Reduce kernel-features.h duplication
Commit Message
This patch reduces duplication between different architectures'
kernel-features.h files by making the architecture-independent file
define various macros unconditionally (instead of only for a
particular list of architectures), with the architecture-specific
files then undefining the macros if necessary.
Specifically, __ASSUME_O_CLOEXEC (O_CLOEXEC flag to open) and
__ASSUME_SOCK_CLOEXEC (SOCK_NONBLOCK and SOCK_CLOEXEC flags to socket)
are supported on all architectures as of 2.6.32 or the minimum kernel
version for the architecture if later. For __ASSUME_IN_NONBLOCK,
__ASSUME_PIPE2, __ASSUME_EVENTFD2, __ASSUME_SIGNALFD4 and
__ASSUME_DUP3, the relevant syscalls were added for alpha in 2.6.33
but otherwise the features are available as of 2.6.32. For
__ASSUME_UTIMES, support is everywhere in 2.6.32 except for
asm-generic architectures and hppa.
Although those were the main cases of duplication among
kernel-features.h files, some other cases of unnecessary definitions
were also cleaned up: the hppa file defined various macros that were
either no longer used at all, or defined by the main file by default
anyway, the ia64 file had duplicative definitions of __ASSUME_PSELECT
and __ASSUME_PPOLL, while mips had such a definition of
__ASSUME_IPC64.
Really, rather than being defined in the main file then undefined for
asm-generic architectures, __ASSUME_UTIMES should become an
hppa-specific macro. Given that __ASSUME_ATFCTS and
__ASSUME_UTIMENSAT are now always true, the only live __ASSUME_UTIMES
conditional is in sysdeps/unix/sysv/linux/utimes.c, which is not used
for asm-generic architectures. I think the desired state would be an
hppa-specific file (that includes sysdeps/unix/sysv/linux/utimes.c if
__ASSUME_UTIMES, and otherwise has fallback code), with the fallback
code being removed from the main utimes.c. But I think that's most
reasonably a separate cleanup once __ASSUME_ATFCTS and
__ASSUME_UTIMESAT have both had conditional code cleaned up.
Given this patch, I think it's straightforward to move non-ex-ports
architectures to having their own kernel-features.h files, like
ex-ports architectures, rather than conditionals in the main file
(i.e., such a move won't require the architecture-specific file to
contain anything that isn't genuinely architecture-specific), and
would encourage architecture maintainers to do so.
Tested x86_64 that the installed shared libraries are unchanged by
this patch. Note that on some architectures this *will* cause
__ASSUME_* macros to be defined in cases where they weren't previously
but should have been (but this is just optimization, not a fix to a
user-visible bug, so doesn't need a bug report in Bugzilla).
2014-05-12 Joseph Myers <joseph@codesourcery.com>
* sysdeps/unix/sysv/linux/kernel-features.h (__ASSUME_UTIMES):
Define unconditionally.
(__ASSUME_O_CLOEXEC): Likewise.
(__ASSUME_SOCK_CLOEXEC): Likewise.
(__ASSUME_IN_NONBLOCK): Likewise.
(__ASSUME_PIPE2): Likewise.
(__ASSUME_EVENTFD2): Likewise.
(__ASSUME_SIGNALFD4): Likewise.
(__ASSUME_DUP3): Likewise.
* sysdeps/unix/sysv/linux/aarch64/kernel-features.h
(__ASSUME_DUP3): Do not define.
(__ASSUME_EVENTFD2): Likewise.
(__ASSUME_IN_NONBLOCK): Likewise.
(__ASSUME_O_CLOEXEC): Likewise.
(__ASSUME_PIPE2): Likewise.
(__ASSUME_SIGNALFD4): Likewise.
(__ASSUME_SOCK_CLOEXEC): Likewise.
(__ASSUME_UTIMES): Undefine.
* sysdeps/unix/sysv/linux/alpha/kernel-features.h
(__ASSUME_UTIMES): Do not define.
(__ASSUME_O_CLOEXEC): Likewise.
(__ASSUME_SOCK_CLOEXEC): Likewise.
(__ASSUME_IN_NONBLOCK): Undefine if [__LINUX_KERNEL_VERSION <
0x020621] instead of defining if [__LINUX_KERNEL_VERSION >=
0x020621].
(__ASSUME_PIPE2): Likewise.
(__ASSUME_EVENTFD2): Likewise.
(__ASSUME_SIGNALFD4): Likewise.
