Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Christophe Leroy <christophe.leroy@c-s.fr>
Subject: [PATCH] powerpc: optimise WARN_ON()
To:     Benjamin Herrenschmidt <benh@kernel.crashing.org>,
        Paul Mackerras <paulus@samba.org>,
        Michael Ellerman <mpe@ellerman.id.au>,
        Segher Boessenkool <segher@kernel.crashing.org>
Cc:     linux-kernel@vger.kernel.org, linuxppc-dev@lists.ozlabs.org
Message-Id: <20190817090442.C5FEF106613@localhost.localdomain>
Date:   Sat, 17 Aug 2019 09:04:42 +0000 (UTC)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Unlike BUG_ON(x), WARN_ON(x) uses !!(x) as the trigger
of the t(d/w)nei instruction instead of using directly the
value of x.

This leads to GCC adding unnecessary pair of addic/subfe. This was
revealed after adding a WARN_ON() on top of clear_page() in order
to detect misaligned destination:

@@ -49,6 +51,8 @@ static inline void clear_page(void *addr)
 {
 	unsigned int i;

+	WARN_ON((unsigned long)addr & (L1_CACHE_BYTES - 1));
+
 	for (i = 0; i < PAGE_SIZE / L1_CACHE_BYTES; i++, addr += L1_CACHE_BYTES)
 		dcbz(addr);
 }

This resulted on:

0000019c <clear_user_page>:
 19c:	54 68 06 fe 	clrlwi  r8,r3,27
 1a0:	31 48 ff ff 	addic   r10,r8,-1
 1a4:	7d 4a 41 10 	subfe   r10,r10,r8
 1a8:	0f 0a 00 00 	twnei   r10,0
 1ac:	39 20 00 80 	li      r9,128
 1b0:	7d 29 03 a6 	mtctr   r9
 1b4:	7c 00 1f ec 	dcbz    0,r3
 1b8:	38 63 00 20 	addi    r3,r3,32
 1bc:	42 00 ff f8 	bdnz    1b4 <clear_user_page+0x18>
 1c0:	7c a3 2b 78 	mr      r3,r5
 1c4:	48 00 00 00 	b       1c4 <clear_user_page+0x28>
			1c4: R_PPC_REL24	flush_dcache_page

By using (x) instead of !!(x) like BUG_ON() does, the additional
instructions go away:

0000019c <clear_user_page>:
 19c:	54 6a 06 fe 	clrlwi  r10,r3,27
 1a0:	0f 0a 00 00 	twnei   r10,0
 1a4:	39 20 00 80 	li      r9,128
 1a8:	7d 29 03 a6 	mtctr   r9
 1ac:	7c 00 1f ec 	dcbz    0,r3
 1b0:	38 63 00 20 	addi    r3,r3,32
 1b4:	42 00 ff f8 	bdnz    1ac <clear_user_page+0x10>
 1b8:	7c a3 2b 78 	mr      r3,r5
 1bc:	48 00 00 00 	b       1bc <clear_user_page+0x20>
			1bc: R_PPC_REL24	flush_dcache_page

Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
---
 arch/powerpc/include/asm/bug.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/powerpc/include/asm/bug.h b/arch/powerpc/include/asm/bug.h
index fed7e6241349..77074582fe65 100644
--- a/arch/powerpc/include/asm/bug.h
+++ b/arch/powerpc/include/asm/bug.h
@@ -107,7 +107,7 @@
 		: : "i" (__FILE__), "i" (__LINE__),		\
 		  "i" (BUGFLAG_WARNING|BUGFLAG_TAINT(TAINT_WARN)),\
 		  "i" (sizeof(struct bug_entry)),		\
-		  "r" (__ret_warn_on));				\
+		  "r" ((__force long)(x)));			\
 	}							\
 	unlikely(__ret_warn_on);				\
 })
-- 
2.17.1