Re: [PATCH v4 14/25] powernv/fadump: process the crashdump by exporting it as /proc/vmcore

On 14/08/19 3:48 PM, Mahesh J Salgaonkar wrote:

On 2019-07-16 17:03:38 Tue, Hari Bathini wrote:

quoted

Add support in the kernel to process the crash'ed kernel's memory
preserved during MPIPL and export it as /proc/vmcore file for the
userland scripts to filter and analyze it later.

Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
---
 arch/powerpc/platforms/powernv/opal-fadump.c |  190 ++++++++++++++++++++++++++
 1 file changed, 187 insertions(+), 3 deletions(-)

[...]

quoted

+		ret = opal_mpipl_query_tag(OPAL_MPIPL_TAG_KERNEL, &addr);
+		if ((ret != OPAL_SUCCESS) || !addr) {
+			pr_err("Failed to get Kernel metadata (%lld)\n", ret);
+			return 1;
+		}
+
+		addr = be64_to_cpu(addr);
+		pr_debug("Kernel metadata addr: %llx\n", addr);
+
+		opal_fdm_active = __va(addr);
+		r_opal_fdm_active = (void *)addr;
+		if (r_opal_fdm_active->version != OPAL_FADUMP_VERSION) {
+			pr_err("FADump active but version (%u) unsupported!\n",
+			       r_opal_fdm_active->version);
+			return 1;
+		}
+
+		/* Kernel regions not registered with f/w  for MPIPL */
+		if (r_opal_fdm_active->registered_regions == 0) {
+			opal_fdm_active = NULL;

What about partial dump capture scenario ? What if opal crashes while
kernel was in middle of registering ranges ? We may have partial dump
captured which won't be useful.
e,g. If we have total of 4 ranges to be registered and opal crashes
after successful registration of only 2 ranges with 2 pending, we will get a
partial dump which needs to be ignored.

I think check shuold be comparing registered_regions against total number of
regions. What do you think ?

Yes, Mahesh.
Taking care of that in 22/25

Thanks
Hari