Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Nick Terrell <nickrterrell@gmail.com>
To:     Borislav Petkov <bp@alien8.de>,
        Thomas Gleixner <tglx@linutronix.de>
Cc:     linux-kernel@vger.kernel.org, Chris Mason <clm@fb.com>,
        linux-kbuild@vger.kernel.org, x86@kernel.org,
        gregkh@linuxfoundation.org, Petr Malat <oss@malat.biz>,
        Kees Cook <keescook@chromium.org>,
        Kernel Team <Kernel-team@fb.com>,
        Nick Terrell <nickrterrell@gmail.com>,
        Adam Borowski <kilobyte@angband.pl>,
        Patrick Williams <patrickw3@fb.com>, rmikey@fb.com,
        mingo@kernel.org, Patrick Williams <patrick@stwcx.xyz>,
        Sedat Dilek <sedat.dilek@gmail.com>,
        Norbert Lange <nolange79@gmail.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Alex Xu <alex_y_xu@yahoo.ca>,
        Masahiro Yamada <masahiroy@kernel.org>,
        Nick Terrell <terrelln@fb.com>
Subject: [PATCH v10 5/8] x86: bump ZO_z_extra_bytes margin for zstd
Date:   Thu, 30 Jul 2020 12:08:38 -0700
Message-Id: <20200730190841.2071656-6-nickrterrell@gmail.com>
In-Reply-To: <20200730190841.2071656-1-nickrterrell@gmail.com>
References: <20200730190841.2071656-1-nickrterrell@gmail.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

From: Nick Terrell <terrelln@fb.com>

Bump the ZO_z_extra_bytes margin for zstd.

Zstd needs 3 bytes per 128 KB, and has a 22 byte fixed overhead.
Zstd needs to maintain 128 KB of space at all times, since that is
the maximum block size. See the comments regarding in-place
decompression added in lib/decompress_unzstd.c for details.

The existing code is written so that all the compression algorithms use
the same ZO_z_extra_bytes. It is taken to be the maximum of the growth
rate plus the maximum fixed overhead. The comments just above this diff
state that:

Reviewed-by: Kees Cook <keescook@chromium.org>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Signed-off-by: Nick Terrell <terrelln@fb.com>
---
 arch/x86/boot/header.S | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/arch/x86/boot/header.S b/arch/x86/boot/header.S
index 735ad7f21ab0..6dbd7e9f74c9 100644
--- a/arch/x86/boot/header.S
+++ b/arch/x86/boot/header.S
@@ -539,8 +539,14 @@ pref_address:		.quad LOAD_PHYSICAL_ADDR	# preferred load addr
 # the size-dependent part now grows so fast.
 #
 # extra_bytes = (uncompressed_size >> 8) + 65536
+#
+# ZSTD compressed data grows by at most 3 bytes per 128K, and only has a 22
+# byte fixed overhead but has a maximum block size of 128K, so it needs a
+# larger margin.
+#
+# extra_bytes = (uncompressed_size >> 8) + 131072
 
-#define ZO_z_extra_bytes	((ZO_z_output_len >> 8) + 65536)
+#define ZO_z_extra_bytes	((ZO_z_output_len >> 8) + 131072)
 #if ZO_z_output_len > ZO_z_input_len
 # define ZO_z_extract_offset	(ZO_z_output_len + ZO_z_extra_bytes - \
 				 ZO_z_input_len)
-- 
2.27.0