Re: [PATCH v5 2/9] watchdog/hpwdt: Remove legacy NMI sourcing.

From: Guenter Roeck
Date: Mon Feb 26 2018 - 20:30:05 EST


On 02/26/2018 05:02 PM, Jerry Hoemann wrote:
On Mon, Feb 26, 2018 at 06:32:30AM -0800, Guenter Roeck wrote:
On 02/26/2018 06:11 AM, Arnd Bergmann wrote:
On Mon, Feb 26, 2018 at 4:22 AM, Jerry Hoemann <jerry.hoemann@xxxxxxx> wrote:
Gen8 and prior Proliant systems supported the "CRU" interface
to firmware. This interfaces allows linux to "call back" into firmware
to source the cause of an NMI. This feature isn't fully utilized
as the actual source of the NMI isn't printed, the driver only
indicates that the source couldn't be determined when the call
fails.

With the advent of Gen9, iCRU replaces the CRU. The call back
feature is no longer available in firmware. To be compatible and
not attempt to call back into firmware on system not supporting CRU,
the SMBIOS table is consulted to determine if it is safe to
make the call back or not.

This results in about half of the driver code being devoted
to either making CRU calls or determing if it is safe to make
CRU calls. As noted, the driver isn't really using the results of
the CRU calls.

Furthermore, as a consequence of the Spectre security issue, the
BIOS/EFI calls are being wrapped into Spectre-disabling section.
Removing the call back in hpwdt_pretimeout assists in this effort.

As the CRU sourcing of the NMI isn't required for handling the
NMI and there are security concerns with making the call back, remove
the legacy (pre Gen9) NMI sourcing and the DMI code to determine if
the system had the CRU interface.

Signed-off-by: Jerry Hoemann <jerry.hoemann@xxxxxxx>

This avoids a warning in mainline kernels, so that's great:

drivers/watchdog/hpwdt.o: warning: objtool: .text+0x24: indirect call
found in RETPOLINE build

I wonder what we do about stable kernels. Are both this patch and the patch
that added the objtool warning message candidates for backports to
stable kernels?


Makes sense to me, but it is really a bit more than a bug fix, so I'll
leave it up to Jerry/HPE to make the call in respect to hpwdt.


Generally speaking, HPE customers who run linux do so through a distro
vendor and pick up patches from them. But I'm sure there are some
customers who do things differently.

The distro vendor's have their own repos and we'll work with them
to back port patches to their code base. So, I typically don't do a lot
of kernel.org stable branch work.

Looks like objtool has been enhanced to find Spectre vulnerable code.
Are the other kernel patches related to Spectre being back ported
to stable release lines? If yes, it probably make sense to do
the hpwdt change as well.


Spectre has been backported to v4.4 and later. I don't know about earlier kernels.

Is just the patch removing the firmware call back wanted/needed? Or the
whole driver rewrite? (The older baseline don't have all the watchdog
features that the patch set uses.)


We would only want to backport this patch. The rest is really out of scope.

Which stable baseline(s) would need to be patched? Priority?

Who does it? (i.e. do you want me to submit patches to the stable baseline?)

We would tag the patch for stable (and submit it into v4.16-rc). Greg would
take care of the rest unless there are conflicts, in which case we get a note
telling us that a backport is needed.

Guenter