Badari Pulavarty <[EMAIL PROTECTED]> wrote:
>
> On Wed, 2005-03-09 at 11:53, Andrew Morton wrote:
> > Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
> > >
> > > >Solaris, which does forcedirectio as a mount option, actually
> > > > will do buffered I/O on the trailing part. Consider it
On Wed, 2005-03-09 at 14:39, Andrew Morton wrote:
> Badari Pulavarty <[EMAIL PROTECTED]> wrote:
> >
> > On Wed, 2005-03-09 at 11:53, Andrew Morton wrote:
> > > Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
> > > >
> > > > > Solaris, which does forcedirectio as a mount option, actually
> >
On Wed, 2005-03-09 at 11:53, Andrew Morton wrote:
> Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
> >
> > > Solaris, which does forcedirectio as a mount option, actually
> > > will do buffered I/O on the trailing part. Consider it like a bounce
> > > buffer. That way they don't DMA the
On Wed, Mar 09, 2005 at 11:53:48AM -0800, Andrew Morton wrote:
> Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
> > If writes/truncates take care of zeroing out the rest of the sector
> > on disk, might we still be OK without having to do the bounce buffer
> > thing ?
>
> We can probably rely
Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
>
> >Solaris, which does forcedirectio as a mount option, actually
> > will do buffered I/O on the trailing part. Consider it like a bounce
> > buffer. That way they don't DMA the trailing data and succeed the I/O.
> > The I/O returns
On Tue, 2005-03-08 at 17:44, Daniel McNeil wrote:
> On Tue, 2005-03-08 at 11:18, Badari Pulavarty wrote:
> > > Andrew, please don't apply the original patch. We shouldn't even attempt
> > > to submit IO beyond the filesize. We should truncate the IO request to
> > > filesize. I will send a patch
On Tue, Mar 08, 2005 at 08:07:57PM -0800, Joel Becker wrote:
> On Tue, Mar 08, 2005 at 03:54:04PM -0800, Badari Pulavarty wrote:
> > > 1. return EINVAL if the DIO goes past EOF.
> > >
> > > 2. truncate the request to file size (which is what your patch does)
> > > and if it works, it works.
>
Le mardi 08 mars 2005 à 20:07 -0800, Joel Becker a écrit :
> On Tue, Mar 08, 2005 at 03:54:04PM -0800, Badari Pulavarty wrote:
> > > 1. return EINVAL if the DIO goes past EOF.
> > >
> > > 2. truncate the request to file size (which is what your patch does)
> > > and if it works, it works.
> >
Le mardi 08 mars 2005 à 20:07 -0800, Joel Becker a écrit :
On Tue, Mar 08, 2005 at 03:54:04PM -0800, Badari Pulavarty wrote:
1. return EINVAL if the DIO goes past EOF.
2. truncate the request to file size (which is what your patch does)
and if it works, it works.
3.
On Tue, Mar 08, 2005 at 08:07:57PM -0800, Joel Becker wrote:
On Tue, Mar 08, 2005 at 03:54:04PM -0800, Badari Pulavarty wrote:
1. return EINVAL if the DIO goes past EOF.
2. truncate the request to file size (which is what your patch does)
and if it works, it works.
3.
On Tue, 2005-03-08 at 17:44, Daniel McNeil wrote:
On Tue, 2005-03-08 at 11:18, Badari Pulavarty wrote:
Andrew, please don't apply the original patch. We shouldn't even attempt
to submit IO beyond the filesize. We should truncate the IO request to
filesize. I will send a patch today to
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
Solaris, which does forcedirectio as a mount option, actually
will do buffered I/O on the trailing part. Consider it like a bounce
buffer. That way they don't DMA the trailing data and succeed the I/O.
The I/O returns actual bytes
On Wed, Mar 09, 2005 at 11:53:48AM -0800, Andrew Morton wrote:
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
If writes/truncates take care of zeroing out the rest of the sector
on disk, might we still be OK without having to do the bounce buffer
thing ?
We can probably rely on the
On Wed, 2005-03-09 at 11:53, Andrew Morton wrote:
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
Solaris, which does forcedirectio as a mount option, actually
will do buffered I/O on the trailing part. Consider it like a bounce
buffer. That way they don't DMA the trailing data and
On Wed, 2005-03-09 at 14:39, Andrew Morton wrote:
Badari Pulavarty [EMAIL PROTECTED] wrote:
On Wed, 2005-03-09 at 11:53, Andrew Morton wrote:
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
Solaris, which does forcedirectio as a mount option, actually
will do buffered
Badari Pulavarty [EMAIL PROTECTED] wrote:
On Wed, 2005-03-09 at 11:53, Andrew Morton wrote:
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
Solaris, which does forcedirectio as a mount option, actually
will do buffered I/O on the trailing part. Consider it like a bounce
On Tue, Mar 08, 2005 at 03:54:04PM -0800, Badari Pulavarty wrote:
> > 1. return EINVAL if the DIO goes past EOF.