[__LINUX_KERNEL_VERSION < 0x020621] (__ASSUME_DUP3): Undefine.
* sysdeps/unix/sysv/linux/arm/kernel-features.h (__ASSUME_UTIMES):
Do not define.
(__ASSUME_EVENTFD2): Likewise.
(__ASSUME_SIGNALFD4): Likewise.
* sysdeps/unix/sysv/linux/hppa/kernel-features.h
(__ASSUME_32BITUIDS): Likewise.
(__ASSUME_TRUNCATE64_SYSCALL): Likewise.
(__ASSUME_IPC64): Likewise.
(__ASSUME_ST_INO_64_BIT): Likewise.
(__ASSUME_GETDENTS64_SYSCALL): Likewise.
[__LINUX_KERNEL_VERSION < 0x030e00] (__ASSUME_UTIMES): Undefine.
* sysdeps/unix/sysv/linux/ia64/kernel-features.h
(__ASSUME_UTIMES): Do not define.
(__ASSUME_PSELECT): Likewise.
(__ASSUME_PPOLL): Likewise.
(__ASSUME_O_CLOEXEC): Likewise.
(__ASSUME_SOCK_CLOEXEC): Likewise.
(__ASSUME_IN_NONBLOCK): Likewise.
(__ASSUME_PIPE2): Likewise.
(__ASSUME_EVENTFD2): Likewise.
(__ASSUME_SIGNALFD4): Likewise.
(__ASSUME_DUP3): Likewise.
* sysdeps/unix/sysv/linux/m68k/kernel-features.h
(__ASSUME_UTIMES): Likewise.
(__ASSUME_O_CLOEXEC): Likewise.
(__ASSUME_SOCK_CLOEXEC): Likewise.
(__ASSUME_IN_NONBLOCK): Likewise.
(__ASSUME_PIPE2): Likewise.
(__ASSUME_EVENTFD2): Likewise.
(__ASSUME_SIGNALFD4): Likewise.
(__ASSUME_DUP3): Likewise.
* sysdeps/unix/sysv/linux/microblaze/kernel-features.h
(__ASSUME_UTIMES): Likewise.
(__ASSUME_O_CLOEXEC): Likewise.
(__ASSUME_SOCK_CLOEXEC): Likewise.
(__ASSUME_IN_NONBLOCK): Likewise.
(__ASSUME_PIPE2): Likewise.
(__ASSUME_EVENTFD2): Likewise.
(__ASSUME_SIGNALFD4): Likewise.
(__ASSUME_DUP3): Likewise.
* sysdeps/unix/sysv/linux/mips/kernel-features.h (__ASSUME_IPC64):
Likewise.
(__ASSUME_UTIMES): Likewise.
(__ASSUME_EVENTFD2): Likewise.
(__ASSUME_SIGNALFD4): Likewise.
* sysdeps/unix/sysv/linux/tile/kernel-features.h
(__ASSUME_O_CLOEXEC): Likewise.
(__ASSUME_SOCK_CLOEXEC): Likewise.
(__ASSUME_IN_NONBLOCK): Likewise.
(__ASSUME_PIPE2): Likewise.
(__ASSUME_EVENTFD2): Likewise.
(__ASSUME_SIGNALFD4): Likewise.
(__ASSUME_DUP3): Likewise.
(__ASSUME_UTIMES): Undefine.
Comments
On Mon, May 12, 2014 at 11:53:24PM +0000, Joseph S. Myers wrote:
> This patch reduces duplication between different architectures'
> kernel-features.h files by making the architecture-independent file
> define various macros unconditionally (instead of only for a
> particular list of architectures), with the architecture-specific
> files then undefining the macros if necessary.
>
> Specifically, __ASSUME_O_CLOEXEC (O_CLOEXEC flag to open) and
> __ASSUME_SOCK_CLOEXEC (SOCK_NONBLOCK and SOCK_CLOEXEC flags to socket)
> are supported on all architectures as of 2.6.32 or the minimum kernel
> version for the architecture if later. For __ASSUME_IN_NONBLOCK,
> __ASSUME_PIPE2, __ASSUME_EVENTFD2, __ASSUME_SIGNALFD4 and
> __ASSUME_DUP3, the relevant syscalls were added for alpha in 2.6.33
> but otherwise the features are available as of 2.6.32. For
> __ASSUME_UTIMES, support is everywhere in 2.6.32 except for
> asm-generic architectures and hppa.
