On 9 February 2015 at 13:58, Simon Horman <horms at verge.net.au> wrote:
> These changes do not seem to be related to each other.
> Please split this into 4 separate patches each with
> a changelog that briefly describes what the problem is.
>
Will split them each into a separate patch.
> On Sun, Feb 01, 2015 at 12:10:08AM +0800, Yousong Zhou wrote:
>>
>> Signed-off-by: Yousong Zhou <yszhou4tech at gmail.com>
>> ---
>> kexec/arch/mips/crashdump-mips.c | 3 ++-
>> kexec/arch/mips/kexec-elf-mips.c | 2 +-
>> kexec/kexec-elf-rel.c | 3 +++
>> kexec/kexec.h | 2 +-
>> 4 files changed, 7 insertions(+), 3 deletions(-)
>>
>> diff --git a/kexec/arch/mips/crashdump-mips.c b/kexec/arch/mips/crashdump-mips.c
>> index e7840e0..98c9f7c 100644
>> --- a/kexec/arch/mips/crashdump-mips.c
>> +++ b/kexec/arch/mips/crashdump-mips.c
>> @@ -22,6 +22,7 @@
>> #include <stdlib.h>
>> #include <errno.h>
>> #include <limits.h>
>> +#include <inttypes.h>
>> #include <elf.h>
>> #include <sys/types.h>
>> #include <sys/stat.h>
>> @@ -52,7 +53,7 @@ static int get_kernel_paddr(struct crash_elf_info *elf_info)
>>
>> if (parse_iomem_single("Kernel code\n", &start, NULL) == 0) {
>> elf_info->kern_paddr_start = start;
>> - dbgprintf("kernel load physical addr start = 0x%lx\n", start);
>> + dbgprintf("kernel load physical addr start = 0x%" PRIu64 "\n", start);
>> return 0;
>> }
>>
>> diff --git a/kexec/arch/mips/kexec-elf-mips.c b/kexec/arch/mips/kexec-elf-mips.c
>> index a27d986..8a6419a 100644
>> --- a/kexec/arch/mips/kexec-elf-mips.c
>> +++ b/kexec/arch/mips/kexec-elf-mips.c
>> @@ -158,7 +158,7 @@ int elf_mips_load(int argc, char **argv, const char *buf, off_t len,
>> if (info->kexec_flags & KEXEC_ON_CRASH)
>> /* In case of crashdump segment[0] is kernel.
>> * Put cmdline just after it. */
>> - cmdline_addr = info->segment[0].mem +
>> + cmdline_addr = (unsigned long)info->segment[0].mem +
>> info->segment[0].memsz;
>
> I wonder if we can resolve this without resorting to a cast.
It's a "pointer to integer without cast" warning. I think keeping
(unsigned long) for a pointer is fine, isnt it?
>
>> else
>> cmdline_addr = 0;
>> diff --git a/kexec/kexec-elf-rel.c b/kexec/kexec-elf-rel.c
>> index c625f30..5a5fcd1 100644
>> --- a/kexec/kexec-elf-rel.c
>> +++ b/kexec/kexec-elf-rel.c
>> @@ -347,6 +347,9 @@ int elf_rel_load(struct mem_ehdr *ehdr, struct kexec_info *info,
>> else if (shdr->sh_type == SHT_RELA) {
>> rel = elf_rela(ehdr, ptr);
>> }
>> + else {
>> + die("Unexpected sh_type: %d\n", shdr->sh_type);
>> + }
>> /* the location to change */
>> location = section->sh_data + rel.r_offset;
>>
This one has already been fixed by explicitly initialising the local
variable. So I will drop it in the next version.
Thank you for your time on this.
yousong
>> diff --git a/kexec/kexec.h b/kexec/kexec.h
>> index 4be2b2f..2c85052 100644
>> --- a/kexec/kexec.h
>> +++ b/kexec/kexec.h
>> @@ -245,7 +245,7 @@ extern int file_types;
>>
>> extern void dbgprint_mem_range(const char *prefix, struct memory_range *mr, int nr_mr);
>> extern void die(const char *fmt, ...)
>> - __attribute__ ((format (printf, 1, 2)));
>> + __attribute__ ((format (printf, 1, 2), noreturn));
>> extern void *xmalloc(size_t size);
>> extern void *xrealloc(void *ptr, size_t size);
>> extern char *slurp_file(const char *filename, off_t *r_size);
>> --
>> 1.7.10.4
>>
