Re: [Qemu-devel] [PATCH 2/3] s390x/css: generate solicited crw for rchp completion signaling

[Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>]

* Cornelia Huck <cohuck@redhat.com> [2017-07-27 13:22:59 +0200]:

[...]

> > diff --git a/hw/s390x/css.c b/hw/s390x/css.c
[...]

> > @@ -2028,7 +2028,7 @@ void css_subch_assign(uint8_t cssid, uint8_t ssid, uint16_t schid,
> >      }
> >  }
> >  
> > -void css_queue_crw(uint8_t rsc, uint8_t erc, int chain, uint16_t rsid)
> > +void css_queue_crw(uint8_t rsc, uint8_t erc, int s, int chain, uint16_t rsid)
> 
> 's' is not a very speaking name...
> 
> >  {
> >      CrwContainer *crw_cont;
> >  
> > @@ -2040,6 +2040,9 @@ void css_queue_crw(uint8_t rsc, uint8_t erc, int chain, uint16_t rsid)
> >          return;
> >      }
> >      crw_cont->crw.flags = (rsc << 8) | erc;
> > +    if (s) {
> > +        crw_cont->crw.flags |= CRW_FLAGS_MASK_S;
> 
> ...as it obviously causes the S flag to be set ;) Let's call it 'solicited'?
Sure. Will adopt.

> 
> > +    }
> >      if (chain) {
> >          crw_cont->crw.flags |= CRW_FLAGS_MASK_C;
> >      }
> > @@ -2086,9 +2089,9 @@ void css_generate_sch_crws(uint8_t cssid, uint8_t ssid, uint16_t schid,
> >      }
> >      chain_crw = (channel_subsys.max_ssid > 0) ||
> >              (channel_subsys.max_cssid > 0);
> > -    css_queue_crw(CRW_RSC_SUBCH, CRW_ERC_IPI, chain_crw ? 1 : 0, schid);
> > +    css_queue_crw(CRW_RSC_SUBCH, CRW_ERC_IPI, 0, chain_crw ? 1 : 0, schid);
> >      if (chain_crw) {
> > -        css_queue_crw(CRW_RSC_SUBCH, CRW_ERC_IPI, 0,
> > +        css_queue_crw(CRW_RSC_SUBCH, CRW_ERC_IPI, 0, 0,
> >                        (guest_cssid << 8) | (ssid << 4));
> >      }
> >      /* RW_ERC_IPI --> clear pending interrupts */
> > @@ -2103,7 +2106,7 @@ void css_generate_chp_crws(uint8_t cssid, uint8_t chpid)
> >  void css_generate_css_crws(uint8_t cssid)
> >  {
> >      if (!channel_subsys.sei_pending) {
> > -        css_queue_crw(CRW_RSC_CSS, CRW_ERC_EVENT, 0, cssid);
> > +        css_queue_crw(CRW_RSC_CSS, CRW_ERC_EVENT, 0, 0, cssid);
> 
> Should we want to support OS-triggered channel path vary (via SCLP or
> otherwise) in the future,
Yes! I had a prototype of series to handle the OS-triggered chp vary,
and that needs...

> we'll probably need a version that generates a solicited crw.
...the new interface with the solicited bit param, which is provided by
patch #3:
void css_generate_chp_crws(uint8_t cssid, uint8_t chpid,
                           int hotplugged, int add, int s);

BTW, I need to renew the legal clearance before sending them out...

> 
> >      }
> >      channel_subsys.sei_pending = true;
> >  }
> > diff --git a/include/hw/s390x/css.h b/include/hw/s390x/css.h
> > index 5c5fe6b202..d03b4ffeac 100644
> > --- a/include/hw/s390x/css.h
> > +++ b/include/hw/s390x/css.h
> > @@ -150,7 +150,7 @@ void copy_scsw_to_guest(SCSW *dest, const SCSW *src);
> >  void css_inject_io_interrupt(SubchDev *sch);
> >  void css_reset(void);
> >  void css_reset_sch(SubchDev *sch);
> > -void css_queue_crw(uint8_t rsc, uint8_t erc, int chain, uint16_t rsid);
> > +void css_queue_crw(uint8_t rsc, uint8_t erc, int s, int chain, uint16_t rsid);
> >  void css_generate_sch_crws(uint8_t cssid, uint8_t ssid, uint16_t schid,
> >                             int hotplugged, int add);
> >  void css_generate_chp_crws(uint8_t cssid, uint8_t chpid);
> 
> Otherwise, patch looks good.
Thanks.

So, I only need to s/s/solicited for the new version of this one?

> 

-- 
Dong Jia Shi