On 29/10/2025 13:25, BALATON Zoltan wrote:
On Wed, 29 Oct 2025, BALATON Zoltan wrote:
On Wed, 29 Oct 2025, Akihiko Odaki wrote:
On 2025/10/29 6:28, BALATON Zoltan wrote:
On Wed, 29 Oct 2025, Akihiko Odaki wrote:
On 2025/10/28 21:59, BALATON Zoltan wrote:
On Tue, 28 Oct 2025, Philippe Mathieu-Daudé wrote:
On 27/10/25 20:47, BALATON Zoltan wrote:
On Mon, 27 Oct 2025, Philippe Mathieu-Daudé wrote:
On 25/10/25 01:31, BALATON Zoltan wrote:
These memory windows are a result of the address decoding in the
Articia S north bridge so better model it there and not in board code.
Suggested-by: Philippe Mathieu-Daudé <[email protected]>
Signed-off-by: BALATON Zoltan <[email protected]>
---
hw/pci-host/articia.c | 15 ++++++++++++++-
hw/ppc/amigaone.c | 28 +++++-----------------------
hw/ppc/pegasos2.c | 13 -------------
3 files changed, 19 insertions(+), 37 deletions(-)
[...]
It looks like we won't be able to come to an agreement before the freeze and I
don't have time now to change this patch but don't want to miss the release with
this series that finishes pegasos renaming because of this. So for this patch I'd
say since this is already how it is now and it does not make it worse and this
object is not user creatable anyway so cannot leak please take it as it is and
we'll do a clean up later after we finish discussion.
As for all of these files I'm the maintainer let me make an executive decision here
to keep this patch without Philippe's reviewed-by for now to be able to move on with
this series before the freeze. Fixing the theoretical leak can be done on top and
since that's a fix it can be done during soft freeze that would give us more time. So
Harsh please go ahead and merge this series too if there are no other concerns. I'll
then address this later together with other similar issues elsewhere.
I think the issue here is that several people have now raised concerns as to the way
you are trying to use MemoryRegions (both here and in the raven series): simply put,
if reviewers didn't see any problems with this approach, your series would already
have review tags.
If you want to suggest a different approach to managing MemoryRegions, then I would
suggest you send a proposal to the mailing list so it can be reviewed by the QOM and
memory people (along with updated code style documentation so we can all use it). But
with freeze coming up in a few days this is certainly out of scope for 10.2.
For now I would suggest the easiest way to get review tags is to stick with the
existing inline MemoryRegion approach for 10.2 to aid getting your patches merged:
it's simply not fair to put time pressure on Harsh to merge patches in this way.
ATB,
Mark.
