Re: [PATCH v3 1/6] argo: lower level of noisy connection-refused log

[dmukhin@xxxxxxxx]

On Mon, Jun 08, 2026 at 03:54:51PM -0400, Daniel P. Smith wrote:
> 
> 
> On 5/26/26 5:58 PM, dmukhin@xxxxxxxx wrote:
> > From: Denis Mukhin <dmukhin@xxxxxxxx>
> > 
> > Switch the log line to argo_dprintk() so it is enabled only in debug
> > environments, as it can spam the logs when a dom0 service using the Argo
> > hypercall tries to communicate with a domain that is still starting up.
> > 
> > Note that this also lowers the log level to debug when the argo_dprintk()
> > facility is enabled.
> > 
> > Signed-off-by: Denis Mukhin <dmukhin@xxxxxxxx>
> > Reviewed-by: Mykola Kvach <mykola_kvach@xxxxxxxx>
> > ---
> > Changes since v2:
> > - updated commit message
> > ---
> >   xen/common/argo.c | 7 +++----
> >   1 file changed, 3 insertions(+), 4 deletions(-)
> > 
> > diff --git a/xen/common/argo.c b/xen/common/argo.c
> > index 28626e00a8cb..98a3db7fd070 100644
> > --- a/xen/common/argo.c
> > +++ b/xen/common/argo.c
> > @@ -2034,10 +2034,9 @@ sendv(struct domain *src_d, xen_argo_addr_t 
> > *src_addr,
> >                                           src_id.domain_id);
> >       if ( !ring_info )
> >       {
> > -        gprintk(XENLOG_ERR,
> > -                "argo: vm%u connection refused, src (vm%u:%x) dst 
> > (vm%u:%x)\n",
> > -                current->domain->domain_id, src_id.domain_id, src_id.aport,
> > -                dst_addr->domain_id, dst_addr->aport);
> > +        argo_dprintk("vm%u connection refused, src (vm%u:%x) dst 
> > (vm%u:%x)\n",
> > +                     current->domain->domain_id, src_id.domain_id, 
> > src_id.aport,
> > +                     dst_addr->domain_id, dst_addr->aport);
> >           ret = -ECONNREFUSED;
> >       }
> 
> My apologies but this is not the wisest approach, hitting this is a real
> error and shouldn't be getting silenced. If you are seeing a lot of these
> messages, I would suggest asking yourself why. Without further context on
> how you are using it, one suggesting is perhaps your connection model might
> need to be revisited.

Thanks for the feedback!

The reason I wrote this patch is because there can be a lot of those messages
from a real domU which is in a boot loop and/or boots for a significant amount
of time.

With -ECONNREFUSED propagated back to the caller having Xen logging the state
of each sendv() with XENLOG_ERR on the shared (dom0) diag console can be a bit
problematic.

Yes, I understand that programs issuing the Argo hypercall could be rewritten
in a particular way so that the Argo hypercall is issued only if the domain is
up and Argo on the other side is initialized and, ideally, domain should not
be boot-looping...

However, I think, that hypervisor should not depend on assumptions made in the
userspace (e.g. retry/logging policy).

De-prioritizing the logline from XENLOG_ERR to XENLOG_DEBUG can be another
potential solution.

--
Denis