Small patch
- From: Dmitri Shubin <sbn@xxxxxxxxxxx>
- To: luajit@xxxxxxxxxxxxx
- Date: Mon, 07 Jan 2013 15:55:25 +0400
Hi Mike,
While upgrading from 2.0.0-beta10 to 2.0.0 and re-applying our
Solaris/x64 support patches I found the following changes that probably
can be safely applied:
$ git diff
diff --git a/src/lj_mcode.c b/src/lj_mcode.c
index 34405b5..7526e99 100644
--- a/src/lj_mcode.c
+++ b/src/lj_mcode.c
@@ -223,8 +223,8 @@ static void *mcode_alloc(jit_State *J, size_t sz)
if (mcode_validptr(p)) {
if ((uintptr_t)p + sz - target < range || target - (uintptr_t)p
< range)
return p;
- mcode_free(J, p, sz); /* Free badly placed area. */
}
+ mcode_free(J, p, sz); /* Free badly placed area. */
}
/* Next try probing pseudo-random addresses. */
do {
diff --git a/src/vm_x86.dasc b/src/vm_x86.dasc
index e2b6a07..cc10ae7 100644
--- a/src/vm_x86.dasc
+++ b/src/vm_x86.dasc
@@ -6151,7 +6151,7 @@ static void emit_asm_debug(BuildCtx *ctx)
"\t.align " SZPTR "\n"
".LEFDE1:\n\n", (int)ctx->codesz - fcofs);
#endif
-#if (defined(__sun__) && defined(__svr4__)) || defined(__solaris_)
+#if (defined(__sun__) && defined(__svr4__)) || defined(__solaris__)
fprintf(ctx->fp, "\t.section .eh_frame,\"aw\",@progbits\n");
#else
fprintf(ctx->fp, "\t.section .eh_frame,\"a\",@progbits\n");
The first one looks like memory leak, and the second one like a typo.
BTW where does this '__solaris__' macro come from, I haven't ever seen
it before?
Thanks!
Other related posts:
