	colinux-daemon.exe --remove-driver
	colinux-daemon.exe --install-driver

Inside coLinux please update modules as follow:

	rm -rf /lib/modules/*-co-*
	tar -xzf modules-*-co-*-20081026.tgz -C /

The autobuild compilations are not official releases of Cooperative Linux
software.  There is no warranty that any autobuild version is stable.  If
use this autobuild version, please give us feedback of your experience.

Job runs on machine with 64 bit version of gcc 4.1.2.
A service from http://gcc.gnu.org/wiki/CompileFarm

-- 
Lots of fun with newest version,
Henry Nestler
------------------------------------------------------------------------
r1132 | henryn | 2008-10-26 23:21:59 +0000 (Sun, 26 Oct 2008) | 1 line
Changed paths:
   M /branches/devel/patch/base-2.6.22.diff
   M /branches/devel/patch/pci-core.diff

* PCI: Remove e820.o from object list. Add only pci_mem_start as dummy.
------------------------------------------------------------------------
r1131 | henryn | 2008-10-26 23:05:46 +0000 (Sun, 26 Oct 2008) | 1 line
Changed paths:
   M /branches/devel/patch/base-2.6.22.diff
   M /branches/devel/patch/pci-core.diff

* Kernel: Move CO_DEVICE_PCI from pci-core.diff to base-2.6.22.diff
------------------------------------------------------------------------
r1130 | henryn | 2008-10-26 22:59:18 +0000 (Sun, 26 Oct 2008) | 1 line
Changed paths:
   M /branches/devel/patch/base-2.6.22.diff
   M /branches/devel/patch/timer-2.6.22.diff
   M /branches/devel/src/colinux/kernel/filesystem.c

* Remove warnings from unused variable and a unused return.
------------------------------------------------------------------------
r1129 | henryn | 2008-10-26 22:43:57 +0000 (Sun, 26 Oct 2008) | 1 line
Changed paths:
   M /branches/devel/src/colinux/os/winnt/kernel/block.c

* Fixed Bug #2013435: cobd supports read-only files now.
------------------------------------------------------------------------
r1128 | henryn | 2008-10-26 22:39:16 +0000 (Sun, 26 Oct 2008) | 1 line
Changed paths:
   M /branches/devel/src/colinux/os/winnt/kernel/block.c

* co_os_file_block_(async_)open: Use pointer directly. Preparation for RO access.
------------------------------------------------------------------------
r1127 | henryn | 2008-10-26 22:31:58 +0000 (Sun, 26 Oct 2008) | 1 line
Changed paths:
   M /branches/devel/NEWS
   M /branches/devel/src/colinux/os/winnt/kernel/fileio.c

* cofs: Bugfix atime and mtime, was swapped on time set.
------------------------------------------------------------------------
r1126 | henryn | 2008-10-26 22:16:51 +0000 (Sun, 26 Oct 2008) | 1 line
Changed paths:
   M /branches/devel/NEWS
   M /branches/devel/doc/cofs
   M /branches/devel/patch/cofs-core.diff
   M /branches/devel/src/colinux/kernel/filesystem.c
   M /branches/devel/src/colinux/kernel/filesystem.h

* cofs: New mount option "noattrib" disables UNIX/FAT-attribute mapping.
------------------------------------------------------------------------
r1125 | henryn | 2008-10-26 21:54:08 +0000 (Sun, 26 Oct 2008) | 1 line
Changed paths:
   M /branches/devel/NEWS
   M /branches/devel/doc/cofs
   M /branches/devel/src/colinux/kernel/filesystem.c
   M /branches/devel/src/colinux/os/winnt/kernel/fileio.c

* cofs: Add flat file modes rwx for owner.
------------------------------------------------------------------------
r1124 | henryn | 2008-10-26 21:18:44 +0000 (Sun, 26 Oct 2008) | 1 line
Changed paths:
   M /branches/devel/src/colinux/kernel/filesystem.c
   M /branches/devel/src/colinux/os/kernel/filesystem.h
   M /branches/devel/src/colinux/os/linux/kernel/module/fs.c
   M /branches/devel/src/colinux/os/winnt/kernel/filesystem.c

* co_os_fs_get_attr: Moved to co_fs_get_attr as OS independent source.
------------------------------------------------------------------------

[coLinux-devel] [ colinux-Bugs-2013435 ] cobd does not work with read-only files

[SourceForge.net <no...@so...>]

Bugs item #2013435, was opened at 2008-07-08 14:22
Message generated for change (Comment added) made by henryn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2013435&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.7.x (release)
Status: Open
>Resolution: Fixed
Priority: 5
Private: No
Submitted By: Adrien Beau (abeau)
>Assigned to: Henry N. (henryn)
Summary: cobd does not work with read-only files

Initial Comment:
I have created an ISO file with the Windows port of cdrtools, and set it read-only (an old habit).

