On Tue, Aug 26, 2008 at 05:13:27PM -0500, Cliff Wickman wrote: > > Hi Vivek, > > I'm having trouble getting a system kernel to load a kdump kernel. > > These are 2.6.26.2 kernels on an x86_64. > The kdump kernel has no modules. > The kdump kernel area is reserved as 64M at 16M. > > This is the way I try to load it: > > /usr/local/sbin/kexec -p /boot/vmlinux-cpwcomm --args-linux --append="root=/dev/sda3 1 irqpoll maxcpus=1 reset_devices" > > This works with -l. But with -p the kernel objects to the 2 lowest > segments, at 0x1000 and 0xa000 which I presume are the startup code > and arguments. > CCing to kexec mailing list. Few questions. - What version of kexec-tools are you using? - Can you please provide output of /proc/iomem. (With crash kernel memory reserved). - Looks like you have compiled your kernel for physical address 16MB? (CONFIG_PHYSICAL_START=0x1000000) - Can you also try loading relocatable bzImage, instead of vmlinux and see what happens. > I have this debugging output from my kexec: > > cpw: elf_x86_64_load returning entry:0x1550 > cpw: after call to file_type[i].load: nr_segments:6 entry:0x1550 > kexec_load: entry = 0x1550 flags = 1 > nr_segments = 6 > segment[0].buf = 0x5237a0 > segment[0].bufsz = 7100 > segment[0].mem = 0x1000 > segment[0].memsz = 9000 > > segment[1].buf = 0x52aaf0 > segment[1].bufsz = 1000 > segment[1].mem = 0xa000 > segment[1].memsz = 1000 > I think above two segments are not being loaded at right place. Looks like kexec-tools decided to load first one at physical address 0x1000 and other at physical address 0xa000. For crash kernel this is not right. It should come out of reserved memory area and that's why you are encountering the error. For crash kernels, kexec-tools sets the value of mem-min and mem-max in such a manner that it is with-in reserved memory area so that any memory allocation in locate_hole() is done with-in reserved area and not outside it. In your case looks like something went wrong. Either memory area was not reserved properly or value of mem-min, mem-max was not set properly etc. Can you please paste the output of /proc/iomem and also do some kexec-tools debugging. Especially value of mem-max and mem-min inside locate_hole() function. Thanks Vivek > segment[2].buf = 0x7f9bd8461010 > segment[2].bufsz = 224668 > segment[2].mem = 0x1000000 > segment[2].memsz = 225000 > > segment[3].buf = 0x7f9bd8686010 > segment[3].bufsz = 75508 > segment[3].mem = 0x1225000 > segment[3].memsz = 76000 > > segment[4].buf = 0x7f9bd8761010 > segment[4].bufsz = c08 > segment[4].mem = 0x129b000 > segment[4].memsz = 1000 > > segment[5].buf = 0x7f9bd87fd010 > segment[5].bufsz = 30004 > segment[5].mem = 0x129c000 > segment[5].memsz = c2000 > kexec_load failed: Cannot assign requested address > > > And I have some debugging in the system kernel: > > cleopatra1:/tmp/cpw # dmesg | tail > > cpw: -p, call kimage_crash_alloc > cpw: kimage_crash_alloc bad entry point 0x1550:0x1000000 0x1550:0x4ffffff > cpw: kimage_crash_alloc segment 0: 0x1000-0x9fff > cpw: kimage_crash_alloc pta EADDRNOTAVAIL 0x1000:0x1000000 0x9fff:0x4ffffff > cpw: kimage_crash_alloc segment 1: 0xa000-0xafff > cpw: kimage_crash_alloc pta EADDRNOTAVAIL 0xa000:0x1000000 0xafff:0x4ffffff > cpw: kimage_crash_alloc segment 2: 0x1000000-0x1224fff > cpw: kimage_crash_alloc segment 3: 0x1225000-0x129afff > cpw: kimage_crash_alloc segment 4: 0x129b000-0x129bfff > cpw: kimage_crash_alloc segment 5: 0x129c000-0x135dfff > > The code that insists that the entry point must be within 64M at 16M, and > that every segment must also be in that range, seems to be old code. > > I don't understand. > Is kexec generating bad segments? Or am I mis-using something? > Can you tell from the above? > > Thanks. > -Cliff
