From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 33C21AD49 for ; Wed, 4 Jan 2023 16:14:53 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 8A16CC433D2; Wed, 4 Jan 2023 16:14:52 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1672848893; bh=sDwQr0Wn1l01rcLgMrwtjowdjVfEvBNikKRTBJ+29/4=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=Zl/Y5L2UorGl49ec98KkHxbK4EuVxQPrVRZgQBA+j9kiSXHTslg8lNzRL+ser/HT8 RgU4Xr1v3g4rFYvWEruSuEa+nhlIaFsdyGKQXrSgVKrkdSPocA0Cb5eoYUexjEik36 XLQNFNMgWskNmhvV91FcNK7LUuZkCk1p35yYWZN8= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Nathan Chancellor , Arnd Bergmann , Nick Desaulniers , "Russell King (Oracle)" Subject: [PATCH 6.1 100/207] ARM: 9256/1: NWFPE: avoid compiler-generated __aeabi_uldivmod Date: Wed, 4 Jan 2023 17:05:58 +0100 Message-Id: <20230104160515.087138892@linuxfoundation.org> X-Mailer: git-send-email 2.39.0 In-Reply-To: <20230104160511.905925875@linuxfoundation.org> References: <20230104160511.905925875@linuxfoundation.org> User-Agent: quilt/0.67 Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit From: Nick Desaulniers commit 3220022038b9a3845eea762af85f1c5694b9f861 upstream. clang-15's ability to elide loops completely became more aggressive when it can deduce how a variable is being updated in a loop. Counting down one variable by an increment of another can be replaced by a modulo operation. For 64b variables on 32b ARM EABI targets, this can result in the compiler generating calls to __aeabi_uldivmod, which it does for a do while loop in float64_rem(). For the kernel, we'd generally prefer that developers not open code 64b division via binary / operators and instead use the more explicit helpers from div64.h. On arm-linux-gnuabi targets, failure to do so can result in linkage failures due to undefined references to __aeabi_uldivmod(). While developers can avoid open coding divisions on 64b variables, the compiler doesn't know that the Linux kernel has a partial implementation of a compiler runtime (--rtlib) to enforce this convention. It's also undecidable for the compiler whether the code in question would be faster to execute the loop vs elide it and do the 64b division. While I actively avoid using the internal -mllvm command line flags, I think we get better code than using barrier() here, which will force reloads+spills in the loop for all toolchains. Link: https://github.com/ClangBuiltLinux/linux/issues/1666 Reported-by: Nathan Chancellor Reviewed-by: Arnd Bergmann Signed-off-by: Nick Desaulniers Tested-by: Nathan Chancellor Cc: stable@vger.kernel.org Signed-off-by: Russell King (Oracle) Signed-off-by: Greg Kroah-Hartman --- arch/arm/nwfpe/Makefile | 6 ++++++ 1 file changed, 6 insertions(+) --- a/arch/arm/nwfpe/Makefile +++ b/arch/arm/nwfpe/Makefile @@ -11,3 +11,9 @@ nwfpe-y += fpa11.o fpa11_cpdo.o fpa11 entry.o nwfpe-$(CONFIG_FPE_NWFPE_XP) += extended_cpdo.o + +# Try really hard to avoid generating calls to __aeabi_uldivmod() from +# float64_rem() due to loop elision. +ifdef CONFIG_CC_IS_CLANG +CFLAGS_softfloat.o += -mllvm -replexitval=never +endif