Re: Regression in xen-unstable due to commit 3802ecbaa9eb36

[Wei Liu <wei.liu2@citrix.com>]

On Thu, May 16, 2019 at 12:45:40PM +0200, Roger Pau Monné wrote:
> On Thu, May 16, 2019 at 11:07:35AM +0200, Olaf Hering wrote:
> > Am Thu, 16 May 2019 10:09:38 +0200
> > schrieb Juergen Gross <jgross@suse.com>:
> > 
> > > The patch "libxl: add helper function to set device_model_version"
> > > breaks creating any domain for me.
> > 
> > The issue is, create_domain will eventually call freemem.
> > If autoballoon is set, due to dom0_mem= for example, all is fine.
> > If memory has to be freed, libxl_domain_need_memory will get an
> > incomplete b_info. Somehow the new libxl__domain_set_device_model
> > must be called for the d_config returned by parse_config_data.
> > 
> > How should this be fixed?
> 
> Having a field in build_info with a default value that depends on
> fields outside of build_info is problematic, since not all callers of
> libxl__domain_build_info_setdefault have access to libxl_domain_config.
> 
> An option would be to pass libxl_domain_config to
> libxl__domain_build_info_setdefault and fixup the callers. That seems
> like the best solution ATM, but it would require reverting the

That will 1) make the name wrong 2) you will have to conjure up a
domain_config structure even if it is not needed...

Wei.

> currently committed patches, since there won't be a reason anymore to
> split the device model selection code out of
> libxl__domain_build_info_setdefault.
> 
> Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel