From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-wm1-f41.google.com (mail-wm1-f41.google.com [209.85.128.41]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 6448E34F27F for ; Mon, 24 Aug 2026 07:56:36 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.128.41 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1787558199; cv=none; b=rKb0VOtaJ7qxQBpegX3EQQ1foKHHTGTZw8kwjidmQwaq9hlg+1+FZLBZ0I7UWlE82xDtERdxODvlR3vY4buYmq9t7MJYc50bkRFX9euBTGRZBGoxPc8ZioaLqw29muZdkgne//DdxidamLrXNMFoYg4pME84nTOYtrYQSRHaoZs= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1787558199; c=relaxed/simple; bh=dqJHfAW5WwMPJOmX5fpLR/yJt+WjJzxp0+Y0nsk5AKI=; h=Date:From:To:Cc:Subject:Message-ID:In-Reply-To:References: MIME-Version:Content-Type; b=KYkZEnszkvCWCHMRCgEKZ2qIWxROf/cp3AOVjYQQQ/Wu/7YzDZAJb35lMZk1UCT2edZ1vlvbvpOQtveu7c/uKF0L22GDthyd2qHj+PEYsI5RZ4Miw11EjnMXIgMRoIve7NNXIzLNC9MmpSEYswBQFJvc3KR+VlstAZU9wOZ+RMU= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com; spf=pass smtp.mailfrom=gmail.com; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b=l8UWUC0U; arc=none smtp.client-ip=209.85.128.41 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=gmail.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b="l8UWUC0U" Received: by mail-wm1-f41.google.com with SMTP id 5b1f17b1804b1-4995b0343c1so22889925e9.3 for ; Mon, 24 Aug 2026 00:56:36 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20251104; t=1787558195; x=1788162995; darn=vger.kernel.org; h=content-transfer-encoding:content-type:mime-version:references :in-reply-to:message-id:subject:cc:to:from:date:from:to:cc:subject :date:message-id:reply-to:content-type; bh=IE8kFIz49p9hLyGen3p2a/60iDUMeHiZ+R4qjtbPgPM=; b=l8UWUC0UfjGeC1UHhHXWmn9qYjqPiI+M/wOmpdGl+dml0yAlip9IwFeTHyw5Q4/zY3 j9P5m7J17pnboG+EYWrRfFYAiawCD85B7j5Zwz26UYQmD2Yr5RVp7dWJlP9NUuCQSfNc XU91xRAi89VOk5N0QE8nPUKu66W0VYekYVXItvD/WGEUtO1BQsXyts6tNqoYO/OfBKDq hK0SOrqTiiLFGgnZQpGB2RXrv8t8ff6hC/aG31FP+qZwc9j8lgPJn9u5uFU1R+ghLOiq FxkP0MNEcie+m4Sqp2UcTHEJVnxf051bu8XkD+MfNwn5Fbw3OoX3gBdZuTC3d9H6Rrte gTOg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1787558195; x=1788162995; h=content-transfer-encoding:content-type:mime-version:references :in-reply-to:message-id:subject:cc:to:from:date:x-gm-gg :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to :content-type; bh=IE8kFIz49p9hLyGen3p2a/60iDUMeHiZ+R4qjtbPgPM=; b=VOu/qrBKuYlFiKbdgnYkdqh/Mh7dnIIiDM/h5eW9T3rnlz58L9EzJ7USFQsIBDA5Ia xYvlJrIRQbjzBj0UrMJzjkYX/f3fdrE38cVT7r0soy3VYDZH5fTo5CpFbLJrCEiSEJib eZfFNcdLJT5W/SzDBjg9ljcQ5+/gCLhvYkJ8X45iqv8WIJVodxWHy9GGXfd5hl1ixYY3 P0SPq2n8bAtmo6lxNJd05+OyZfVI31EQudNO6d/bG0tyFgiGGo/YCZ5RKRCFepqyK+x1 sbCxQz4TdEszzC6SkHyij5Lv9PXPzQeAGX0ASdNAd4M7MYAZrg/2VhNoiVeuhx8SMRUq MXjg== X-Forwarded-Encrypted: i=1; AHgh+RqSrm15K6Wbi1rUwTEIF9O5G8V6+azn7vt4wtit7FGtFeQ2VIOPxtTZ8W61t1DMFs5grKSGessU3h+/37I=@vger.kernel.org X-Gm-Message-State: AFuF++lzO5OvPWiJRypDKjs+xSLsxkyenqcO0ca5WTDktcNmt+ebQ13T 040QDpkyPcj2IwNRLKklxU8AKNA5c5S8XEdDh/KgGuw8eNTtxd/pg1gr X-Gm-Gg: AR+sD13TAICaM3yzPuUFx6LaubMkvGkDuhQoOgISGQF9lKdRD8UuKnEWkq1uzM8OUkB mbpATnqtKHk9Mu4RzH8WDSljQfiPURMkl3R5xTog6k+9yOVG5dMsX5EyreToQqq6j4TTPYnRNl2 nT2D/VT1KdoEc24wPuGjZs7dnJPBqr9Ui4FaLYGFGtMPEZfDY05sNsqybvMf0iOX8DIq8I1eaFu vZstGKQhuPM5QxxNxxFuP+C/OlI6sxe3VGuTQKY0N10AxU4z2XzZAv+xOrWXw6eMGIXctODqrxB 0tURPDy3bd8SGu+dp/h+FZkymh9xh+NFXw9RpF7U3Pv1FcIGPjy1TmocNJIhfYQ0O+ixMmDv9TP 300eE8CPA70AoPioZ0AQgEymwyYJOCHg2fosVDhwzLSb4wBd3+OA2wXdZWLRXGdGsh1ixBhYeF4 rw0aL/NhbOehIG9sanxXvV817lSRYLxoyDC/J8yRoNx7pcZYJbc8ze2PhlAnNN3VoxYw/4JyHHD L8NPlbvtdsJiFlD2Ga4F+OZZw== X-Received: by 2002:a05:600c:1914:b0:499:7a19:408b with SMTP id 5b1f17b1804b1-499c3302280mr147710725e9.11.1787558195033; Mon, 24 Aug 2026 00:56:35 -0700 (PDT) Received: from pumpkin (82-69-66-36.dsl.in-addr.zen.co.uk. [82.69.66.36]) by smtp.gmail.com with ESMTPSA id 5b1f17b1804b1-499b914bfcbsm114555395e9.9.2026.08.24.00.56.34 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 24 Aug 2026 00:56:34 -0700 (PDT) Date: Mon, 24 Aug 2026 08:56:32 +0100 From: David Laight To: Karl Mehltretter Cc: Miquel Raynal , Richard Weinberger , Vignesh Raghavendra , Andy Shevchenko , Lukas Wunner , linux-mtd@lists.infradead.