----- "Paul-Kenji Cahier Furuya" <pkc@xxxxxxxxxxxx> wrote: > I tried with the system map, but it did not help, still got the same > error. I don't know why you would have gotten "the same error" when you used the System.map file. Originally you saw: <readmem: c148a9a0, KVADDR, "kernel_config_data", 32768, (ROE), 8bad3d8> crash: read error: kernel virtual address: c148a9a0 type: "kernel_config_data" WARNING: cannot read kernel_config_data <readmem: c1487e28, KVADDR, "cpu_possible_mask", 4, (FOE), bfed4bbc> crash: read error: kernel virtual address: c1487e28 type: "cpu_possible_mask" And that's because the vmlinux file that you were using had compiled-in kernel virtual addresses that started with "c1xxxxxx". If the System.map file was put on the command line, it would have overwritten those kernel addresses with the ones found in the System.map, i.e., those that begin with "c0xxxxxx". Did the System.map file that you used contain symbols that started with "c1xxxxxx" or "c0xxxxxx"? And if they did start with "c0xxxxxx", can you attach the "crash -d8 ..." output? Dave -- Crash-utility mailing list Crash-utility@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/crash-utility