I then tried to mount it in the guest OS (a Slackware 12.1 installation initrd that I patched to include coLinux devices and kernel modules):

~# mount /dev/hdc /cdrom
mount: you must specify the filesystem type
~# mount -t iso9660 /dev/hdc /cdrom
mount: /dev/hdc: can't read superblock

After solving other problems with my hda and hdb devices (it turns out I was trying to use whole-disk images instead of filesystem images) and discovering hdc had a different problem after all, I was a bit lucky to try first to remove the read-only flag from the ISO file in Windows. It solved my problem immediately:

~# mount /dev/hdc /cdrom
~# mount
proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
/dev/hdc on /var/log/mount type iso9660 (rw)

(Note that in the Slackware installation initrd, /cdrom is a symlink to /var/log/mount.)

My configuration, in case I missed a parameter (having read config.c, I doubt it):

# --- configuration start ---
kernel="C:\Linux\colinux\vmlinux"
hda1="C:\Linux\slackware\colinux\slackware-12.1-colinux.partition"
hdb1="C:\Linux\slackware\colinux\swap.partition"
hdc="C:\Linux\slackware\colinux\install\slackware-12.1.iso"
initrd="C:\Linux\slackware\colinux\install\initrd.img"
root=/dev/hda1
rw
# --- configuration end ---

This is a minor bug, but I believe the current behavior should be documented at the very least.

----------------------------------------------------------------------

>Comment By: Henry N. (henryn)
Date: 2008-10-26 23:48

Message:
Thanks for this report.

Read only mount for such files are supported in next coLinux version now.
Committed as SVN revision r1129 for the devel version 0.8.0 and available
from next autobuild.


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2013435&group_id=98788

[coLinux-devel] [ colinux-Bugs-2152550 ] rsync cannot set mtime correctly on cofs

[SourceForge.net <no...@so...>]

Bugs item #2152550, was opened at 2008-10-08 05:29
Message generated for change (Comment added) made by henryn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2152550&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Xuefer (xuefer)
Assigned to: Nobody/Anonymous (nobody)
Summary: rsync cannot set mtime correctly on cofs

Initial Comment:
$ cd ~
(native mount, non-cofs, good result)
$ rm -f a b; touch -d 1970-1-1 a ; rsync -a a b;  ls --full-time a b ; rm -f a b-rw-r--r-- 1 xuefer xuefer 0 1970-01-01 00:00:00.000000000 +0800 a
-rw-r--r-- 1 xuefer xuefer 0 1970-01-01 00:00:00.000000000 +0800 b

$ cd /c/tmp
(cofs mount, wrong result)
$ rm -f a b; touch -d 1970-1-1 a ; rsync -a a b;  ls --full-time a b ; rm -f a b-rwxr-xr-x 1 xuefer users 0 1970-01-01 00:00:00.000000000 +0800 a*
-rwxr-xr-x 1 xuefer users 0 2008-10-08 11:29:22.000000000 +0800 b*


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-26 21:55

Message:
This is a duplicate of bug# 2176188 "file sync wrong on cofs".

Rsync creates temporally files like ".b..b8jlLdx". Than rsync tries to
change attributes after rename.
The rename is not complete from async file close, so the file can't reopen
shortly for time modifications.

As workaround you can add "nocache" to the mount options, then it works.


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2152550&group_id=98788

[coLinux-devel] [ colinux-Bugs-2176188 ] file sync wrong on cofs

[SourceForge.net <no...@so...>]

Bugs item #2176188, was opened at 2008-10-18 07:20
Message generated for change (Comment added) made by nobody
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: file sync wrong on cofs

Initial Comment:
I found some problem in cofs filesystem.
ex:
mv t1 t2
stat t2

many times I got "t2 not found".
it see like the second command exec before the first finished.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-23 01:18

Message:
Close inode cache possible loss performance, it is a little bug in rename
only.
I fix this bug yesterday. and I will use the my new kernel some days,
if that has other problem, I can fix it.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-22 11:41

Message:
The inode cache is default enabled for 1 Second. You can disable it by
adding "nocache" to mount options, for example:

  mount -o nocache -t cofs 0 /mnt/windows

More cosf options, see cofs.txt in your coLinux installation or here:
 
http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/doc/cofs

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-21 11:48

Message:
I trace the code, it see like rename bug, the filename change, but inode
not.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-18 07:24

Message:
Excuse me, I think this bug same as #2152550

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

[coLinux-devel] [ colinux-Bugs-2176188 ] file sync wrong on cofs

[SourceForge.net <no...@so...>]

Bugs item #2176188, was opened at 2008-10-18 09:20
Message generated for change (Comment added) made by henryn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: v0.8.x (devel)
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: file sync wrong on cofs

Initial Comment:
I found some problem in cofs filesystem.
ex:
mv t1 t2
stat t2

many times I got "t2 not found".
it see like the second command exec before the first finished.


