x86/boot: Provide more slack space during decompression

The current slack space is not enough for LZ4, which has a worst case
overhead of 0.4% for data that cannot be further compressed. With
an LZ4 compressed kernel with an embedded initrd, the output is likely
to overwrite the input.

Increase the slack space to avoid that.

Signed-off-by: Jan H. Schönherr <jschoenh@amazon.de>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1503842124-29718-1-git-send-email-jschoenh@amazon.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>

diff --git a/arch/x86/boot/header.S b/arch/x86/boot/header.S
index 2ed8f0c25def4b01dd50107425c5b6e54d60f2eb..1bb08ecffd24a61e1d9bccf2094e70996f280bce 100644 (file)
--- a/arch/x86/boot/header.S
+++ b/arch/x86/boot/header.S
@@ -520,8 +520,14 @@ pref_address:              .quad LOAD_PHYSICAL_ADDR        # preferred load addr
 # the description in lib/decompressor_xxx.c for specific information.
 #
 # extra_bytes = (uncompressed_size >> 12) + 65536 + 128
+#
+# LZ4 is even worse: data that cannot be further compressed grows by 0.4%,
+# or one byte per 256 bytes. OTOH, we can safely get rid of the +128 as
+# the size-dependent part now grows so fast.
+#
+# extra_bytes = (uncompressed_size >> 8) + 65536
 
-#define ZO_z_extra_bytes       ((ZO_z_output_len >> 12) + 65536 + 128)
+#define ZO_z_extra_bytes       ((ZO_z_output_len >> 8) + 65536)
 #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)