>
> Although those were the main cases of duplication among
> kernel-features.h files, some other cases of unnecessary definitions
> were also cleaned up: the hppa file defined various macros that were
> either no longer used at all, or defined by the main file by default
> anyway, the ia64 file had duplicative definitions of __ASSUME_PSELECT
> and __ASSUME_PPOLL, while mips had such a definition of
> __ASSUME_IPC64.
>
> Really, rather than being defined in the main file then undefined for
> asm-generic architectures, __ASSUME_UTIMES should become an
> hppa-specific macro. Given that __ASSUME_ATFCTS and
> __ASSUME_UTIMENSAT are now always true, the only live __ASSUME_UTIMES
> conditional is in sysdeps/unix/sysv/linux/utimes.c, which is not used
> for asm-generic architectures. I think the desired state would be an
> hppa-specific file (that includes sysdeps/unix/sysv/linux/utimes.c if
> __ASSUME_UTIMES, and otherwise has fallback code), with the fallback
> code being removed from the main utimes.c. But I think that's most
> reasonably a separate cleanup once __ASSUME_ATFCTS and
> __ASSUME_UTIMESAT have both had conditional code cleaned up.
>
> Given this patch, I think it's straightforward to move non-ex-ports
> architectures to having their own kernel-features.h files, like
> ex-ports architectures, rather than conditionals in the main file
> (i.e., such a move won't require the architecture-specific file to
> contain anything that isn't genuinely architecture-specific), and
> would encourage architecture maintainers to do so.
>
> Tested x86_64 that the installed shared libraries are unchanged by
> this patch. Note that on some architectures this *will* cause
> __ASSUME_* macros to be defined in cases where they weren't previously
> but should have been (but this is just optimization, not a fix to a
> user-visible bug, so doesn't need a bug report in Bugzilla).
>
looks ok.
@@ -25,14 +25,10 @@
features. */
#define __ASSUME_ACCEPT4_SYSCALL 1
-#define __ASSUME_DUP3 1
-#define __ASSUME_EVENTFD2 1
-#define __ASSUME_IN_NONBLOCK 1
-#define __ASSUME_O_CLOEXEC 1
-#define __ASSUME_PIPE2 1
#define __ASSUME_RECVMMSG_SYSCALL 1
#define __ASSUME_SENDMMSG_SYSCALL 1
-#define __ASSUME_SIGNALFD4 1
-#define __ASSUME_SOCK_CLOEXEC 1
#include_next <kernel-features.h>
+
+/* asm-generic architectures do not have the utimes syscall. */
+#undef __ASSUME_UTIMES
@@ -20,26 +20,6 @@
#ifndef _KERNEL_FEATURES_H
#define _KERNEL_FEATURES_H 1
-#define __ASSUME_UTIMES 1
-
-/* Support for the O_CLOEXEC flag was added for alpha in 2.6.23. */
-#define __ASSUME_O_CLOEXEC 1
-
-/* Support for various CLOEXEC and NONBLOCK flags was added for alpha after
- 2.6.33-rc1. */
-#if __LINUX_KERNEL_VERSION >= 0x020621
-# define __ASSUME_SOCK_CLOEXEC 1
-# define __ASSUME_IN_NONBLOCK 1
-#endif
-
-/* Support for the pipe2, eventfd2, signalfd4 syscalls was added for alpha
- after 2.6.33-rc1. */
-#if __LINUX_KERNEL_VERSION >= 0x020621
-# define __ASSUME_PIPE2 1
-# define __ASSUME_EVENTFD2 1
-# define __ASSUME_SIGNALFD4 1
-#endif
-
/* Support for recvmmsg was added for alpha in 2.6.33. */
#if __LINUX_KERNEL_VERSION >= 0x020621
# define __ASSUME_RECVMMSG_SYSCALL 1
@@ -67,10 +47,15 @@
/* Support for fsyncdata was added for alpha after 2.6.21. */
#define __ASSUME_FDATASYNC 1
