Simplify close(2) EINTR handling.
This doesn't affect code like Chrome that correctly ignores EINTR on
close, makes code that tries TEMP_FAILURE_RETRY work (where before it might
have closed a different fd and appeared to succeed, or had a bogus EBADF),
and makes "goto fail" code work (instead of mistakenly assuming that EINTR
means that the close failed).
Who loses? Anyone actively trying to detect that they caught a signal while
in close(2). I don't think those people exist, and I think they have better
alternatives available.
Bug: https://code.google.com/p/chromium/issues/detail?id=269623
Bug: http://b/20501816
Change-Id: I11e2f66532fe5d1b0082b2433212e24bdda8219b
diff --git a/libc/arch-x86/syscalls/close.S b/libc/arch-x86/syscalls/___close.S
similarity index 89%
rename from libc/arch-x86/syscalls/close.S
rename to libc/arch-x86/syscalls/___close.S
index f6cce62..796944b 100644
--- a/libc/arch-x86/syscalls/close.S
+++ b/libc/arch-x86/syscalls/___close.S
@@ -2,7 +2,7 @@
#include <private/bionic_asm.h>
-ENTRY(close)
+ENTRY(___close)
pushl %ebx
.cfi_def_cfa_offset 8
.cfi_rel_offset ebx, 0
@@ -18,4 +18,5 @@
1:
popl %ebx
ret
-END(close)
+END(___close)
+.hidden ___close