This is _not_ an upstream commit and just for 5.4.y only.

kernel test robot reported a 5.4.y build issue found by randconfig [1] after backporting commit 89b158635ad7 ("lib/lz4: explicitly support in-place decompression"") due to "undefined reference to `memmove'".

However, upstream and 5.10 LTS seem fine. After digging further, I found commit a510b616131f ("MIPS: Add support for ZSTD-compressed kernels") introduced memmove() occasionally and it has been included since v5.10.

This partially cherry-picks the memmove() part of commit a510b616131f to fix the reported build regression since we don't need the whole patch for 5.4 LTS at all.

[1] https://lore.kernel.org/r/202107070120.6dOj1kB7-lkp@intel.com/ Fixes: defcc2b5e54a ("lib/lz4: explicitly support in-place decompression") # 5.4.y Reported-by: kernel test robot lkp@intel.com Signed-off-by: Gao Xiang hsiangkao@linux.alibaba.com --- v2: "just submit the fix and say _why_ this is not an upstream commit, do not attempt to emulate an upstream commit like your change did." as Greg suggested...

arch/mips/boot/compressed/string.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+)

diff --git a/arch/mips/boot/compressed/string.c b/arch/mips/boot/compressed/string.c index 43beecc3587c..0b593b709228 100644 --- a/arch/mips/boot/compressed/string.c +++ b/arch/mips/boot/compressed/string.c @@ -5,6 +5,7 @@ * Very small subset of simple string routines */

+#include <linux/compiler_attributes.h> #include <linux/types.h>

void *memcpy(void *dest, const void *src, size_t n) @@ -27,3 +28,19 @@ void *memset(void *s, int c, size_t n) ss[i] = c; return s; } + +void * __weak memmove(void *dest, const void *src, size_t n) +{ + unsigned int i; + const char *s = src; + char *d = dest; + + if ((uintptr_t)dest < (uintptr_t)src) { + for (i = 0; i < n; i++) + d[i] = s[i]; + } else { + for (i = n; i > 0; i--) + d[i - 1] = s[i - 1]; + } + return dest; +}