-/* Support for preadv and pwritev was added for alpha in 2.6.33. */
+/* Support for various syscalls was added for alpha in 2.6.33. */
#if __LINUX_KERNEL_VERSION < 0x020621
# undef __ASSUME_PREADV
# undef __ASSUME_PWRITEV
+# undef __ASSUME_IN_NONBLOCK
+# undef __ASSUME_PIPE2
+# undef __ASSUME_EVENTFD2
+# undef __ASSUME_SIGNALFD4
+# undef __ASSUME_DUP3
#endif
#endif /* _KERNEL_FEATURES_H */
@@ -17,16 +17,9 @@
License along with the GNU C Library. If not, see
<http://www.gnu.org/licenses/>. */
-/* The utimes syscall was added before 2.6.1. */
-#define __ASSUME_UTIMES 1
-
/* The signal frame layout changed in 2.6.18. */
#define __ASSUME_SIGFRAME_V2 1
-/* Support for the eventfd2 and signalfd4 syscalls was added in 2.6.27. */
-#define __ASSUME_EVENTFD2 1
-#define __ASSUME_SIGNALFD4 1
-
/* Support for the recvmmsg syscall was added in 2.6.33. */
#if __LINUX_KERNEL_VERSION >= 0x020621
# define __ASSUME_RECVMMSG_SYSCALL 1
@@ -18,15 +18,6 @@
<http://www.gnu.org/licenses/>. */
-/* There are an infinite number of PA-RISC kernel versions numbered
- 2.4.0. But they've not really been released as such. We require
- and expect the final version here. */
-#define __ASSUME_32BITUIDS 1
-#define __ASSUME_TRUNCATE64_SYSCALL 1
-#define __ASSUME_IPC64 1
-#define __ASSUME_ST_INO_64_BIT 1
-#define __ASSUME_GETDENTS64_SYSCALL 1
-
/* PA-RISC 2.6.9 kernels had the first LWS CAS support */
#define __ASSUME_LWS_CAS 1
@@ -42,3 +33,8 @@
#endif
#include_next <kernel-features.h>
+
+/* hppa did not get the utimes syscall until 3.14. */
+#if __LINUX_KERNEL_VERSION < 0x030e00
+# undef __ASSUME_UTIMES
+#endif
@@ -20,25 +20,6 @@
#ifndef _KERNEL_FEATURES_H
#define _KERNEL_FEATURES_H 1
-/* The utimes syscall has been available for some architectures
- forever. */
-#define __ASSUME_UTIMES 1
-
-/* pselect/ppoll were introduced just after 2.6.16-rc1. */
-#define __ASSUME_PSELECT 1
-#define __ASSUME_PPOLL 1
-
-/* Support for various CLOEXEC and NONBLOCK flags was added in 2.6.23. */
-#define __ASSUME_O_CLOEXEC 1
-
-/* Support for various CLOEXEC and NONBLOCK flags was added in 2.6.27. */
-#define __ASSUME_SOCK_CLOEXEC 1
-#define __ASSUME_IN_NONBLOCK 1
-#define __ASSUME_PIPE2 1
-#define __ASSUME_EVENTFD2 1
-#define __ASSUME_SIGNALFD4 1
-#define __ASSUME_DUP3 1
-
/* Support for the recvmmsg syscall was added in 2.6.33. */
#if __LINUX_KERNEL_VERSION >= 0x020621
# define __ASSUME_RECVMMSG_SYSCALL 1
@@ -76,14 +76,7 @@
forever. For x86 it was introduced after 2.5.75, for x86-64,
ppc, and ppc64 it was introduced in 2.6.0-test3, for s390 it was
introduced in 2.6.21-rc5. */
-#if defined __sparc__ \
- || defined __i386__ \
- || defined __x86_64__ \
- || defined __powerpc__ \
- || defined __sh__ \
- || defined __s390__
-# define __ASSUME_UTIMES 1
-#endif
+#define __ASSUME_UTIMES 1
/* pselect/ppoll were introduced just after 2.6.16-rc1. On x86_64 and
SH this appeared first in 2.6.19-rc1, on ia64 in 2.6.22-rc1. */
@@ -123,27 +116,21 @@
only after 2.6.23-rc1. */
#define __ASSUME_FALLOCATE 1
-/* Support for various CLOEXEC and NONBLOCK flags was added for x86,
- x86-64, PPC, IA-64, SPARC< and S390 in 2.6.23. */
-#if (defined __i386__ || defined __x86_64__ || defined __powerpc__ \
- || defined __sparc__ || defined __s390__)
-# define __ASSUME_O_CLOEXEC 1
-#endif
