Hello Suzuki,

On 4/6/2016 9:53 AM, Suzuki K Poulose wrote:

On 06/04/16 16:12, Tyler Baicar wrote:

quoted

+    hest_hdr = (struct acpi_hest_header *)generic;
+    if (hest_hdr->type == ACPI_HEST_TYPE_GENERIC_ERROR_V2) {
+        ghes->generic_v2 = (struct acpi_hest_generic_v2 *)generic;
+        rc = apei_map_generic_address(
+            &ghes->generic_v2->read_ack_register);
+        if (rc)
+            goto err_unmap;
+    } else
+        ghes->generic_v2 = NULL;

...

quoted

  err_unmap:
apei_unmap_generic_address(&generic->error_status_address);
+    if (ghes->generic_v2)
+        apei_unmap_generic_address(
+            &ghes->generic_v2->read_ack_register);
  err_free:
      kfree(ghes);
      return ERR_PTR(rc);

@@ -279,6 +295,9 @@ static void ghes_fini(struct ghes *ghes)
  {
      kfree(ghes->estatus);

apei_unmap_generic_address(&ghes->generic->error_status_address);
+    if (ghes->generic_v2)
+        apei_unmap_generic_address(
+            &ghes->generic_v2->error_status_address);

I am not familiar with the APEI code, but is this error_status_address or
read_ack_register ? We don't seem to be mapping error_status_address 
in generic_v2 header
which is introduced in this patch ? Am I missing something ?

Suzuki

Thank you for your feedback. This does look like an error; it should be 
&ghes->generic_v2->read_ack_register. The variable 
&ghes->generic_v2->error_status_address is the same as 
&ghes->generic->error_status_address which is unmapped on the line above 
the if statement here.

Thanks,
Tyler

-- 
Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project