----------------------------------------------------------------------

Comment By: Henry N. (henryn)
Date: 2008-10-22 13:41

Message:
The inode cache is default enabled for 1 Second. You can disable it by
adding "nocache" to mount options, for example:

  mount -o nocache -t cofs 0 /mnt/windows

More cosf options, see cofs.txt in your coLinux installation or here:
 
http://colinux.svn.sourceforge.net/viewvc/colinux/branches/devel/doc/cofs

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-21 13:48

Message:
I trace the code, it see like rename bug, the filename change, but inode
not.

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-10-18 09:24

Message:
Excuse me, I think this bug same as #2152550

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=622063&aid=2176188&group_id=98788

Re: [coLinux-devel] [about cofs rename bug]

[Paolo M. <pao...@gm...>]

Hi,
I was trying to understand the problem related to rename on cofs (and
check if the patch was OK for me).
To do it I write a simple script to run on cofs.
This is the script (I call it test.sh)

dd if=/dev/zero of=CCCC count=5 2>/dev/null
mv CCCC DDDD
cont=1
while [ $cont = 1 ]
do
   stat DDDD 2>&1 > /dev/null
   if [ $? != 0 ]; then
       echo -n ERROR
   else
       echo OK
       cont=0
   fi
done

If I run it, it does not show me the problem related to rename on cofs.
On my PC I don't see the problem. This is my output.

root@darkstar:/win/d# ./test.sh
OK
root@darkstar:/win/d# ./test.sh
OK
root@darkstar:/win/d# ./test.sh
OK
root@darkstar:/win/d# ./test.sh
OK
root@darkstar:/win/d# ./test.sh
OK
root@darkstar:/win/d# ./test.sh
OK
root@darkstar:/win/d# ./test.sh
OK
root@darkstar:/win/d# ./test.sh
OK

*But I see an other problem*

If I run it several times and I relanch it ***quickly*** after
termination I see these message, that sometimes show an error.
This error, if I understand correctly, is different from the #2176188.

root@darkstar:/win/d# ./test.sh
mv: `CCCC' and `DDDD' are the same file
OK
root@darkstar:/win/d# ./test.sh
OK
root@darkstar:/win/d# ./test.sh
mv: `CCCC' and `DDDD' are the same file
OK
root@darkstar:/win/d# ./test.sh
mv: `CCCC' and `DDDD' are the same file
OK
root@darkstar:/win/d# ./test.sh
OK
root@darkstar:/win/d# ./test.sh
mv: `CCCC' and `DDDD' are the same file
OK
root@darkstar:/win/d# ./test.sh
OK
root@darkstar:/win/d# ./test.sh
mv: `CCCC' and `DDDD' are the same file
OK

I use 0.7.3.

Regards
Paolo

[coLinux-devel] [about cofs rename bug]

[chen <che...@16...>]

Hi,

    I simple fix this(#2176188), it work fine in my computer, can anyone
do more check?
    Thanks!

patch:
linux-2.6.22-source/fs/cofusefs/dir.c

static int fuse_rename(struct inode *olddir, struct dentry *oldent,
		       struct inode *newdir, struct dentry *newent)
{
	struct fuse_conn *fc = INO_FC(olddir);
	struct fuse_in in = FUSE_IN_INIT;
	struct fuse_out out = FUSE_OUT_INIT;
	struct fuse_rename_in inarg;

	memset(&inarg, 0, sizeof(inarg));
	inarg.newdir = newdir->i_ino;

	in.h.opcode = FUSE_RENAME;
	in.h.ino = olddir->i_ino;
	in.numargs = 3;
	in.args[0].size = sizeof(inarg);
	in.args[0].value = &inarg;
	in.args[1].size = oldent->d_name.len + 1;
	in.args[1].value = oldent->d_name.name;
	in.args[2].size = newent->d_name.len + 1;
	in.args[2].value = newent->d_name.name;
	request_send(fc, &in, &out);

	if (!out.h.error) {
		uncache_dir(olddir);
+       d_move(oldent, newent);
		if (olddir != newdir)
			uncache_dir(newdir);
	}

	return out.h.error;
}



　　　　　　　　chen
　　　　　　　　che...@16...
　　　　　　　　　　2008-10-22