policycoreutils: setfiles: FIXME switch from stat to

PATCH In review Daniel J Walsh 3 participants 6 messages Sep 15, 2011 – Sep 16, 2011 SELinux security module

lore

Share & tools

lore

Daniel J Walsh

author

2011-09-15 19:16 UTC #0 lore raw

PATCH 11/67 policycoreutils: setfiles: FIXME switch from stat to

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

This patch looks good to me. acked.

I think Steven Smalley also agrees with this patch. He can NACK it now or forever hold his peace...
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk5yTyIACgkQrlYvE4MpobPvTACfZdTv/nZhO8v0/3z4JmFSTq7+ qYsAoLffVmxlhRt8RMWA6MlNgH2jeSrP
=op1/
-----END PGP SIGNATURE-----

Guido Trentalancia

2011-09-16 07:13 UTC #1 lore raw

Quoted 7 lines · Daniel J Walsh

From 1268e66e94286a55c399383d5959734d3597792d Mon Sep 17 00:00:00 2001
From: Eric Paris <[email protected]>
Date: Sun, 10 Jul 2011 16:54:25 +0200
Subject: [PATCH 11/67] policycoreutils: setfiles: FIXME switch from stat to
stat64

Quoted 7 lines

This looks bad. glibc takes care of this. We should do send this upstream but I would like to know why you did it in Fedora....
NOT-Signed-off-by Eric Paris <[email protected]>
---
 policycoreutils/setfiles/restore.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-) 

Perhaps you could exploit a few #ifdef's here ?

Such as _LARGEFILE_SOURCE _LARGEFILE64_SOURCE and/or _FILE_OFFSET_BITS=64 ?

More available from:

info libc 'Feature Test Macros'

Guido

--
This message was distributed to subscribers of the selinux mailing list. If you no longer wish to subscribe, send mail to [email protected] with the words "unsubscribe selinux" without quotes as the message.

Stephen Smalley

2011-09-16 12:41 UTC #2 lore raw

Quoted 16 lines · Guido Trentalancia

Quoted 7 lines · Daniel J Walsh
From 1268e66e94286a55c399383d5959734d3597792d Mon Sep 17 00:00:00 2001
From: Eric Paris <[email protected]>
Date: Sun, 10 Jul 2011 16:54:25 +0200
Subject: [PATCH 11/67] policycoreutils: setfiles: FIXME switch from stat to
stat64
Quoted 7 lines
This looks bad. glibc takes care of this. We should do send this upstream but I would like to know why you did it in Fedora....
NOT-Signed-off-by Eric Paris <[email protected]>
---
 policycoreutils/setfiles/restore.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-) 

Quoted 8 lines

Perhaps you could exploit a few #ifdef's here ?
Such as _LARGEFILE_SOURCE _LARGEFILE64_SOURCE and/or _FILE_OFFSET_BITS=64 ?
More available from:
info libc 'Feature Test Macros'

We were doing that before converting setfiles from using nftw(3) to using fts(3) for the file tree walk. But fts.h has this gem:
/* The fts interface is incompatible with the LFS interface which
transparently uses the 64-bit file access functions. */
#ifdef __USE_FILE_OFFSET64

# error "<fts.h> cannot be used with -D_FILE_OFFSET_BITS==64"
#endif

Signature · 8 lines

Stephen Smalley
National Security Agency


--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to [email protected] with
the words "unsubscribe selinux" without quotes as the message.

Guido Trentalancia

2011-09-16 13:07 UTC #3 lore raw

Quoted 26 lines · Stephen Smalley

Quoted 16 lines · Guido Trentalancia
Quoted 7 lines · Daniel J Walsh
From 1268e66e94286a55c399383d5959734d3597792d Mon Sep 17 00:00:00 2001
From: Eric Paris <[email protected]>
Date: Sun, 10 Jul 2011 16:54:25 +0200
Subject: [PATCH 11/67] policycoreutils: setfiles: FIXME switch from stat to
stat64
Quoted 7 lines
This looks bad. glibc takes care of this. We should do send this upstream but I would like to know why you did it in Fedora....
NOT-Signed-off-by Eric Paris <[email protected]>
---
 policycoreutils/setfiles/restore.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-) 
Quoted 8 lines
Perhaps you could exploit a few #ifdef's here ?
Such as _LARGEFILE_SOURCE _LARGEFILE64_SOURCE and/or _FILE_OFFSET_BITS=64 ?
More available from:
info libc 'Feature Test Macros'

Quoted 7 lines

We were doing that before converting setfiles from using nftw(3) to using fts(3) for the file tree walk. But fts.h has this gem:
/* The fts interface is incompatible with the LFS interface which
transparently uses the 64-bit file access functions. */
#ifdef __USE_FILE_OFFSET64
# error "<fts.h> cannot be used with -D_FILE_OFFSET_BITS==64"
#endif

That must be the reason why other projects are moving in the opposite direction, see for example:

http://sourceforge.net/tracker/index.php?func=detail&aid=3126942&group_id=3382&atid=103382

Guido

Stephen Smalley

2011-09-16 15:06 UTC #4 lore raw

Quoted 35 lines · Guido Trentalancia

Quoted 26 lines · Stephen Smalley
Quoted 16 lines · Guido Trentalancia
Quoted 7 lines · Daniel J Walsh
From 1268e66e94286a55c399383d5959734d3597792d Mon Sep 17 00:00:00 2001
From: Eric Paris <[email protected]>
Date: Sun, 10 Jul 2011 16:54:25 +0200
Subject: [PATCH 11/67] policycoreutils: setfiles: FIXME switch from stat to
stat64
Quoted 7 lines
This looks bad. glibc takes care of this. We should do send this upstream but I would like to know why you did it in Fedora....
NOT-Signed-off-by Eric Paris <[email protected]>
---
 policycoreutils/setfiles/restore.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-) 
Quoted 8 lines
Perhaps you could exploit a few #ifdef's here ?
Such as _LARGEFILE_SOURCE _LARGEFILE64_SOURCE and/or _FILE_OFFSET_BITS=64 ?
More available from:
info libc 'Feature Test Macros'
Quoted 7 lines
We were doing that before converting setfiles from using nftw(3) to using fts(3) for the file tree walk. But fts.h has this gem:
/* The fts interface is incompatible with the LFS interface which
transparently uses the 64-bit file access functions. */
#ifdef __USE_FILE_OFFSET64
# error "<fts.h> cannot be used with -D_FILE_OFFSET_BITS==64"
#endif

Quoted 4 lines

That must be the reason why other projects are moving in the opposite direction, see for example:
http://sourceforge.net/tracker/index.php?func=detail&aid=3126942&group_id=3382&atid=103382

The discussion of the change for setfiles can be found in the list archives, e.g.
http://marc.info/?t=124639081000002&r=1&w=2
http://marc.info/?l=selinux&m=124688830500777&w=2

Signature · 8 lines

Stephen Smalley
National Security Agency


--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to [email protected] with
the words "unsubscribe selinux" without quotes as the message.

Guido Trentalancia

2011-09-16 15:40 UTC #5 lore raw

Quoted 39 lines · Stephen Smalley

Quoted 38 lines · Guido Trentalancia
Quoted 26 lines · Stephen Smalley
Quoted 16 lines · Guido Trentalancia
Quoted 7 lines · Daniel J Walsh
From 1268e66e94286a55c399383d5959734d3597792d Mon Sep 17 00:00:00 2001
From: Eric Paris <[email protected]>
Date: Sun, 10 Jul 2011 16:54:25 +0200
Subject: [PATCH 11/67] policycoreutils: setfiles: FIXME switch from stat to
stat64
Quoted 7 lines
This looks bad. glibc takes care of this. We should do send this upstream but I would like to know why you did it in Fedora....
NOT-Signed-off-by Eric Paris <[email protected]>
---
 policycoreutils/setfiles/restore.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-) 
Quoted 8 lines
Perhaps you could exploit a few #ifdef's here ?
Such as _LARGEFILE_SOURCE _LARGEFILE64_SOURCE and/or _FILE_OFFSET_BITS=64 ?
More available from:
info libc 'Feature Test Macros'
Quoted 7 lines
We were doing that before converting setfiles from using nftw(3) to using fts(3) for the file tree walk. But fts.h has this gem:
/* The fts interface is incompatible with the LFS interface which
transparently uses the 64-bit file access functions. */
#ifdef __USE_FILE_OFFSET64
# error "<fts.h> cannot be used with -D_FILE_OFFSET_BITS==64"
#endif
That must be the reason why other projects are moving in the opposite direction, see for example:

Quoted 6 lines

Quoted 1 line
http://sourceforge.net/tracker/index.php?func=detail&aid=3126942&group_id=3382&atid=103382
The discussion of the change for setfiles can be found in the list archives, e.g.
http://marc.info/?t=124639081000002&r=1&w=2
http://marc.info/?l=selinux&m=124688830500777&w=2

Hi Stephen,

thanks for getting back.

Yes, I read that already. I still cannot understand the point since it says about the benchmark results: "so not much of a change".

Last thing I can add is that I'm pretty sure I've seen around code designed to fall back to the other scheme when _FILE_OFFSET_BITS=64 is defined.

Regards,

Guido