> >
> > 2. truncate the request to file size (which is what your patch does)
> > and if it works, it works.
> >
> > 3. truncate the request to a size that actually works - like a
On Tue, 2005-03-08 at 11:18, Badari Pulavarty wrote:
> > Andrew, please don't apply the original patch. We shouldn't even attempt
> > to submit IO beyond the filesize. We should truncate the IO request to
> > filesize. I will send a patch today to fix this.
> >
>
> Well, spoke too soon. This is
On Tue, 2005-03-08 at 11:18, Badari Pulavarty wrote:
> > Andrew, please don't apply the original patch. We shouldn't even attempt
> > to submit IO beyond the filesize. We should truncate the IO request to
> > filesize. I will send a patch today to fix this.
> >
>
> Well, spoke too soon. This is
On Tue, 2005-03-08 at 15:27, Daniel McNeil wrote:
> On Tue, 2005-03-08 at 11:18, Badari Pulavarty wrote:
> > > Andrew, please don't apply the original patch. We shouldn't even attempt
> > > to submit IO beyond the filesize. We should truncate the IO request to
> > > filesize. I will send a patch
> Andrew, please don't apply the original patch. We shouldn't even attempt
> to submit IO beyond the filesize. We should truncate the IO request to
> filesize. I will send a patch today to fix this.
>
Well, spoke too soon. This is an ugly corner case :( But I have
a ugly hack to fix it :)
Let
On Tue, 2005-03-08 at 01:09, Suparna Bhattacharya wrote:
>
> Hmm, shouldn't dio->result ideally have been adjusted to be within
> i_size at the time of io submission, so we don't have to deal with
> this during completion ? We are creating bios with the right size
> after all.
>
> We have
Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
>
> But, I'm wondering if this solution just covers up the real reason
> for the single thread problem ? We shouldn't even be issuing IOs
> beyond i_size into the user-space buffer !
That's true.
-
To unsubscribe from this list: send the line
On Tue, Mar 08, 2005 at 01:41:41AM -0800, Andrew Morton wrote:
> Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
> >
> > > > can you spot what is going wrong here that we have to try and
> > > > workaround this later ?
> > >
> > > Good question. Do we have the i_sem coverage to prevent a
Le mardi 08 mars 2005 à 01:18 -0800, Andrew Morton a écrit :
> Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
> >
> > ...
> >
> > Bugs in this area seem never-ending don't they - plug one, open up
> > another - hard to be confident/verify :( - someday we'll have to
> > rewrite a part of this
Le mardi 08 mars 2005 à 01:18 -0800, Andrew Morton a écrit :
> Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
> >
> > ...
> >
> > Bugs in this area seem never-ending don't they - plug one, open up
> > another - hard to be confident/verify :( - someday we'll have to
> > rewrite a part of this
Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
>
> > > can you spot what is going wrong here that we have to try and
> > > workaround this later ?
> >
> > Good question. Do we have the i_sem coverage to prevent a concurrent
> > truncate?
> >
> > But from Sebastien's description it
On Tue, Mar 08, 2005 at 01:18:14AM -0800, Andrew Morton wrote:
> Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
> >
> > ...
> >
> > Bugs in this area seem never-ending don't they - plug one, open up
> > another - hard to be confident/verify :( - someday we'll have to
> > rewrite a part of this
Suparna Bhattacharya <[EMAIL PROTECTED]> wrote:
>
> ...
>
> Bugs in this area seem never-ending don't they - plug one, open up
> another - hard to be confident/verify :( - someday we'll have to
> rewrite a part of this code.
It's solving a complex problem. Any rewrite would probably end up just
On Mon, Mar 07, 2005 at 10:39:17PM -0800, Andrew Morton wrote:
> 6
> Lines: 76
>
> Sébastien Dugué <[EMAIL PROTECTED]> wrote:
> >
> > When reading a file in async mode (using kernel AIO), and the file
> > size is lower than the requested size (short read), the direct IO
> > layer reports an
On Mon, Mar 07, 2005 at 10:39:17PM -0800, Andrew Morton wrote:
6
Lines: 76
Sébastien Dugué [EMAIL PROTECTED] wrote:
When reading a file in async mode (using kernel AIO), and the file
size is lower than the requested size (short read), the direct IO
layer reports an incorrect number
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
...
