Hi,
I just started using Coccinelle yesterday, and I can already see
it saving me a lot of time and agony. Thank you.
I ran into a problem, and I think that this patch might fix it,
hopefully without causing other problems. I did some sanity
checking on my kernel tree, and it seems to be fine.
There are quite a few variables named "far" in the kernel, e.g.,
arch/arm64/kvm/inject_fault.c. Coccinelle has special treatment
for "far" as being linkage related, which causes parse errors in
their presence. I've grepped for "far" in the kernel tree, and
haven't noticed where it's used like that, but I could have
missed it.
To reproduce:
cat > test.c << EOF
int main(void)
{
int far = 0;
int x;
x = 10;
return x;
}
EOF
cat > test.cocci << EOF
@@
identifier x;
@@
- x
+ y
EOF
spatch --sp-file test.cocci test.c --debug --verbose-parsing
Signed-off-by: Fuad Tabba <ta...@google.com>
---
scripts/coccicheck/cocci/notand.h | 1 -
standard.h | 1 -
2 files changed, 2 deletions(-)
diff --git a/scripts/coccicheck/cocci/notand.h
b/scripts/coccicheck/cocci/notand.h
index 3da8c303..91fa6c96 100644
--- a/scripts/coccicheck/cocci/notand.h
+++ b/scripts/coccicheck/cocci/notand.h
@@ -302,7 +302,6 @@
#define fastcall
#define asmlinkage
-#define far
#define SK_FAR
// pb
diff --git a/standard.h b/standard.h
index 7a7f96ea..936b19c3 100644
--- a/standard.h
+++ b/standard.h
@@ -298,7 +298,6 @@
#define fastcall
#define asmlinkage
-#define far
#define SK_FAR
// pb
--
2.31.1.368.gbe11c130af-goog
_______________________________________________
Cocci mailing list
Cocci@systeme.lip6.fr
https://systeme.lip6.fr/mailman/listinfo/cocci
