kdb: Remove stack dump when entering kgdb due to NMI
authorDaniel Thompson <daniel.thompson@linaro.org>
Thu, 6 Nov 2014 15:32:13 +0000 (15:32 +0000)
committerJason Wessel <jason.wessel@windriver.com>
Thu, 19 Feb 2015 18:39:02 +0000 (12:39 -0600)
Issuing a stack dump feels ergonomically wrong when entering due to NMI.

Entering due to NMI is normally a reaction to a user request, either the
NMI button on a server or a "magic knock" on a UART. Therefore the
backtrace behaviour on entry due to NMI should be like SysRq-g (no stack
dump) rather than like oops.

Note also that the stack dump does not offer any information that
cannot be trivial retrieved using the 'bt' command.

Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
kernel/debug/kdb/kdb_main.c

index 60f6bb8..0a22f45 100644 (file)
@@ -1247,7 +1247,6 @@ static int kdb_local(kdb_reason_t reason, int error, struct pt_regs *regs,
                kdb_printf("due to NonMaskable Interrupt @ "
                           kdb_machreg_fmt "\n",
                           instruction_pointer(regs));
-               kdb_dumpregs(regs);
                break;
        case KDB_REASON_SSTEP:
        case KDB_REASON_BREAK: