Re: [PATCH v2 16/16] objtool/powerpc: Add --mcount specific implementation

Le 05/09/2022 à 22:43, Segher Boessenkool a écrit :

Hi!

On Mon, Sep 05, 2022 at 04:15:07PM +0530, Naveen N. Rao wrote:

quoted

Segher Boessenkool wrote:

quoted

quoted

quoted

+		if ((insn & 3) == 1) {
+			*type = INSN_CALL;
+			*immediate = insn & 0x3fffffc;
+			if (*immediate & 0x2000000)
+				*immediate -= 0x4000000;
+		}
+		break;
+	}

Does this handle AA=1 correctly at all?  That is valid both with and
without relocations, just like AA=0.  Same for AA=1 LK=0 btw.

If you only handle AA=0, the code should explicitly test for that.

The code does test for AA=0 LK=1 with the if statement there?

Yes, but that is not what I said :-)

It may be fine to not *handle* AA=1 at all, but the code should at least
scream bloody murder when it encounters it anyway :-)

By the way, I proposed a cleanup patch that handles it, see 
https://patchwork.ozlabs.org/project/linuxppc-dev/patch/ebe11b73d1015a17034a2c4bedf093fa57f5d29f.1662032631.git.christophe.leroy@csgroup.eu/

Christophe