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH] mtd: cfi_cmdset_0001: shrink do_write_buffer() stack frame Message-ID: <20260824085632.2dd08ba9@pumpkin> In-Reply-To: <20260824014703.56502-1-kmehltretter@gmail.com> References: <20260824014703.56502-1-kmehltretter@gmail.com> X-Mailer: Claws Mail 4.1.1 (GTK 3.24.38; arm-unknown-linux-gnueabihf) Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit On Mon, 24 Aug 2026 03:47:03 +0200 Karl Mehltretter wrote: > arm32 allmodconfig fails to build with gcc: > > drivers/mtd/chips/cfi_cmdset_0001.c:1883:1: error: the frame size of 1296 > bytes is larger than 1280 bytes [-Werror=frame-larger-than=] > > With MTD_MAP_BANK_WIDTH_32 a map_word is 32 bytes, and with KASAN_STACK > every temporary gets its own redzoned slot. do_write_buffer() builds a > map_word for each of the twelve commands it issues, and those temporaries > dominate its frame. Looks to me like pretty much all of mtd/map.h should be real functions and not #defines or inlines. The bloat from all those expansions is going to be significant. They only time they get optimised to anything small is when only CONFIG_MTD_MAP_BANK_WIDTH_1 is set. The execution time of all those that do real io (especially reads) will be dominated by the io access itself. David > > Issue the commands through a small noinline helper that takes the > command as a plain value and builds the map_word in its own frame. The > sequence of reads and writes to the chip is unchanged. > > The helper must not be inlined, and reusing one map_word local is not > enough: clang cannot assume the callee does not observe the local, so it > still returns each CMD() into a per-call-site temporary, and the frame > grows (1760 -> 1824 bytes in cfi_intelext_writev(), into which clang > inlines do_write_buffer()). > > Frame sizes with MTD_MAP_BANK_WIDTH_32 and KASAN_STACK: > > gcc 15.2 do_write_buffer() 1296 -> 520 > clang 21 cfi_intelext_writev() 1760 -> 896 > > The helper itself takes 168 (gcc) / 160 (clang) bytes. > > Reported-by: Andy Shevchenko > Closes: https://lore.kernel.org/all/20260408211407.2295175-1-andriy.shevchenko@linux.intel.com/ > Suggested-by: Miquel Raynal > Link: https://lore.kernel.org/all/87ik9cfm6g.fsf@bootlin.com/ > Assisted-by: LLM > Signed-off-by: Karl Mehltretter > --- > drivers/mtd/chips/cfi_cmdset_0001.c | 42 ++++++++++++++++++----------- > 1 file changed, 26 insertions(+), 16 deletions(-) > > diff --git a/drivers/mtd/chips/cfi_cmdset_0001.c b/drivers/mtd/chips/cfi_cmdset_0001.c > index b73596a8e021..a7c77c4e76ae 100644 > --- a/drivers/mtd/chips/cfi_cmdset_0001.c > +++ b/drivers/mtd/chips/cfi_cmdset_0001.c > @@ -1716,12 +1716,24 @@ static int cfi_intelext_write_words (struct mtd_info *mtd, loff_t to , size_t le > } > > > +/* > + * Keep noinline: inlined, the map_word temporaries put do_write_buffer() over > + * the frame-size limit with MTD_MAP_BANK_WIDTH_32 and KASAN_STACK. > + */ > +static noinline void __xipram cfi_write_cmd(struct map_info *map, > + unsigned long cmd, unsigned long adr) > +{ > + struct cfi_private *cfi = map->fldrv_priv; > + > + map_write(map, CMD(cmd), adr); > +} > + > static int __xipram do_write_buffer(struct map_info *map, struct