checksyscalls.sh is run at every "make" run while building the kernel,
even if no files have changed. I looked at where we spend time in
a trivial empty rebuild and found checksyscalls.sh to be a source
of noticeable overhead, as it spawns a lot of child processes just
to call 'cat' copying from stdin to stdout, once for each of the
over 400 x86 syscalls.
Using a shell-builtin (echo) instead of the external command gives
us a 13x speedup:
Before After
real 0m1.018s real 0m0.077s
user 0m0.068s user 0m0.048s
sys 0m0.156s sys 0m0.024s
The time it took to rebuild a single file on my machine dropped
from 5.5 seconds to 4.5 seconds.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
sys_arch_prctl is only provided on x86, and there is no reason
to add it elsewhere. However, including it on the 32-bit syscall
table caused a warning for most configurations on non-x86:
:1328:2: warning: #warning syscall arch_prctl not implemented [-Wcpp]
This adds an exception to the syscall table checking script.
Fixes: 79170fda31 ("x86/syscalls/32: Wire up arch_prctl on x86-32")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Kyle Huey <khuey@kylehuey.com>
Link: http://lkml.kernel.org/r/20170323151904.706286-1-arnd@arndb.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
The build time generated syscall definitions are entry code related, move
them into the arch/x86/entry/ directory.
Cc: Borislav Petkov <bp@alien8.de>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The new renameat2 syscall provides all the functionality of renameat
with an additional flags argument, so make renameat optional so that
future architectures can omit it without getting a warning.
This patch doesn't affect existing architectures.
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Cc: linux-arch@vger.kernel.org
"echo" doesn't read from stdin, therefore the checksyscalls script didn't
warn about not implemented system calls anymore since 29dc54c6
("checksyscalls: Use arch/x86/syscalls/syscall_32.tbl as source").
Use "cat" instead of "echo" which handles this correctly.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Michal Marek <mmarek@suse.cz>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Cc: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Use the new arch/x86/syscalls/syscall_32.tbl file as source instead of
arch/x86/include/asm/unistd_32.h.
Cc: Michal Marek <mmarek@suse.cz>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
The usage help in the comments
- refers to the wrong script name,
- doesn't mention that $srctree must be set.
Hence correct the script name, and derive the source tree path from the script
path, so we no longer need to rely on $srctree being set by the caller.
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Michal Marek <mmarek@suse.cz>
The initial pass at the generic ABI assumed that wait4() could be
easily expressed using waitid(). Although it's true that wait4()
can be built on waitid(), it's awkward enough that it makes more
sense to continue to include wait4 in the generic syscall ABI.
Since there is already a deprecated wait4 in the ABI, this change
converts that wait4 into old_wait, and puts wait4 in the next
available slot for new supported syscalls, after the platform-specific
syscalls at number 260.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
A new architecture should only define a minimal set of system
calls while still providing the full functionality. This version
of unistd.h has gone through intensive review to make sure that
by default it only enables syscalls that do not already have
a more featureful replacement.
It is modeled after the x86-64 version of unistd.h, which unifies
the syscall number definition and the actual system call table
in a single file, in order to keep them synchronized much more
easily.
This first version still keeps legacy system call definitions
around, guarded by various #ifdefs, and with numbers larger
than 1024. The idea behind this is to make it easier for
new architectures to transition from a full list to the reduced
set. In particular, the new microblaze architecture that should
migrate to using the generic ABI headers can at least use an
existing uClibc source tree without major rewrites during the
conversion.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Make the checksyscalls script work even on systems where sed is non-gnu.
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Move the headers to include/asm-x86 and fixup the
header install make rules
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Not all the world is an i386. Many architectures need 64-bit arguments to be
aligned in suitable pairs of registers, and the original
sys_sync_file_range(int, loff_t, loff_t, int) was therefore wasting an
argument register for padding after the first integer. Since we don't
normally have more than 6 arguments for system calls, that left no room for
the final argument on some architectures.
Fix this by introducing sys_sync_file_range2(int, int, loff_t, loff_t) which
all fits nicely. In fact, ARM already had that, but called it
sys_arm_sync_file_range. Move it to fs/sync.c and rename it, then implement
the needed compatibility routine. And stop the missing syscall check from
bitching about the absence of sys_sync_file_range() if we've implemented
sys_sync_file_range2() instead.
Tested on PPC32 and with 32-bit and 64-bit userspace on PPC64.
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Acked-by: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Most system calls seems to get added to i386 first. This patch
automatically generates a warning for any new system call which is
implemented on i386 but not the architecture currently being compiled.
On PowerPC at the moment, for example, it results in these warnings:
init/missing_syscalls.h:935:3: warning: #warning syscall sync_file_range not implemented
init/missing_syscalls.h:947:3: warning: #warning syscall getcpu not implemented
init/missing_syscalls.h:950:3: warning: #warning syscall epoll_pwait not implemented
The file scripts/checksyscalls.sh list a number of legacy system calls
that are ignored because they only makes sense on i386 systems.
Other contributors to this patch are Russell King <rmk+lkml@arm.linux.org.uk>
and Stéphane Jourdois <kwisatz@rubis.org>
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
