|
»
»
Odd write usage
|
View:
New views
7 Messages
—
Rating Filter:
Alert me
|
 |
Odd write usage
by Daniel Benamy-3
::
Rate this Message:
Hi,
In my filesystem, I don't pass the length of buf in size. Instead I specify one of a couple of constants as the size which tells the fs what to do. Buf is a shorter, null terminated string (shorter than the value passed as size). This works correctly; valgrind doesn't show memory errors. Is this behavior supported? Ie, should it continue to work or might it suddenly break in a future version? Thanks again, Dan ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ fuse-devel mailing list fuse-devel@... https://lists.sourceforge.net/lists/listinfo/fuse-devel |
|
|
Re: Odd write usage
by Miklos Szeredi
::
Rate this Message:
On Fri, 18 Jul 2008, Daniel Benamy wrote:
> In my filesystem, I don't pass the length of buf in size. Instead I > specify one of a couple of constants as the size which tells the fs > what to do. Buf is a shorter, null terminated string (shorter than > the value passed as size). This works correctly; valgrind doesn't > show memory errors. Is this behavior supported? Ie, should it > continue to work or might it suddenly break in a future version? Sounds pretty weird. Why not pass the "command" as the first few bytes of the buffer, instead of encoding it in the length? Miklos ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ fuse-devel mailing list fuse-devel@... https://lists.sourceforge.net/lists/listinfo/fuse-devel |
|
|
Re: Odd write usage
by Daniel Benamy-3
::
Rate this Message:
On Saturday 19 July 2008 03:54:14 Miklos Szeredi wrote:
> On Fri, 18 Jul 2008, Daniel Benamy wrote: > > In my filesystem, I don't pass the length of buf in size. Instead I > > specify one of a couple of constants as the size which tells the fs > > what to do. Buf is a shorter, null terminated string (shorter than > > the value passed as size). This works correctly; valgrind doesn't > > show memory errors. Is this behavior supported? Ie, should it > > continue to work or might it suddenly break in a future version? > > Sounds pretty weird. > > Why not pass the "command" as the first few bytes of the buffer, > instead of encoding it in the length? There's an existing kernel filesystem which has this interface (for some reason unknown to me) and I'm trying to support the same interface. Dan ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ fuse-devel mailing list fuse-devel@... https://lists.sourceforge.net/lists/listinfo/fuse-devel |
|
|
Re: Odd write usage
by Miklos Szeredi
::
Rate this Message:
On Tue, 22 Jul 2008, Daniel Benamy wrote:
> On Saturday 19 July 2008 03:54:14 Miklos Szeredi wrote: > > On Fri, 18 Jul 2008, Daniel Benamy wrote: > > > In my filesystem, I don't pass the length of buf in size. Instead I > > > specify one of a couple of constants as the size which tells the fs > > > what to do. Buf is a shorter, null terminated string (shorter than > > > the value passed as size). This works correctly; valgrind doesn't > > > show memory errors. Is this behavior supported? Ie, should it > > > continue to work or might it suddenly break in a future version? > > > > Sounds pretty weird. > > > > Why not pass the "command" as the first few bytes of the buffer, > > instead of encoding it in the length? > > There's an existing kernel filesystem which has this interface Curious, which one? Miklos ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ fuse-devel mailing list fuse-devel@... https://lists.sourceforge.net/lists/listinfo/fuse-devel |
|
|
Re: Odd write usage
by Daniel Benamy-3
::
Rate this Message:
On Wednesday 23 July 2008 02:49:24 Miklos Szeredi wrote:
> On Tue, 22 Jul 2008, Daniel Benamy wrote: > > On Saturday 19 July 2008 03:54:14 Miklos Szeredi wrote: > > > On Fri, 18 Jul 2008, Daniel Benamy wrote: > > > > In my filesystem, I don't pass the length of buf in size. Instead I > > > > specify one of a couple of constants as the size which tells the fs > > > > what to do. Buf is a shorter, null terminated string (shorter than > > > > the value passed as size). This works correctly; valgrind doesn't > > > > show memory errors. Is this behavior supported? Ie, should it > > > > continue to work or might it suddenly break in a future version? > > > > > > Sounds pretty weird. > > > > > > Why not pass the "command" as the first few bytes of the buffer, > > > instead of encoding it in the length? > > > > There's an existing kernel filesystem which has this interface > > Curious, which one? Sorry I wasn't clear. We have an existing linux kernel vmblock driver/fs which uses this interface. If you're really interested, check out http://sourceforge.net/project/showfiles.php?group_id=204462 Dan ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ fuse-devel mailing list fuse-devel@... https://lists.sourceforge.net/lists/listinfo/fuse-devel |
|
|
Re: Odd write usage
by Miklos Szeredi
::
Rate this Message:
On Wed, 23 Jul 2008, Daniel Benamy wrote:
> Sorry I wasn't clear. We have an existing linux kernel vmblock > driver/fs which uses this interface. If you're really interested, > check out > http://sourceforge.net/project/showfiles.php?group_id=204462 Ah, I see. Well, it might work most of the time with fuse, and the comand constants being around 100 may make it even likely to work. But in reality it will only work assuming that: a) the string will fit in that 98 or so bytes b) the unused bytes in that range happen to fall within mapped virtual memory And b) isn't guaranteed at all, unless userspace is using specially prepared buffers. So, my advice is to migrate to a new, sane interface as quickly as possible. Miklos ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ fuse-devel mailing list fuse-devel@... https://lists.sourceforge.net/lists/listinfo/fuse-devel |
|
|
Re: Odd write usage
by Daniel Benamy-3
::
Rate this Message:
On Wednesday 23 July 2008 10:43:19 Miklos Szeredi wrote:
> On Wed, 23 Jul 2008, Daniel Benamy wrote: > > Sorry I wasn't clear. We have an existing linux kernel vmblock > > driver/fs which uses this interface. If you're really interested, > > check out > > http://sourceforge.net/project/showfiles.php?group_id=204462 > > Ah, I see. Well, it might work most of the time with fuse, and the > comand constants being around 100 may make it even likely to work. > But in reality it will only work assuming that: > > a) the string will fit in that 98 or so bytes > > b) the unused bytes in that range happen to fall within mapped > virtual memory > > And b) isn't guaranteed at all, unless userspace is using specially > prepared buffers. So, my advice is to migrate to a new, sane > interface as quickly as possible. Ok. Thanks a lot for taking a look! Dan ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ fuse-devel mailing list fuse-devel@... https://lists.sourceforge.net/lists/listinfo/fuse-devel |
| Free Forum Powered by Nabble | Forum Help |