flchip *chip, > unsigned long adr, const struct kvec **pvec, > unsigned long *pvec_seek, int len) > { > struct cfi_private *cfi = map->fldrv_priv; > - map_word status, write_cmd, datum; > + map_word status, datum; > unsigned long cmd_adr; > int ret, wbufsize, word_gap, words; > const struct kvec *vec; > @@ -1740,9 +1752,6 @@ static int __xipram do_write_buffer(struct map_info *map, struct flchip *chip, > if (is_LH28F640BF(cfi)) > cmd_adr = adr; > > - /* Let's determine this according to the interleave only once */ > - write_cmd = (cfi->cfiq->P_ID != P_ID_INTEL_PERFORMANCE) ? CMD(0xe8) : CMD(0xe9); > - > mutex_lock(&chip->mutex); > ret = get_chip(map, chip, cmd_adr, FL_WRITING); > if (ret) { > @@ -1759,7 +1768,7 @@ static int __xipram do_write_buffer(struct map_info *map, struct flchip *chip, > So we must check here and reset those bits if they're set. Otherwise > we're just pissing in the wind */ > if (chip->state != FL_STATUS) { > - map_write(map, CMD(0x70), cmd_adr); > + cfi_write_cmd(map, 0x70, cmd_adr); > chip->state = FL_STATUS; > } > status = map_read(map, cmd_adr); > @@ -1767,21 +1776,22 @@ static int __xipram do_write_buffer(struct map_info *map, struct flchip *chip, > xip_enable(map, chip, cmd_adr); > printk(KERN_WARNING "SR.4 or SR.5 bits set in buffer write (status %lx). Clearing.\n", status.x[0]); > xip_disable(map, chip, cmd_adr); > - map_write(map, CMD(0x50), cmd_adr); > - map_write(map, CMD(0x70), cmd_adr); > + cfi_write_cmd(map, 0x50, cmd_adr); > + cfi_write_cmd(map, 0x70, cmd_adr); > } > > chip->state = FL_WRITING_TO_BUFFER; > - map_write(map, write_cmd, cmd_adr); > + cfi_write_cmd(map, (cfi->cfiq->P_ID != P_ID_INTEL_PERFORMANCE) ? 0xe8 : 0xe9, > + cmd_adr); > ret = WAIT_TIMEOUT(map, chip, cmd_adr, 0, 0); > if (ret) { > /* Argh. Not ready for write to buffer */ > map_word Xstatus = map_read(map, cmd_adr); > - map_write(map, CMD(0x70), cmd_adr); > + cfi_write_cmd(map, 0x70, cmd_adr); > chip->state = FL_STATUS; > status = map_read(map, cmd_adr); > - map_write(map, CMD(0x50), cmd_adr); > - map_write(map, CMD(0x70), cmd_adr); > + cfi_write_cmd(map, 0x50, cmd_adr); > + cfi_write_cmd(map, 0x70, cmd_adr); > xip_enable(map, chip, cmd_adr); > printk(KERN_ERR "%s: Chip not ready for buffer write. Xstatus = %lx, status = %lx\n", > map->name, Xstatus.x[0], status.x[0]); > @@ -1800,7 +1810,7 @@ static int __xipram do_write_buffer(struct map_info *map, struct flchip *chip, > } > > /* Write length of data to come */ > - map_write(map, CMD(words), cmd_adr ); > + cfi_write_cmd(map, words, cmd_adr); > > /* Write data */ > vec = *pvec; > @@ -1837,7 +1847,7 @@ static int __xipram do_write_buffer(struct map_info *map, struct flchip *chip, > *pvec_seek = vec_seek; > > /* GO GO GO */ > - map_write(map, CMD(0xd0), cmd_adr); > + cfi_write_cmd(map, 0xd0, cmd_adr); > chip->state = FL_WRITING; > > ret = INVAL_CACHE_AND_WAIT(map, chip, cmd_adr, > @@ -1845,7 +1855,7 @@ static int __xipram do_write_buffer(struct map_info *map, struct flchip *chip, > chip->buffer_write_time, > chip->buffer_write_time_max); > if (ret) { > - map_write(map, CMD(0x70), cmd_adr); > + cfi_write_cmd(map, 0x70, cmd_adr); > chip->state = FL_STATUS; > xip_enable(map, chip, cmd_adr); > printk(KERN_ERR "%s: buffer write error (status timeout)\n", map->name); > @@ -1858,8 +1868,8 @@ static int __xipram do_write_buffer(struct map_info *map, struct flchip *chip, > unsigned long chipstatus = MERGESTATUS(status); > > /* reset status */ > - map_write(map, CMD(0x50), cmd_adr); > - map_write(map, CMD(0x70), cmd_adr); > + cfi_write_cmd(map, 0x50, cmd_adr); > + cfi_write_cmd(map, 0x70, cmd_adr); > xip_enable(map, chip, cmd_adr); > > if (chipstatus & 0x02) {