[V3] powerpc/powernv: Add a kmsg_dumper that flushes console output on panic

[Michael Ellerman]

On Tue, 2016-01-12 at 15:17 +1100, Russell Currey wrote:
> On Tue, 2016-01-12 at 14:44 +1100, Stewart Smith wrote:
> > Michael Ellerman <mpe at ellerman.id.au> writes:
> > > On Fri, 2015-27-11 at 06:23:07 UTC, Russell Currey wrote:
> > > > On BMC machines, console output is controlled by the OPAL firmware and is
> > > > only flushed when its pollers are called.  When the kernel is in a panic
> > > > state, it no longer calls these pollers and thus console output does not
> > > > completely flush, causing some output from the panic to be lost.
> > > > 
> > > > Output is only actually lost when the kernel is configured to not power
> > > > off
> > > > or reboot after panic (i.e. CONFIG_PANIC_TIMEOUT is set to 0) since OPAL
> > > > flushes the console buffer as part of its power down routines.  Before
> > > > this
> > > > patch, however, only partial output would be printed during the timeout
> > > > wait.
> > > > 
> > > > This patch adds a new kmsg_dumper which gets called at panic time to
> > > > ensure
> > > > panic output is not lost.  It accomplishes this by calling
> > > > OPAL_CONSOLE_FLUSH
> > > > in the OPAL API, and if that is not available, the pollers are called
> > > > enough
> > > > times to (hopefully) completely flush the buffer.
> > > > 
> > > > The flushing mechanism will only affect output printed at and before the
> > > > kmsg_dump call in kernel/panic.c:panic().  As such, the "end Kernel
> > > > panic"
> > > > message may still be truncated as follows:
> > > > 

> > > > > Call Trace:
> > > > > [c000000f1f603b00] [c0000000008e9458] dump_stack+0x90/0xbc (unreliable)
> > > > > [c000000f1f603b30] [c0000000008e7e78] panic+0xf8/0x2c4
> > > > > [c000000f1f603bc0] [c000000000be4860] mount_block_root+0x288/0x33c
> > > > > [c000000f1f603c80] [c000000000be4d14] prepare_namespace+0x1f4/0x254
> > > > > [c000000f1f603d00] [c000000000be43e8] kernel_init_freeable+0x318/0x350
> > > > > [c000000f1f603dc0] [c00000000000bd74] kernel_init+0x24/0x130
> > > > > [c000000f1f603e30] [c0000000000095b0] ret_from_kernel_thread+0x5c/0xac
> > > > > ---[ end Kernel panic - not
> > > > 
> > > > This functionality is implemented as a kmsg_dumper as it seems to be the
> > > > most sensible way to introduce platform-specific functionality to the
> > > > panic function.
> > > > 
> > > > Signed-off-by: Russell Currey <ruscur at russell.cc>
> > > > Reviewed-by: Andrew Donnellan <andrew.donnellan at au1.ibm.com>
> > > 
> > > Applied to powerpc next, thanks.
> > > 
> > > https://git.kernel.org/powerpc/c/affddff69c55eb68969448f35f
> > 
> > The firmware interface changed slightly since this kernel patch[1], it
> > added a parameter to OPAL_CONSOLE_FLUSH which accepted the terminal
> > number to flush, theoretically allowing this to be plumbed into TTY
> > layer or something too.
> > 
> > So, we'll either have to update this patch or replace it with an updated
> > one.
> > 
> > [1] i'm pushing the accepted skiboot patch now.
> > 
> I'm working on an updated kernel patch to use the new parameter and additional
> return values, so I suppose it's up to mpe whether or not this patch gets
> merged now and another gets sent later to amend it, or if this patch gets
> reverted in next and I can send a V4 adding the new stuff.

Doh. I'd rather not revert it, unless we have to.

Basically we're passing junk in r3, which skiboot is expecting to be the
terminal number.

So running the current kernel code on the updated skiboot shouldn't crash and
burn, it just won't actually work the way it's supposed to.

So my preference would be just an incremental patch ASAP to fix the kernel to
do the right thing with the new interface.

cheers