| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <20110620143131.9c6e040c.akpm@linux-foundation.org>
Date: Mon, 20 Jun 2011 14:31:31 -0700
From: Andrew Morton <akpm@...ux-foundation.org>
To: drepper@...adia.org
Cc: kosaki.motohiro@...fujitsu.com, linux-kernel@...r.kernel.org,
rientjes@...gle.com, torvalds@...ux-foundation.org,
viro@...iv.linux.org.uk, wilsons@...rt.ca,
linux-man@...r.kernel.org,
Yoshinori Sato <ysato@...rs.sourceforge.jp>,
Geert Uytterhoeven <geert@...ux-m68k.org>,
Chris Zankel <chris@...kel.net>
Subject: Re: [PATCH] Add cloexec information to fdinfo
On Thu, 9 Jun 2011 23:55:08 -0400
drepper@...adia.org wrote:
> There is one piece of information about a file descriptor which is
> currently not visible from the outside: the close-on-exec flag. The
> /proc/PID/fdinfo/* files have the mode information but this is
> missing. Is the following patch acceptable?
>
> What I don't know is whether the RCU locking is needed given that
> real locks are taken. Someone with more knowledge could just
> remove those two lines.
The locking looks OK to me.
Hopefully /prod/pid/fdinfo is documented somewhere.
linux-man@...r.kernel.org appears to be the email address we use to rub
the documentation lamp.
> diff --git a/fs/proc/base.c b/fs/proc/base.c
> index 14def99..bda3651 100644
> --- a/fs/proc/base.c
> +++ b/fs/proc/base.c
> @@ -1924,12 +1924,23 @@ static int proc_fd_info(struct inode *inode, struct path *path, char *info)
> *path = file->f_path;
> path_get(&file->f_path);
> }
> - if (info)
> + if (info) {
> + int cloexec;
> + struct fdtable *fdt;
> +
> + rcu_read_lock();
> + fdt = files_fdtable(files);
> + cloexec = FD_ISSET(fd, fdt->close_on_exec);
Does FD_ISSET return 0 or 1? Or 0 or non-zero?
For x86 it's the former.
<checks the architectures>
arch/h8300/include/asm/posix_types.h: busted
#define __FD_ISSET(d, set) ((set)->fds_bits[__FDELT(d)] & __FDMASK(d))
arch/m68k/include/asm/posix_types.h: busted
#define __FD_ISSET(d, set) ((set)->fds_bits[__FDELT(d)] & __FDMASK(d))
arch/xtensa/include/asm/posix_types.h: busted
#define __FD_ISSET(d, set) ((set)->fds_bits[__FDELT(d)] & __FDMASK(d))
From: Andrew Morton <akpm@...ux-foundation.org>
Harmonise these return values with other architectures. In some cases
this affects all compilers and in other cases non-gcc compilers only.
Cc: Yoshinori Sato <ysato@...rs.sourceforge.jp>
Cc: Geert Uytterhoeven <geert@...ux-m68k.org>
Cc: Chris Zankel <chris@...kel.net>
Cc: Ulrich Drepper <drepper@...il.com>
Signed-off-by: Andrew Morton <akpm@...ux-foundation.org>
---
arch/h8300/include/asm/posix_types.h | 2 +-
arch/m68k/include/asm/posix_types.h | 2 +-
arch/xtensa/include/asm/posix_types.h | 2 +-
3 files changed, 3 insertions(+), 3 deletions(-)
diff -puN arch/h8300/include/asm/posix_types.h~h8300-m68k-xtensa-__fd_isset-should-return-0-1 arch/h8300/include/asm/posix_types.h
--- a/arch/h8300/include/asm/posix_types.h~h8300-m68k-xtensa-__fd_isset-should-return-0-1
+++ a/arch/h8300/include/asm/posix_types.h
@@ -50,7 +50,7 @@ typedef struct {
#define __FD_CLR(d, set) ((set)->fds_bits[__FDELT(d)] &= ~__FDMASK(d))
#undef __FD_ISSET
-#define __FD_ISSET(d, set) ((set)->fds_bits[__FDELT(d)] & __FDMASK(d))
+#define __FD_ISSET(d, set) (!!((set)->fds_bits[__FDELT(d)] & __FDMASK(d)))
#undef __FD_ZERO
#define __FD_ZERO(fdsetp) (memset (fdsetp, 0, sizeof(*(fd_set *)fdsetp)))
diff -puN arch/m68k/include/asm/posix_types.h~h8300-m68k-xtensa-__fd_isset-should-return-0-1 arch/m68k/include/asm/posix_types.h
--- a/arch/m68k/include/asm/posix_types.h~h8300-m68k-xtensa-__fd_isset-should-return-0-1
+++ a/arch/m68k/include/asm/posix_types.h
@@ -51,7 +51,7 @@ typedef struct {
#define __FD_CLR(d, set) ((set)->fds_bits[__FDELT(d)] &= ~__FDMASK(d))
#undef __FD_ISSET
-#define __FD_ISSET(d, set) ((set)->fds_bits[__FDELT(d)] & __FDMASK(d))
+#define __FD_ISSET(d, set) (!!((set)->fds_bits[__FDELT(d)] & __FDMASK(d)))
#undef __FD_ZERO
#define __FD_ZERO(fdsetp) (memset (fdsetp, 0, sizeof(*(fd_set *)fdsetp)))
diff -puN arch/xtensa/include/asm/posix_types.h~h8300-m68k-xtensa-__fd_isset-should-return-0-1 arch/xtensa/include/asm/posix_types.h
--- a/arch/xtensa/include/asm/posix_types.h~h8300-m68k-xtensa-__fd_isset-should-return-0-1
+++ a/arch/xtensa/include/asm/posix_types.h
@@ -58,7 +58,7 @@ typedef struct {
#define __FD_SET(d, set) ((set)->fds_bits[__FDELT(d)] |= __FDMASK(d))
#define __FD_CLR(d, set) ((set)->fds_bits[__FDELT(d)] &= ~__FDMASK(d))
-#define __FD_ISSET(d, set) ((set)->fds_bits[__FDELT(d)] & __FDMASK(d))
+#define __FD_ISSET(d, set) (!!((set)->fds_bits[__FDELT(d)] & __FDMASK(d)))
#define __FD_ZERO(set) \
((void) memset ((void *) (set), 0, sizeof (__kernel_fd_set)))
_
> + rcu_read_unlock();
> +
> snprintf(info, PROC_FDINFO_MAX,
> "pos:\t%lli\n"
> - "flags:\t0%o\n",
> + "flags:\t0%o\n"
> + "cloexec: %d\n",
Should use a tab here for consistency.
--- a/fs/proc/base.c~proc-pid-fdinfo-add-cloexec-information-fix
+++ a/fs/proc/base.c
@@ -1936,7 +1936,7 @@ static int proc_fd_info(struct inode *in
snprintf(info, PROC_FDINFO_MAX,
"pos:\t%lli\n"
"flags:\t0%o\n"
- "cloexec: %d\n",
+ "cloexec:\t%d\n",
(long long) file->f_pos,
file->f_flags,
cloexec);
_
> (long long) file->f_pos,
> - file->f_flags);
> + file->f_flags,
> + cloexec);
> + }
> spin_unlock(&files->file_lock);
> put_files_struct(files);
> return 0;
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists