[PATCH] powerpc/64s: Fix possible corruption on big endian due to pgd/pud_present()

Michael Ellerman mpe at ellerman.id.au
Wed Feb 20 22:18:38 AEDT 2019


Segher Boessenkool <segher at kernel.crashing.org> writes:
> On Mon, Feb 18, 2019 at 11:49:18AM +1100, Michael Ellerman wrote:
>> Balbir Singh <bsingharora at gmail.com> writes:
>> > Fair enough, my point was that the compiler can help out. I'll see what
>> > -Wconversion finds on my local build :)
>> 
>> I get about 43MB of warnings here :)
>
> Yes, -Wconversion complains about a lot of things that are idiomatic C.
> There is a reason -Wconversion is not in -Wall or -Wextra.

Actually a lot of those go away when I add -Wno-sign-conversion.

And what's left seems mostly reasonable, they all indicate the
possibility of a bug I think.

In fact this works and would have caught the bug:

diff --git a/arch/powerpc/include/asm/book3s/64/pgtable.h b/arch/powerpc/include/asm/book3s/64/pgtable.h
index d8c8d7c9df15..3114e3f368e2 100644
--- a/arch/powerpc/include/asm/book3s/64/pgtable.h
+++ b/arch/powerpc/include/asm/book3s/64/pgtable.h
@@ -904,7 +904,12 @@ static inline int pud_none(pud_t pud)
 
 static inline int pud_present(pud_t pud)
 {
+	__diag_push();
+	__diag_warn(GCC, 8, "-Wconversion", "ulong -> int");
+
 	return !!(pud_raw(pud) & cpu_to_be64(_PAGE_PRESENT));
+
+	__diag_pop();
 }
 
 extern struct page *pud_page(pud_t pud);



Obviously we're not going to instrument every function like that. But we
could start instrumenting particular files.

cheers


More information about the Linuxppc-dev mailing list