----- Original Message ----- Good catch. Apparently the ARM users and maintainers on this list don't ever run "live" as you are doing. And since I don't have any ARM hardware, I can only test it on supplied dumpfiles with an x86 binary built with "make target=ARM". The problem is this patch that went into crash-7.0.9: Improve the method for determining whether a 32-bit ARM vmlinux is an LPAE enabled kernel by first checking whether CONFIG_ARM_LPAE exists in the vmcoreinfo data, and if it does not, by then checking whether the next higher symbol above "swapper_pg_dir" is 0x5000 bytes higher in value. (sdu.liu(a)huawei.com) diff --git a/arm.c b/arm.c index cb7d841..e7d3dbc 100644 --- a/arm.c +++ b/arm.c @@ -190,6 +190,8 @@ void arm_init(int when) { ulong vaddr; + char *string; + struct syment *sp; #if defined(__i386__) || defined(__x86_64__) if (ACTIVE()) @@ -229,8 +231,13 @@ arm_init(int when) * LPAE requires an additional page for the PGD, * so PG_DIR_SIZE = 0x5000 for LPAE */ - if ((symbol_value("_text") - symbol_value("swapper_pg_dir")) == 0x5000) + if ((string = pc->read_vmcoreinfo("CONFIG_ARM_LPAE"))) { machdep->flags |= PAE; + free(string); + } else if ((sp = next_symbol("swapper_pg_dir", NULL)) && + (sp->value - symbol_value("swapper_pg_dir")) == 0x5000) + machdep->flags |= PAE; + machdep->kvbase = symbol_value("_stext") & ~KVBASE_MASK; machdep->identity_map_base = machdep->kvbase; machdep->is_kvaddr = arm_is_kvaddr; The pc->read_vmcoreinfo method is only initialized for ELF kdumps and compressed kdumps. So either a dummy function should be put in there that returns a NULL or arm_init() should check for its existence. I appreciate the bug report -- I'll post something today for crash-7.1.0. Thanks, Dave