On Sat, 2008-01-26 at 22:03 -0800, Andrew Morton wrote:
> > On Wed, 23 Jan 2008 10:29:37 -0800 Matt Helsley <[EMAIL PROTECTED]> wrote:
> >
> > Andrew, please consider this patch for inclusion in -mm.
> >
> > ...
> >
>
> Can't say that we're particularly exercised about mvfs's problems, but the
On Sat, 2008-01-26 at 22:03 -0800, Andrew Morton wrote:
On Wed, 23 Jan 2008 10:29:37 -0800 Matt Helsley [EMAIL PROTECTED] wrote:
Andrew, please consider this patch for inclusion in -mm.
...
Can't say that we're particularly exercised about mvfs's problems, but the
current way of
> On Wed, 23 Jan 2008 10:29:37 -0800 Matt Helsley <[EMAIL PROTECTED]> wrote:
>
> Andrew, please consider this patch for inclusion in -mm.
>
> ...
>
Can't say that we're particularly exercised about mvfs's problems, but the
current way of doing /proc/pid/exe is indeed a nasty hack.
>
>
In message <[EMAIL PROTECTED]>, Jan Engelhardt writes:
>
> On Jan 23 2008 10:29, Matt Helsley wrote:
> >
> >For executables on the stackable MVFS filesystem the current procfs
> >methods for implementing a task's exe symlink do not point to the
> >correct file and applications relying on the
Matt Helsley <[EMAIL PROTECTED]> wrote:
> To solve the problem this patch changes the way that the kernel resolves a
> task's exe symlink. Instead of walking the VMAs to find the first
> executable file-backed VMA we store a reference to the exec'd file in the
> mm_struct -- /foo/bar/jvm/bin/java
On Wed, Jan 23, 2008 at 10:29:37AM -0800, Matt Helsley wrote:
> For executables on the stackable MVFS filesystem the current procfs methods
> for
> implementing a task's exe symlink do not point to the correct file and
> applications relying on the symlink fail (see the java example below).
Dou
On Jan 23 2008 10:29, Matt Helsley wrote:
>
>For executables on the stackable MVFS filesystem the current procfs
>methods for implementing a task's exe symlink do not point to the
>correct file and applications relying on the symlink fail (see the
>java example below).
This reminds me of
The kernel implements readlink of /proc/pid/exe by getting the file from the
first executable VMA. Then the path to the file is reconstructed and reported as
the result. While this method is often correct it does not always identify the
correct path.
Some applications may play games with the
The kernel implements readlink of /proc/pid/exe by getting the file from the
first executable VMA. Then the path to the file is reconstructed and reported as
the result. While this method is often correct it does not always identify the
correct path.
Some applications may play games with the
On Jan 23 2008 10:29, Matt Helsley wrote:
For executables on the stackable MVFS filesystem the current procfs
methods for implementing a task's exe symlink do not point to the
correct file and applications relying on the symlink fail (see the
java example below).
This reminds me of unoionfs -
On Wed, Jan 23, 2008 at 10:29:37AM -0800, Matt Helsley wrote:
For executables on the stackable MVFS filesystem the current procfs methods
for
implementing a task's exe symlink do not point to the correct file and
applications relying on the symlink fail (see the java example below).
Dou you
Matt Helsley [EMAIL PROTECTED] wrote:
To solve the problem this patch changes the way that the kernel resolves a
task's exe symlink. Instead of walking the VMAs to find the first
executable file-backed VMA we store a reference to the exec'd file in the
mm_struct -- /foo/bar/jvm/bin/java in
12 matches
Mail list logo