sparc syscall stuff
sparc syscall stuff - killing pointless wrappers, conversions to
{COMPAT_,}SYSCALL_DEFINE.
Conflicts in arch/sparc/kernel/sys_sparc32.c are not hard to resolve,
but keep in mind that casts to u64 are needed there. I'd pushed a variant of resolution into #test-merge.
The following changes since commit 0c8efd610b58cb23cefdfa12015799079aef94ae:
Linux 4.16-rc5 (2018-03-11 17:25:09 -0700)
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git misc.sparc
for you to fetch changes up to 47db69752b22de2ac315d094fd09fac31b0d3bfc:
sparc: get rid of asm wrapper for nis_syscall() (2018-03-20 12:05:17 -0400)
----------------------------------------------------------------
Al Viro (11):
sparc: get rid of sys_sparc_pipe() wrappers
sparc: kill useless SIGN... wrappers
sparc: get rid of remaining SIGN... wrappers
sparc: bury a zombie extern that had been that way for twenty years
sparc: trivial conversions to {COMPAT_,}SYSCALL_DEFINE()
sparc: get rid of memory_ordering(2) wrapper
switch sparc_remap_file_pages() to SYSCALL_DEFINE convert compat sync_file_range() to COMPAT_SYSCALL_DEFINE
sparc: switch compat pread64 and pwrite64 to COMPAT_SYSCALL_DEFINE
sparc: switch compat {f,}truncate64() to COMPAT_SYSCALL_DEFINE
sparc: get rid of asm wrapper for nis_syscall()
arch/sparc/kernel/entry.S | 15 ------- arch/sparc/kernel/sys32.S | 38 ------------------ arch/sparc/kernel/sys_sparc32.c | 87 ++++++++++++++++------------------------ arch/sparc/kernel/sys_sparc_32.c | 31 +++++++------- arch/sparc/kernel/sys_sparc_64.c | 18 ++++----- arch/sparc/kernel/syscalls.S | 9 ----- arch/sparc/kernel/systbls.h | 53 ++++++++++++------------ arch/sparc/kernel/systbls_32.S | 2 +- arch/sparc/kernel/systbls_64.S | 24 +++++------ 9 files changed, 97 insertions(+), 180 deletions(-)
Quoted 4 lines · Al Viro
Conflicts in
arch/sparc/kernel/sys_sparc32.care not hard to resolve,
but keep in mind that casts to u64 are needed there. I'd pushed a variant of resolution into #test-merge.
You say "keep in mind that casts to u64 are needed there", but then you seem to be missing them yourself in your test merge (see [f]truncate64).
I only noticed because I decided to check my merge against yours. I strive to always do the merge resolution, but if people give me a test merge, I then double-check against it after-the-fack.
Hmm?
Linus
Quoted 10 lines · Linus Torvalds
Quoted 4 lines · Al Viro
Conflicts in
arch/sparc/kernel/sys_sparc32.care not hard to resolve,
but keep in mind that casts to u64 are needed there. I'd pushed a variant of resolution into #test-merge.You say "keep in mind that casts to u64 are needed there", but then you seem to be missing them yourself in your test merge (see [f]truncate64).
Quoted 5 lines
I only noticed because I decided to check my merge against yours. I strive to always do the merge resolution, but if people give me a test merge, I then double-check against it after-the-fack.
Hmm?
Merge fuckup; mine ;-/ FWIW, the corresponding part in my branch (in 3d0e354e4e51) is
-asmlinkage long sys32_truncate64(const char __user * path, unsigned long high, unsigned long low) +COMPAT_SYSCALL_DEFINE3(truncate64, const char __user *, path, u32, high, u32, low)
{
- if ((int)high < 0)
- return -EINVAL;
- else
- return sys_truncate(path, (high << 32) | low);
+ return sys_truncate(path, ((u64)high << 32) | low);
}
linux-next resolution is the right one -
git diff v4.16..HEAD -- arch/sparc/kernel/sys_sparc32.c
gives
-asmlinkage long sys32_truncate64(const char __user * path, unsigned long high, unsigned long low) +COMPAT_SYSCALL_DEFINE3(truncate64, const char __user *, path, u32, high, u32, low)
{
- if ((int)high < 0)
- return -EINVAL;
- else
- return sys_truncate(path, (high << 32) | low);
+ return ksys_truncate(path, ((u64)high << 32) | low);
}
etc. in there. Again, my apologies.