Re: [PATCH 2/5] conf: Introduce pstore device
On Mon, Jul 22, 2024 at 02:23:30PM GMT, Michal Privoznik wrote:
+Pstore
+~~~~~~~~~
+
+Pstore is an oops/panic logger that writes its logs to a block device and
+non-block device before the system crashes. Currently only ACPI Error Record
+Serialization Table, ERST, is supported. This feature is designed for storing
+error records in persistent storage for future reference and/or debugging.
+:since:`Since v10.6.0`
+
+::
+
+ ...
+ <pstore backend='acpi-erst'>
+ <path>/tmp/guest_acpi_esrt</path>
+ <size unit='KiB'>8</size>
+ <address type='pci' domain='0x0000' bus='0x02'
slot='0x01' function='0x0'/>
+ </pstore>
+ ...
+
+The ``pstore`` element has one mandatory attribute ``backend`` which selects
+desired backend (only ``acpi-erst`` is accepted for now). Then it has the
+following child elements:
+
+``path``
+ Represents a path in the host that backs the pstore device in the guest. It
+ is mandatory.
I know, very late feedback, but couldn't we fill this in for the user
if no value has been provided, e.g.
/var/lib/libvirt/qemu/nvram/mydomain_PSTORE.raw
or something like that?
I suppose using a well-known directory would help avoid file
permission issues and SELinux denials.
Also IIUC we're leaving creation of the file to QEMU instead of
handling that part ourselves, which is what we do in other cases such
as UEFI VARS storage. Are we okay with that?
--
Andrea Bolognani / Red Hat / Virtualization