Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-28 Thread Stefano Stabellini
On Mon, 28 Mar 2022, Daniel P. Smith wrote: > On 3/25/22 17:05, Stefano Stabellini wrote: > > On Fri, 25 Mar 2022, Julien Grall wrote: > >> So to me, the idea of switching to a "fake" domain or bypassing the check > >> is > >> more appealing. I have a preference for the "fake" domain here. > > >

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-28 Thread Daniel P. Smith
On 3/25/22 17:05, Stefano Stabellini wrote: > On Fri, 25 Mar 2022, Julien Grall wrote: >> So to me, the idea of switching to a "fake" domain or bypassing the check is >> more appealing. I have a preference for the "fake" domain here. > > As a maintainer, I am not opposed to the

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-28 Thread Daniel P. Smith
On 3/25/22 17:05, Stefano Stabellini wrote: > On Fri, 25 Mar 2022, Julien Grall wrote: >> So to me, the idea of switching to a "fake" domain or bypassing the check is >> more appealing. I have a preference for the "fake" domain here. > > As a maintainer, I am not opposed to the

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-28 Thread Julien Grall
Hi Stefano, On 25/03/2022 21:05, Stefano Stabellini wrote: On Fri, 25 Mar 2022, Julien Grall wrote: > As a contributor, sadly I won't be able to spend a lot of time on this in the following months. If a significant rework is required, I don't think I'll be able to do it, at least not for this

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-28 Thread Daniel P. Smith
On 3/25/22 12:52, Jason Andryuk wrote: > On Fri, Mar 25, 2022 at 11:46 AM Daniel P. Smith > wrote: >> >> On 3/24/22 20:30, Stefano Stabellini wrote: >>> On Wed, 23 Mar 2022, Jan Beulich wrote: On 23.03.2022 01:22, Stefano Stabellini wrote: > The existing XSM check in

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-25 Thread Stefano Stabellini
On Fri, 25 Mar 2022, Julien Grall wrote: > So to me, the idea of switching to a "fake" domain or bypassing the check is > more appealing. I have a preference for the "fake" domain here. As a maintainer, I am not opposed to the "fake"/"contructor" domain concept. It all depends on how many

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-25 Thread Julien Grall
Hi Stefano, On 25/03/2022 00:30, Stefano Stabellini wrote: On Wed, 23 Mar 2022, Jan Beulich wrote: On 23.03.2022 01:22, Stefano Stabellini wrote: On Tue, 15 Mar 2022, Daniel P. Smith wrote: On 1/28/22 16:33, Stefano Stabellini wrote: From: Luca Miccio The xenstore event channel will be

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-25 Thread Julien Grall
Hi Daniel, On 25/03/2022 15:45, Daniel P. Smith wrote: The existing XSM check in evtchn_alloc_unbound cannot work and should not work: it is based on the current domain having enough privileges to create the event channel. In this case, we have no current domain at all. The current domain is

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-25 Thread Jason Andryuk
On Fri, Mar 25, 2022 at 11:46 AM Daniel P. Smith wrote: > > On 3/24/22 20:30, Stefano Stabellini wrote: > > On Wed, 23 Mar 2022, Jan Beulich wrote: > >> On 23.03.2022 01:22, Stefano Stabellini wrote: > >>> The existing XSM check in evtchn_alloc_unbound cannot work and should > >>> not work: it is

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-25 Thread Daniel P. Smith
On 3/24/22 20:30, Stefano Stabellini wrote: > On Wed, 23 Mar 2022, Jan Beulich wrote: >> On 23.03.2022 01:22, Stefano Stabellini wrote: >>> On Tue, 15 Mar 2022, Daniel P. Smith wrote: On 1/28/22 16:33, Stefano Stabellini wrote: > From: Luca Miccio > > The xenstore event channel

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-25 Thread Jan Beulich
On 25.03.2022 01:30, Stefano Stabellini wrote: > Maybe, instead of exporting get_free_port, we could create a new > function in xen/common/event_channel.c and mark it as __init? This way: > - we don't need to expose get_free_port > - the new function would only be __init anyway, so no risk of

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-24 Thread Stefano Stabellini
On Wed, 23 Mar 2022, Jan Beulich wrote: > On 23.03.2022 01:22, Stefano Stabellini wrote: > > On Tue, 15 Mar 2022, Daniel P. Smith wrote: > >> On 1/28/22 16:33, Stefano Stabellini wrote: > >>> From: Luca Miccio > >>> > >>> The xenstore event channel will be allocated for dom0less domains. It is >

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-24 Thread Daniel P. Smith
On 3/22/22 20:22, Stefano Stabellini wrote: > On Tue, 15 Mar 2022, Daniel P. Smith wrote: >> On 1/28/22 16:33, Stefano Stabellini wrote: >>> From: Luca Miccio >>> >>> The xenstore event channel will be allocated for dom0less domains. It is >>> necessary to have access to the evtchn_alloc_unbound

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-23 Thread Jan Beulich
On 23.03.2022 01:22, Stefano Stabellini wrote: > On Tue, 15 Mar 2022, Daniel P. Smith wrote: >> On 1/28/22 16:33, Stefano Stabellini wrote: >>> From: Luca Miccio >>> >>> The xenstore event channel will be allocated for dom0less domains. It is >>> necessary to have access to the

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-22 Thread Stefano Stabellini
On Tue, 15 Mar 2022, Daniel P. Smith wrote: > On 1/28/22 16:33, Stefano Stabellini wrote: > > From: Luca Miccio > > > > The xenstore event channel will be allocated for dom0less domains. It is > > necessary to have access to the evtchn_alloc_unbound function to do > > that, so make

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-03-15 Thread Daniel P. Smith
On 1/28/22 16:33, Stefano Stabellini wrote: > From: Luca Miccio > > The xenstore event channel will be allocated for dom0less domains. It is > necessary to have access to the evtchn_alloc_unbound function to do > that, so make evtchn_alloc_unbound public. > > Add a skip_xsm parameter to allow

Re: [PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-01-31 Thread Jan Beulich
On 28.01.2022 22:33, Stefano Stabellini wrote: > From: Luca Miccio > > The xenstore event channel will be allocated for dom0less domains. It is > necessary to have access to the evtchn_alloc_unbound function to do > that, so make evtchn_alloc_unbound public. > > Add a skip_xsm parameter to

[PATCH v3 2/5] xen: make evtchn_alloc_unbound public

2022-01-28 Thread Stefano Stabellini
From: Luca Miccio The xenstore event channel will be allocated for dom0less domains. It is necessary to have access to the evtchn_alloc_unbound function to do that, so make evtchn_alloc_unbound public. Add a skip_xsm parameter to allow disabling the XSM check in evtchn_alloc_unbound