qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Qemu-devel] KVM call minutes for July 27


From: Chris Wright
Subject: [Qemu-devel] KVM call minutes for July 27
Date: Tue, 27 Jul 2010 08:03:28 -0700
User-agent: Mutt/1.5.20 (2009-08-17)

0.13
- -rc0 tagged, propagating now
- no more features, bug fix only
- although a few things, like shared memory device, are still feasible

qemu64 cpu model
- currently model 2
  - this cpu simply does not exist at all in the real world
- model 13 or higher windows 32bit will enable MSI/-X support
  - anyone aware of issues with simply bumping the model
- must retain compatibility with -M
- cpu models fully configurable in config files
  - should move default to config files
- raises a couple questions
  - does "qemu64" need to have a single stable definition?
  - does default cpu make sense
- also the physical models are broken (Conroe, Penryn, etc..)
  - these are simply broken and need to change
- can create versions of base model (qemu64-v1/0.13/whatever)

probed_raw
- 79368c81 closed security hole
- qraw addressing theoretical issue and has too much magic
- any further discussion, list is ---> that way

qemu -help parsing
- libvirt current issues
  - qemu -help/-version was changed and broke libvirt (fixed 0.8.2)
  - libvirt improperly parses cache= (fixed 0.8.3)
- reverting version string change: f75ca1ae
- apply bruce's cache -help patch
- no further significant -help changes
- libvirt uses version only (and maintains the version meaning for
  downstreams)
- eventually capabilities fixes this
- minimal "info capabilities" that usable tomorrow?
  - becomes a suppored interface, deprecating it will be complicated
  - unclear if it buys anything, -help parsing working now, the interim
    sol'n would be thrown away
- could isolate libvirt -help parser and toss it into qemu as a test



reply via email to

[Prev in Thread] Current Thread [Next in Thread]