Mailing List Archive

[PATCH] libxl: User defined max_maptrack_frames in a stub domain
Hi,

Implementing qrexec+usbip+qemu in Linux-based stub domain leads me to
an issue where a device model stub domain doesn't have maptrack entries.

Would it be possible to apply a user defined max_maptrack_frames value
to dm_config in the same way as for max_grant_frames?

Signed-off-by: Dmitry Fedorov <d.fedorov@tabit.pro>
---
 tools/libxl/libxl_dm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/libxl/libxl_dm.c b/tools/libxl/libxl_dm.c
index f2dc5696b9..f044f2566c 100644
--- a/tools/libxl/libxl_dm.c
+++ b/tools/libxl/libxl_dm.c
@@ -2292,7 +2292,7 @@ void libxl__spawn_stub_dm(libxl__egc *egc,
libxl__stub_dm_spawn_state *sdss)
     dm_config->b_info.target_memkb = dm_config->b_info.max_memkb;

     dm_config->b_info.max_grant_frames =
guest_config->b_info.max_grant_frames;
-    dm_config->b_info.max_maptrack_frames = 0;
+    dm_config->b_info.max_maptrack_frames =
guest_config->b_info.max_maptrack_frames;

     dm_config->b_info.u.pv.features = "";

--
2.26.2
Re: [PATCH] libxl: User defined max_maptrack_frames in a stub domain [ In reply to ]
On 14/09/2020 15:50, Dmitry Fedorov wrote:
> Hi,
>
> Implementing qrexec+usbip+qemu in Linux-based stub domain leads me to
> an issue where a device model stub domain doesn't have maptrack entries.
>
> Would it be possible to apply a user defined max_maptrack_frames value
> to dm_config in the same way as for max_grant_frames?
>
> Signed-off-by: Dmitry Fedorov <d.fedorov@tabit.pro>

This looks entirely reasonable.

CC'ing the maintainers for their opinion.

~Andrew

> ---
>  tools/libxl/libxl_dm.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/tools/libxl/libxl_dm.c b/tools/libxl/libxl_dm.c
> index f2dc5696b9..f044f2566c 100644
> --- a/tools/libxl/libxl_dm.c
> +++ b/tools/libxl/libxl_dm.c
> @@ -2292,7 +2292,7 @@ void libxl__spawn_stub_dm(libxl__egc *egc,
> libxl__stub_dm_spawn_state *sdss)
>      dm_config->b_info.target_memkb = dm_config->b_info.max_memkb;
>
>      dm_config->b_info.max_grant_frames =
> guest_config->b_info.max_grant_frames;
> -    dm_config->b_info.max_maptrack_frames = 0;
> +    dm_config->b_info.max_maptrack_frames =
> guest_config->b_info.max_maptrack_frames;
>
>      dm_config->b_info.u.pv.features = "";
>
> --
> 2.26.2
>
>
Re: [PATCH] libxl: User defined max_maptrack_frames in a stub domain [ In reply to ]
On Mon, Sep 14, 2020 at 04:27:45PM +0100, Andrew Cooper wrote:
> On 14/09/2020 15:50, Dmitry Fedorov wrote:
> > Hi,
> >
> > Implementing qrexec+usbip+qemu in Linux-based stub domain leads me to
> > an issue where a device model stub domain doesn't have maptrack entries.
> >
> > Would it be possible to apply a user defined max_maptrack_frames value
> > to dm_config in the same way as for max_grant_frames?
> >
> > Signed-off-by: Dmitry Fedorov <d.fedorov@tabit.pro>
>
> This looks entirely reasonable.
>
> CC'ing the maintainers for their opinion.
>

Looks fine to me.

Acked-by: Wei Liu <wl@xen.org>

> ~Andrew
>
> > ---
> > ?tools/libxl/libxl_dm.c | 2 +-
> > ?1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/tools/libxl/libxl_dm.c b/tools/libxl/libxl_dm.c
> > index f2dc5696b9..f044f2566c 100644
> > --- a/tools/libxl/libxl_dm.c
> > +++ b/tools/libxl/libxl_dm.c
> > @@ -2292,7 +2292,7 @@ void libxl__spawn_stub_dm(libxl__egc *egc,
> > libxl__stub_dm_spawn_state *sdss)
> > ???? dm_config->b_info.target_memkb = dm_config->b_info.max_memkb;
> >
> > ???? dm_config->b_info.max_grant_frames =
> > guest_config->b_info.max_grant_frames;
> > -??? dm_config->b_info.max_maptrack_frames = 0;
> > +??? dm_config->b_info.max_maptrack_frames =
> > guest_config->b_info.max_maptrack_frames;
> >
> > ???? dm_config->b_info.u.pv.features = "";
> >
> > --
> > 2.26.2
> >
> >
>
Re: [PATCH] libxl: User defined max_maptrack_frames in a stub domain [ In reply to ]
On Tue, Sep 22, 2020 at 9:31 AM Wei Liu <wl@xen.org> wrote:
>
> On Mon, Sep 14, 2020 at 04:27:45PM +0100, Andrew Cooper wrote:
> > On 14/09/2020 15:50, Dmitry Fedorov wrote:
> > > Hi,
> > >
> > > Implementing qrexec+usbip+qemu in Linux-based stub domain leads me to
> > > an issue where a device model stub domain doesn't have maptrack entries.
> > >
> > > Would it be possible to apply a user defined max_maptrack_frames value
> > > to dm_config in the same way as for max_grant_frames?
> > >
> > > Signed-off-by: Dmitry Fedorov <d.fedorov@tabit.pro>
> >
> > This looks entirely reasonable.
> >
> > CC'ing the maintainers for their opinion.
> >
>
> Looks fine to me.
>
> Acked-by: Wei Liu <wl@xen.org>

Hi,

Wei, looks like you Acked but did not apply this patch. And after the
libs rename, the file paths no longer match. Do you want to fix this
up, or should Dmitry re-submit?

Regards,
Jason

> > ~Andrew
> >
> > > ---
> > > tools/libxl/libxl_dm.c | 2 +-
> > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/tools/libxl/libxl_dm.c b/tools/libxl/libxl_dm.c
> > > index f2dc5696b9..f044f2566c 100644
> > > --- a/tools/libxl/libxl_dm.c
> > > +++ b/tools/libxl/libxl_dm.c
> > > @@ -2292,7 +2292,7 @@ void libxl__spawn_stub_dm(libxl__egc *egc,
> > > libxl__stub_dm_spawn_state *sdss)
> > > dm_config->b_info.target_memkb = dm_config->b_info.max_memkb;
> > >
> > > dm_config->b_info.max_grant_frames =
> > > guest_config->b_info.max_grant_frames;
> > > - dm_config->b_info.max_maptrack_frames = 0;
> > > + dm_config->b_info.max_maptrack_frames =
> > > guest_config->b_info.max_maptrack_frames;
> > >
> > > dm_config->b_info.u.pv.features = "";
> > >
> > > --
> > > 2.26.2
> > >
> > >
> >
>
Re: [PATCH] libxl: User defined max_maptrack_frames in a stub domain [ In reply to ]
On 10/04/2021 16:55, Jason Andryuk wrote:
> On Tue, Sep 22, 2020 at 9:31 AM Wei Liu <wl@xen.org> wrote:
>> On Mon, Sep 14, 2020 at 04:27:45PM +0100, Andrew Cooper wrote:
>>> On 14/09/2020 15:50, Dmitry Fedorov wrote:
>>>> Hi,
>>>>
>>>> Implementing qrexec+usbip+qemu in Linux-based stub domain leads me to
>>>> an issue where a device model stub domain doesn't have maptrack entries.
>>>>
>>>> Would it be possible to apply a user defined max_maptrack_frames value
>>>> to dm_config in the same way as for max_grant_frames?
>>>>
>>>> Signed-off-by: Dmitry Fedorov <d.fedorov@tabit.pro>
>>> This looks entirely reasonable.
>>>
>>> CC'ing the maintainers for their opinion.
>>>
>> Looks fine to me.
>>
>> Acked-by: Wei Liu <wl@xen.org>
> Hi,
>
> Wei, looks like you Acked but did not apply this patch. And after the
> libs rename, the file paths no longer match. Do you want to fix this
> up, or should Dmitry re-submit?

Apologies for the process failure here.

I have committed
https://xenbits.xen.org/gitweb/?p=xen.git;a=commitdiff;h=80714e55042b2242cd82eccfb7405d671ecbebda

I take it that it wants flagging for backports?

~Andrew
Re: [PATCH] libxl: User defined max_maptrack_frames in a stub domain [ In reply to ]
On Tue, Apr 13, 2021 at 10:23 AM Andrew Cooper
<andrew.cooper3@citrix.com> wrote:
>
> On 10/04/2021 16:55, Jason Andryuk wrote:
> > On Tue, Sep 22, 2020 at 9:31 AM Wei Liu <wl@xen.org> wrote:
> >> On Mon, Sep 14, 2020 at 04:27:45PM +0100, Andrew Cooper wrote:
> >>> On 14/09/2020 15:50, Dmitry Fedorov wrote:
> >>>> Hi,
> >>>>
> >>>> Implementing qrexec+usbip+qemu in Linux-based stub domain leads me to
> >>>> an issue where a device model stub domain doesn't have maptrack entries.
> >>>>
> >>>> Would it be possible to apply a user defined max_maptrack_frames value
> >>>> to dm_config in the same way as for max_grant_frames?
> >>>>
> >>>> Signed-off-by: Dmitry Fedorov <d.fedorov@tabit.pro>
> >>> This looks entirely reasonable.
> >>>
> >>> CC'ing the maintainers for their opinion.
> >>>
> >> Looks fine to me.
> >>
> >> Acked-by: Wei Liu <wl@xen.org>
> > Hi,
> >
> > Wei, looks like you Acked but did not apply this patch. And after the
> > libs rename, the file paths no longer match. Do you want to fix this
> > up, or should Dmitry re-submit?
>
> Apologies for the process failure here.

No worries.

> I have committed
> https://xenbits.xen.org/gitweb/?p=xen.git;a=commitdiff;h=80714e55042b2242cd82eccfb7405d671ecbebda

Thank you, Andrew.

> I take it that it wants flagging for backports?

The Qubes qrexec+usbip+qemu feature isn't merged yet, but they have a
backport to 4.14 as part of the series.

Regards,
Jason