From: Steve Ellcey <sellcey@caviumnetworks.com>
To: <libc-alpha@sourceware.org>, <ynorov@caviumnetworks.com>
Subject: [PATCH] Updated version of define for __ASSUME_ST_INO_64_BIT
Date: Thu, 27 Oct 2016 20:42:00 -0000 [thread overview]
Message-ID: <1477600905.15823.9.camel@caviumnetworks.com> (raw)
[-- Attachment #1: Type: text/plain, Size: 1493 bytes --]
Yury,
I took a look at your __ASSUME_ST_INO_64_BIT patch at:
        https://sourceware.org/ml/libc-alpha/2016-09/msg00075.html
I think that, rather than ifdef'ing the setting of
__ASSUME_ST_INO_64_BIT in sysdeps/unix/sysv/linux/kernel-features.h,
the setting should be left as unconditional and then sh and alpha can
undef and re-define it to zero after including linux/kernel-features.h.
Here is my modified version of your patch, I tested it on x86 just toÂ
make sure I didn't break anything obvious, but I don't have hardware to
test sh or alpha.
Can someone approve this for checkin?
Steve Ellcey
sellcey@caviumnetworks.com
2016-10-27  Yury Norov  <ynorov@caviumnetworks.com>
            Steve Ellcey <sellcey@caviumnetworks.com>
        * sysdeps/unix/sysv/linux/alpha/kernel-features.h: #define
          __ASSUME_ST_INO_64_BIT as 0
        * sysdeps/unix/sysv/linux/sh/kernel-features.h: Likewise.
        * sysdeps/unix/sysv/linux/fxstat64.c: Replace #ifdef with #if
          over the code where __ASSUME_ST_INO_64_BIT is used.
        * sysdeps/unix/sysv/linux/lxstat64.c: Likewise.
        * sysdeps/unix/sysv/linux/xstat64.c: Likewise.
        * sysdeps/unix/sysv/linux/xstatconv.c: Likewise.
[-- Attachment #2: assume.patch --]
[-- Type: text/x-patch, Size: 3216 bytes --]
diff --git a/sysdeps/unix/sysv/linux/alpha/kernel-features.h b/sysdeps/unix/sysv/linux/alpha/kernel-features.h
index 732749e..f09222d 100644
--- a/sysdeps/unix/sysv/linux/alpha/kernel-features.h
+++ b/sysdeps/unix/sysv/linux/alpha/kernel-features.h
@@ -23,6 +23,7 @@
#include_next <kernel-features.h>
#undef __ASSUME_ST_INO_64_BIT
+#define __ASSUME_ST_INO_64_BIT 0
/* There never has been support for fstat64. */
#undef __ASSUME_STATFS64
diff --git a/sysdeps/unix/sysv/linux/fxstat64.c b/sysdeps/unix/sysv/linux/fxstat64.c
index 5468dd6..8d10282 100644
--- a/sysdeps/unix/sysv/linux/fxstat64.c
+++ b/sysdeps/unix/sysv/linux/fxstat64.c
@@ -33,7 +33,7 @@ ___fxstat64 (int vers, int fd, struct stat64 *buf)
{
int result;
result = INLINE_SYSCALL (fstat64, 2, fd, buf);
-#if defined _HAVE_STAT64___ST_INO && !defined __ASSUME_ST_INO_64_BIT
+#if defined _HAVE_STAT64___ST_INO && !__ASSUME_ST_INO_64_BIT
if (__builtin_expect (!result, 1) && buf->__st_ino != (__ino_t) buf->st_ino)
buf->st_ino = buf->__st_ino;
#endif
diff --git a/sysdeps/unix/sysv/linux/lxstat64.c b/sysdeps/unix/sysv/linux/lxstat64.c
index 09a678a..4d8c0a6 100644
--- a/sysdeps/unix/sysv/linux/lxstat64.c
+++ b/sysdeps/unix/sysv/linux/lxstat64.c
@@ -32,7 +32,7 @@ ___lxstat64 (int vers, const char *name, struct stat64 *buf)
{
int result;
result = INLINE_SYSCALL (lstat64, 2, name, buf);
-#if defined _HAVE_STAT64___ST_INO && !defined __ASSUME_ST_INO_64_BIT
+#if defined _HAVE_STAT64___ST_INO && !__ASSUME_ST_INO_64_BIT
if (__builtin_expect (!result, 1) && buf->__st_ino != (__ino_t) buf->st_ino)
buf->st_ino = buf->__st_ino;
#endif
diff --git a/sysdeps/unix/sysv/linux/sh/kernel-features.h b/sysdeps/unix/sysv/linux/sh/kernel-features.h
index c5240fa..ea4fdbc 100644
--- a/sysdeps/unix/sysv/linux/sh/kernel-features.h
+++ b/sysdeps/unix/sysv/linux/sh/kernel-features.h
@@ -44,6 +44,7 @@
/* SH does not have a 64-bit inode field. */
#undef __ASSUME_ST_INO_64_BIT
+#define __ASSUME_ST_INO_64_BIT 0
/* SH4 ABI does not really require argument alignment for 64-bits, but
the kernel interface for p{read,write}64 adds a dummy long argument
diff --git a/sysdeps/unix/sysv/linux/xstat64.c b/sysdeps/unix/sysv/linux/xstat64.c
index 2075383..00a4459 100644
--- a/sysdeps/unix/sysv/linux/xstat64.c
+++ b/sysdeps/unix/sysv/linux/xstat64.c
@@ -33,7 +33,7 @@ ___xstat64 (int vers, const char *name, struct stat64 *buf)
{
int result;
result = INLINE_SYSCALL (stat64, 2, name, buf);
-#if defined _HAVE_STAT64___ST_INO && !defined __ASSUME_ST_INO_64_BIT
+#if defined _HAVE_STAT64___ST_INO && !__ASSUME_ST_INO_64_BIT
if (__builtin_expect (!result, 1) && buf->__st_ino != (__ino_t) buf->st_ino)
buf->st_ino = buf->__st_ino;
#endif
diff --git a/sysdeps/unix/sysv/linux/xstatconv.c b/sysdeps/unix/sysv/linux/xstatconv.c
index c5fbf3d..27b528c 100644
--- a/sysdeps/unix/sysv/linux/xstatconv.c
+++ b/sysdeps/unix/sysv/linux/xstatconv.c
@@ -190,7 +190,7 @@ __xstat32_conv (int vers, struct stat64 *kbuf, struct stat *buf)
buf->__pad1 = 0;
#endif
#ifdef _HAVE_STAT64___ST_INO
-# ifndef __ASSUME_ST_INO_64_BIT
+# if !__ASSUME_ST_INO_64_BIT
if (kbuf->st_ino == 0)
buf->st_ino = kbuf->__st_ino;
else
next reply other threads:[~2016-10-27 20:42 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-27 20:42 Steve Ellcey [this message]
2016-11-07 16:42 ` Steve Ellcey
2016-11-07 20:09 ` Andreas Schwab
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1477600905.15823.9.camel@caviumnetworks.com \
--to=sellcey@caviumnetworks.com \
--cc=libc-alpha@sourceware.org \
--cc=ynorov@caviumnetworks.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).