Hi Lei, Thank you for the patch. The almost same patch has been merged into the devel git-tree, does the following patch fix the prblem you said ? http://makedumpfile.git.sourceforge.net/git/gitweb.cgi?p=makedumpfile/makedumpfile;a=commit;h=63f720a7784debce82c8942970d5c0d8412f209f Thanks Ken'ichi Ohmichi On Sun, 4 Sep 2011 09:02:39 -0700 Lei Wen <leiwen at marvell.com> wrote: > > If the area reserved for the kdump is at the end of system memory, then > it would occur such case that the second sector is turly 0 size. And as > the beginnig of the second sector is already out of system memory range. > Try to access this range, would cause makedumpfile failed. > > Signed-off-by: Lei Wen <leiwen at marvell.com> > --- > makedumpfile.c | 4 ++++ > 1 files changed, 4 insertions(+), 0 deletions(-) > > diff --git a/makedumpfile.c b/makedumpfile.c > index 7d0c663..3f8d9af 100644 > --- a/makedumpfile.c > +++ b/makedumpfile.c > @@ -1336,6 +1336,8 @@ get_elf_info(void) > } > if (phdr.p_type != PT_LOAD) > continue; > + if (phdr.p_filesz == 0) > + continue; > > if (j == 0) { > info->offset_load_memory = phdr.p_offset; > @@ -5756,6 +5758,8 @@ write_elf_pages(struct cache_data *cd_header, struct cache_data *cd_page) > > if (load.p_type != PT_LOAD) > continue; > + if (load.p_filesz == 0) > + continue; > > off_memory= load.p_offset; > paddr = load.p_paddr; > -- > 1.7.0.4 > > > _______________________________________________ > kexec mailing list > kexec at lists.infradead.org > http://lists.infradead.org/mailman/listinfo/kexec
