[PATCH] powerpc/powernv: Fallback to old HMI handling behavior for old firmware

From: Mahesh Salgaonkar <redacted>

Recently we moved HMI handling into Linux kernel instead of taking
HMI directly in OPAL. This new change is dependent on new OPAL call
for HMI recovery which was introduced in newer firmware. While this new
change works fine with latest OPAL firmware, we broke the HMI handling
if we run newer kernel on old OPAL firmware that results in system hang.

This patch fixes this issue by falling back to old HMI behavior on older
OPAL firmware.

This patch introduces a check for opal token OPAL_HANDLE_HMI to see
if we are running on newer firmware or old firmware. On newer firmware
this check would return OPAL_TOKEN_PRESENT, otherwise we are running on
old firmware and fallback to old HMI behavior.

This patch depends on opal check token patch posted at ppc-devel
https://lists.ozlabs.org/pipermail/linuxppc-dev/2014-August/120224.html

Signed-off-by: Mahesh Salgaonkar <redacted>
---
 arch/powerpc/platforms/powernv/opal.c |   18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/arch/powerpc/platforms/powernv/opal.c b/arch/powerpc/platforms/powernv/opal.c
index b44eec3..2768cd3 100644
--- a/arch/powerpc/platforms/powernv/opal.c
+++ b/arch/powerpc/platforms/powernv/opal.c

@@ -194,6 +194,24 @@ static int __init opal_register_exception_handlers(void)
 	 * fwnmi area at 0x7000 to provide the glue space to OPAL
 	 */
 	glue = 0x7000;
+
+	/* Check if we are running on newer firmware that exports
+	 * OPAL_HANDLE_HMI token. If yes, then don't ask opal to patch
+	 * HMI interrupt and we catch it directly in Linux kernel.
+	 *
+	 * For older firmware we will fallback to old behavior and
+	 * let OPAL patch the HMI vector and handle it inside OPAL
+	 * firmware.
+	 */
+	if (opal_check_token(OPAL_HANDLE_HMI) != OPAL_TOKEN_PRESENT) {
+		/* We are on old firmware. fallback to old behavior. */
+		pr_info("%s: Falling back to old HMI handling behavior.\n",
+			__func__);
+		opal_register_exception_handler(
+				OPAL_HYPERVISOR_MAINTENANCE_HANDLER,
+				0, glue);
+		glue += 128;
+	}
 	opal_register_exception_handler(OPAL_SOFTPATCH_HANDLER, 0, glue);
 #endif
 

Mahesh J Salgaonkar [off-list ref] writes:

From: Mahesh Salgaonkar <redacted>

Recently we moved HMI handling into Linux kernel instead of taking
HMI directly in OPAL. This new change is dependent on new OPAL call
for HMI recovery which was introduced in newer firmware. While this new
change works fine with latest OPAL firmware, we broke the HMI handling
if we run newer kernel on old OPAL firmware that results in system hang.

This patch fixes this issue by falling back to old HMI behavior on older
OPAL firmware.

This patch introduces a check for opal token OPAL_HANDLE_HMI to see
if we are running on newer firmware or old firmware. On newer firmware
this check would return OPAL_TOKEN_PRESENT, otherwise we are running on
old firmware and fallback to old HMI behavior.

This patch depends on opal check token patch posted at ppc-devel
https://lists.ozlabs.org/pipermail/linuxppc-dev/2014-August/120224.html

(just reviewed that patch before this one...)

So.. that all looks fine, but another note: why are we doing the
OPAL_SOFTPATCH_HANDLER and why is it all #if 0 out in skiboot?

for this patch:

Reviewed-by: Stewart Smith <redacted>

On Mon, 2014-06-10 at 09:34:19 UTC, Mahesh Salgaonkar wrote:

From: Mahesh Salgaonkar <redacted>

Hi Mahesh,

Recently we moved HMI handling into Linux kernel instead of taking
HMI directly in OPAL. This new change is dependent on new OPAL call
for HMI recovery which was introduced in newer firmware. While this new
change works fine with latest OPAL firmware, we broke the HMI handling
if we run newer kernel on old OPAL firmware that results in system hang.

This patch fixes this issue by falling back to old HMI behavior on older
OPAL firmware.

It sounds like "older" firmware is actually "the currently released firmware".
The "newer" firmware is still in development, is that right?

If so please update the comment and changelog to better reflect that.

Please format your long comments like:

/*
 * Check if we are ..
 */

I know some of our code uses the other style but this is the commonly accepted
style.

OPAL_TOKEN_PRESENT was dropped from the API. Just use:

	if (!opal_check_token(OPAL_HANDLE_HMI)) {

Please just use "opal: " rather than __func__.

And rather than the user having to know what the old vs new behaviour is, can
you make it explicit in the message, eg:

"opal: Old firmware detected, letting OPAL handle HMIs."

Newline here please.

 	opal_register_exception_handler(OPAL_SOFTPATCH_HANDLER, 0, glue);
 #endif

cheers