All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH 0/1] acpi: Implement ACPI ERST support for guests
@ 2020-10-26 20:19 Eric DeVolder
  2020-10-26 20:19 ` [PATCH 1/1] " Eric DeVolder
                   ` (3 more replies)
  0 siblings, 4 replies; 7+ messages in thread
From: Eric DeVolder @ 2020-10-26 20:19 UTC (permalink / raw)
  To: qemu-devel
  Cc: ehabkost, mst, konrad.wilk, pbonzini, imammedo, boris.ostrovsky, rth

This changeset introduces support for the ACPI Error Record
Serialization Table, ERST.

The change to hw/acpi/meson.build simply adds in the new .c file
for compilation.

The change to hw/i386/acpi-build.c calls out the building of the
ERST table (and also creates the associated device).

The new file hw/acpi/erst.c contains the building of the ERST
table, as well as the simple device for exchanging error records.

The new file include/hw/acpi/erst.h contains associated definitions
and declarations for ERST.

The primary description of this changeset is in the patch commit
message.

NOTES: When reviewing, I would especially appreciate feedback
on the following topics:

- The hope is to have ERST always present if ACPI is enabled, however,
  I have found it difficult to devise a method for passing the base
  address that does not require the workaround at the bottom of
  build_erst(). The issues I encountered are:
  - desire to keep this is common ACPI code
  - the device requires a qdev_new(), this needs to happen early,
    thus the workaround in build_erst()
  - the base address is machine/arch specific (eg ARM vs x86)
  I've not found a nice way to thread this needle, so what I've settled
  on is to simply lump ERST on to the CONFIG_ACPI (rather than a
  separate CONFIG_ACPI_ERST), and the workaround at the bottom of
  build_erst(). I suspect there is a better way for a built-in/
  always present device. This does not support "-device acpi-erst,...".

- I found a base address that "worked", but would like an address
  that would be known to be availabe, and then to document/reserve
  it for ERST. This takes into account that the base address can be
  different for x86 vs ARM.

- I've run this through checkpatch, and all issues addressed except
  for the long lines in build_erst(). For readable I left the long
  lines, but will change if asked.

- What else do I need to provide?

Signed-off-by: Eric DeVolder <eric.devolder@oracle.com>

---
 hw/acpi/erst.c         | 909 +++++++++++++++++++++++++++++++++++++++++++++++++
 hw/acpi/meson.build    |   1 +
 hw/i386/acpi-build.c   |   4 +
 include/hw/acpi/erst.h |  97 ++++++
 4 files changed, 1011 insertions(+)
 create mode 100644 hw/acpi/erst.c
 create mode 100644 include/hw/acpi/erst.h



^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2021-02-09  0:43 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-10-26 20:19 [PATCH 0/1] acpi: Implement ACPI ERST support for guests Eric DeVolder
2020-10-26 20:19 ` [PATCH 1/1] " Eric DeVolder
2020-11-03 21:16   ` Paolo Bonzini
2021-02-08 21:08   ` Eric Devolder
2020-10-26 20:54 ` [PATCH 0/1] " no-reply
2020-11-03 14:57 ` Igor Mammedov
2021-02-08 21:07 ` Eric Devolder

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.