+/* Support for various CLOEXEC and NONBLOCK flags was added in
+ 2.6.23. */
+#define __ASSUME_O_CLOEXEC 1
/* Support for ADJ_OFFSET_SS_READ was added in 2.6.24. */
#define __ASSUME_ADJ_OFFSET_SS_READ 1
-/* Support for various CLOEXEC and NONBLOCK flags was added for x86,
- x86-64, PPC, IA-64, and SPARC in 2.6.27. */
-#if (defined __i386__ || defined __x86_64__ || defined __powerpc__ \
- || defined __sparc__ || defined __s390__)
-# define __ASSUME_SOCK_CLOEXEC 1
-# define __ASSUME_IN_NONBLOCK 1
-# define __ASSUME_PIPE2 1
-# define __ASSUME_EVENTFD2 1
-# define __ASSUME_SIGNALFD4 1
-# define __ASSUME_DUP3 1
-#endif
+/* Support for various CLOEXEC and NONBLOCK flags was added in
+ 2.6.27. */
+#define __ASSUME_SOCK_CLOEXEC 1
+#define __ASSUME_IN_NONBLOCK 1
+#define __ASSUME_PIPE2 1
+#define __ASSUME_EVENTFD2 1
+#define __ASSUME_SIGNALFD4 1
+#define __ASSUME_DUP3 1
/* Support for accept4 functionality was added in 2.6.28, but for some
architectures using a separate syscall rather than socketcall that
@@ -20,20 +20,6 @@
/* m68k uses socketcall. */
#define __ASSUME_SOCKETCALL 1
-/* Many syscalls were added in 2.6.10 for m68k. */
-#define __ASSUME_UTIMES 1
-
-/* Support for various CLOEXEC and NONBLOCK flags was added 2.6.23. */
-#define __ASSUME_O_CLOEXEC 1
-
-/* Support for various CLOEXEC and NONBLOCK flags was added in 2.6.27. */
-#define __ASSUME_SOCK_CLOEXEC 1
-#define __ASSUME_IN_NONBLOCK 1
-#define __ASSUME_PIPE2 1
-#define __ASSUME_EVENTFD2 1
-#define __ASSUME_SIGNALFD4 1
-#define __ASSUME_DUP3 1
-
#include_next <kernel-features.h>
/* These syscalls were added only in 3.0 for m68k. */
@@ -19,16 +19,6 @@
/* MicroBlaze uses socketcall. */
#define __ASSUME_SOCKETCALL 1
-/* MicroBlaze glibc support starts with 2.6.30, guaranteeing many kernel features. */
-#define __ASSUME_UTIMES 1
-#define __ASSUME_O_CLOEXEC 1
-#define __ASSUME_SOCK_CLOEXEC 1
-#define __ASSUME_IN_NONBLOCK 1
-#define __ASSUME_PIPE2 1
-#define __ASSUME_EVENTFD2 1
-#define __ASSUME_SIGNALFD4 1
-#define __ASSUME_DUP3 1
-
/* Support for the accept4 and recvmmsg syscalls was added in 2.6.33. */
#if __LINUX_KERNEL_VERSION >= 0x020621
# define __ASSUME_ACCEPT4_SYSCALL 1
@@ -19,16 +19,6 @@
#include <sgidefs.h>
-/* MIPS platforms had IPC64 all along. */
-#define __ASSUME_IPC64 1
-
-/* MIPS had the utimes syscall by 2.6.0. */
-#define __ASSUME_UTIMES 1
-
-/* Support for the eventfd2 and signalfd4 syscalls was added in 2.6.27. */
-#define __ASSUME_EVENTFD2 1
-#define __ASSUME_SIGNALFD4 1
-
/* Support for the accept4 syscall was added in 2.6.31. */
#define __ASSUME_ACCEPT4_SYSCALL 1
@@ -18,14 +18,7 @@
/* TILE glibc support starts with 2.6.36, guaranteeing many kernel features. */
-#define __ASSUME_O_CLOEXEC 1
-#define __ASSUME_SOCK_CLOEXEC 1
-#define __ASSUME_IN_NONBLOCK 1
-#define __ASSUME_PIPE2 1
-#define __ASSUME_EVENTFD2 1
-#define __ASSUME_SIGNALFD4 1
#define __ASSUME_ACCEPT4_SYSCALL 1
-#define __ASSUME_DUP3 1
#define __ASSUME_RECVMMSG_SYSCALL 1
/* Support for the sendmmsg syscall was added in 3.0. */
@@ -35,6 +28,9 @@
#include_next <kernel-features.h>
+/* asm-generic architectures do not have the utimes syscall. */
+#undef __ASSUME_UTIMES
+
/* Define this if your 32-bit syscall API requires 64-bit register
pairs to start with an even-number register. */
#define __ASSUME_ALIGNED_REGISTER_PAIRS 1