On Fri, 24 May 2013 13:05:07 -0400 Vivek Goyal <vgoyal at redhat.com> wrote: > On Fri, May 24, 2013 at 06:46:53PM +0200, Michael Holzheu wrote: > > On Fri, 24 May 2013 11:28:49 -0400 > > Vivek Goyal <vgoyal at redhat.com> wrote: > > > > > On Fri, May 24, 2013 at 05:06:26PM +0200, Michael Holzheu wrote: > > > > [snip] > > > > > As /proc/vmcore is the most used and useful interface, I prefer > > > that we swap memory and put that info in elf headers. > > > For /dev/oldme, I don't mind if we leave it as it is. If somebody > > > really cares, then I guess we need to write a new command line > > > option which /dev/mem can parse and which tells it about swaps so > > > that /dev/oldmem can map things correctly. (This is better than > > > hardcoding things). > > > > Besides of the potential /dev/oldmem issue, I still do not > > understand the option of doing the swap in the elf header. Looks > > like I missed here a fundamental design point of kdump :( > > > > Is that done by specifying different virtual and physical addresses > > in the ELF header? > > Nope. We keep the virtual to physical address mapping same. We just > modify the p_offset in PT_LOAD elf header to represent where actually > the memory is present physically. And when /proc/vmcore reads the > data, it reads it from p_offset. > > IOW, p_offset and p_paddr will be different for swapped memory but > should be same for memory which has not been swapped. Hello Vivek, Ok, now I got it :) It worked for me by specifying a PT_LOAD with: phdr->p_offset = OLDMEM_BASE; phdr->p_vaddr = phdr->p_paddr = 0; phdr->p_filesz = phdr->p_memsz = OLDMEM_SIZE; Best Regards, Michael