Bugs in this area seem never-ending don't they - plug one, open up
another - hard to be confident/verify :( - someday we'll have to
rewrite a part of this code.
It's solving a complex problem. Any rewrite would probably end up just as
On Tue, Mar 08, 2005 at 01:18:14AM -0800, Andrew Morton wrote:
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
...
Bugs in this area seem never-ending don't they - plug one, open up
another - hard to be confident/verify :( - someday we'll have to
rewrite a part of this code.
It's
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
can you spot what is going wrong here that we have to try and
workaround this later ?
Good question. Do we have the i_sem coverage to prevent a concurrent
truncate?
But from Sebastien's description it doesn't soound as if a
Le mardi 08 mars 2005 à 01:18 -0800, Andrew Morton a écrit :
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
...
Bugs in this area seem never-ending don't they - plug one, open up
another - hard to be confident/verify :( - someday we'll have to
rewrite a part of this code.
It's
Le mardi 08 mars 2005 à 01:18 -0800, Andrew Morton a écrit :
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
...
Bugs in this area seem never-ending don't they - plug one, open up
another - hard to be confident/verify :( - someday we'll have to
rewrite a part of this code.
It's
On Tue, Mar 08, 2005 at 01:41:41AM -0800, Andrew Morton wrote:
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
can you spot what is going wrong here that we have to try and
workaround this later ?
Good question. Do we have the i_sem coverage to prevent a concurrent
Suparna Bhattacharya [EMAIL PROTECTED] wrote:
But, I'm wondering if this solution just covers up the real reason
for the single thread problem ? We shouldn't even be issuing IOs
beyond i_size into the user-space buffer !
That's true.
-
To unsubscribe from this list: send the line
On Tue, 2005-03-08 at 01:09, Suparna Bhattacharya wrote:
Hmm, shouldn't dio-result ideally have been adjusted to be within
i_size at the time of io submission, so we don't have to deal with
this during completion ? We are creating bios with the right size
after all.
We have this:
Andrew, please don't apply the original patch. We shouldn't even attempt
to submit IO beyond the filesize. We should truncate the IO request to
filesize. I will send a patch today to fix this.
Well, spoke too soon. This is an ugly corner case :( But I have
a ugly hack to fix it :)
Let me
On Tue, 2005-03-08 at 15:27, Daniel McNeil wrote:
On Tue, 2005-03-08 at 11:18, Badari Pulavarty wrote:
Andrew, please don't apply the original patch. We shouldn't even attempt
to submit IO beyond the filesize. We should truncate the IO request to
filesize. I will send a patch today to
On Tue, 2005-03-08 at 11:18, Badari Pulavarty wrote:
Andrew, please don't apply the original patch. We shouldn't even attempt
to submit IO beyond the filesize. We should truncate the IO request to
filesize. I will send a patch today to fix this.
Well, spoke too soon. This is an ugly
On Tue, 2005-03-08 at 11:18, Badari Pulavarty wrote:
Andrew, please don't apply the original patch. We shouldn't even attempt
to submit IO beyond the filesize. We should truncate the IO request to
filesize. I will send a patch today to fix this.
Well, spoke too soon. This is an ugly
On Tue, Mar 08, 2005 at 03:54:04PM -0800, Badari Pulavarty wrote:
1. return EINVAL if the DIO goes past EOF.
2. truncate the request to file size (which is what your patch does)
and if it works, it works.
3. truncate the request to a size that actually works - like a multiple
Sébastien Dugué <[EMAIL PROTECTED]> wrote:
>
> When reading a file in async mode (using kernel AIO), and the file
> size is lower than the requested size (short read), the direct IO
> layer reports an incorrect number of bytes read (transferred).
>
> That case is handled for the sync path in
Hi,
When reading a file in async mode (using kernel AIO), and the file
size is lower than the requested size (short read), the direct IO
layer reports an incorrect number of bytes read (transferred).
That case is handled for the sync path in 'direct_io_worker'
(fs/direct-io.c) where a check
Hi,
When reading a file in async mode (using kernel AIO), and the file
size is lower than the requested size (short read), the direct IO
layer reports an incorrect number of bytes read (transferred).
That case is handled for the sync path in 'direct_io_worker'
(fs/direct-io.c) where a check
Sébastien Dugué [EMAIL PROTECTED] wrote:
When reading a file in async mode (using kernel AIO), and the file
size is lower than the requested size (short read), the direct IO
layer reports an incorrect number of bytes read (transferred).
That case is handled for the sync path in
48 matches
Mail list logo