Re: [PATCH net-next v2 7/7] net/smc: manage system EID in SMC stack instead of ISM driver

[Wen Gu]

On 2023/11/27 22:04, Alexandra Winter wrote:
> On 24.11.23 15:42, Wen Gu wrote:
> > The System EID (SEID) is an internal EID that is used by the SMCv2
> > software stack that has a predefined and constant value representing
> > the s390 physical machine that the OS is executing on. So it should
> > be managed by SMC stack instead of ISM driver and be consistent for
> > all ISMv2 device (including virtual ISM devices) on s390 architecture.
> >
> > Suggested-by: Alexandra Winter <wintera@xxxxxxxxxxxxx>
> > Signed-off-by: Wen Gu <guwen@xxxxxxxxxxxxxxxxx>
> > ---
>
> Yes, this is what I had in mind. Thank you Wen Gu.

:)
> [...]
>
> > diff --git a/drivers/s390/net/ism.h b/drivers/s390/net/ism.h
> > index 70c5bbd..49ccbd68 100644
> > --- a/drivers/s390/net/ism.h
> > +++ b/drivers/s390/net/ism.h
>
> Please remove ISM_IDENT_MASK from drivers/s390/net/ism.h
> [...]

Thanks for reminding. I will remove it.

> > --- a/drivers/s390/net/ism_drv.c
> > +++ b/drivers/s390/net/ism_drv.c
> > @@ -36,6 +36,7 @@
> [...]
> > -static void ism_create_system_eid(void)
> > -{
> > -	struct cpuid id;
> > -	u16 ident_tail;
> > -	char tmp[5];
> > -
> > -	get_cpu_id(&id);
> > -	ident_tail = (u16)(id.ident & ISM_IDENT_MASK);
> > -	snprintf(tmp, 5, "%04X", ident_tail);
> > -	memcpy(&SYSTEM_EID.serial_number, tmp, 4);
> > -	snprintf(tmp, 5, "%04X", id.machine);
> > -	memcpy(&SYSTEM_EID.type, tmp, 4);
> > -}
> > -
> [...]
> > @@ -560,7 +535,7 @@ static int ism_dev_init(struct ism_dev *ism)
> >   
> >   	if (!ism_add_vlan_id(ism, ISM_RESERVED_VLANID))
> >   		/* hardware is V2 capable */
> > -		ism_create_system_eid();
> > +		ism_v2_capable = true;
> >   
>
> Please assign 'false' in the else path.
> This is required here for backwards compatibility. Hardware that only supports v1,
> will reject ISM_RESERVED_VLANID.

OK. I will assign false in the else path to explicitly express the meaning. Thank you.

> [...]
>
>
> > --- a/net/smc/smc_ism.c
> > +++ b/net/smc/smc_ism.c
> [...]
> > @@ -70,6 +91,11 @@ bool smc_ism_is_v2_capable(void)
> >   	return smc_ism_v2_capable;
> >   }
> >   
> > +void smc_ism_set_v2_capable(void)
> > +{
> > +	smc_ism_v2_capable = true;
> > +}
> > +
> >   /* Set a connection using this DMBE. */
> >   void smc_ism_set_conn(struct smc_connection *conn)
> >   {
> > @@ -431,14 +457,8 @@ static void smcd_register_dev(struct ism_dev *ism)
> >   
> >   	mutex_lock(&smcd_dev_list.mutex);
> >   	if (list_empty(&smcd_dev_list.list)) {
> > -		u8 *system_eid = NULL;
> > -
> > -		system_eid = smcd->ops->get_system_eid();
> > -		if (smcd->ops->supports_v2()) {
> > -			smc_ism_v2_capable = true;
> > -			memcpy(smc_ism_v2_system_eid, system_eid,
> > -			       SMC_MAX_EID_LEN);
> > -		}
> > +		if (smcd->ops->supports_v2())
> > +			smc_ism_set_v2_capable();
>
> I don't see the benefit in declaring smc_ism_set_v2_capable() and exporting it in smc_ism.h,
> when it is used only once and only here.
> Why don't you just set
> 	smc_ism_v2_capable = true;
> here?

Yes.. it may be confusing if readers only look at this patch set.

It is because loopback-ism (or other kinds of ISMv2.1) will also use this helper in such
as smc_loopback.c to set smc_ism_v2_capable (defined in smc_ism.c) as true. So I think it
may be better to introduce a helper here instead of exposing smc_ism_v2_capable variable.

Thanks.

> [...]
> > diff --git a/net/smc/smc_ism.h b/net/smc/smc_ism.h
> > index 0e5e563..6903cd5 100644
> > --- a/net/smc/smc_ism.h
> > +++ b/net/smc/smc_ism.h
> > @@ -16,6 +16,7 @@
> >   #include "smc.h"
> >   
> >   #define SMC_VIRTUAL_ISM_CHID_MASK	0xFF00
> > +#define SMC_ISM_IDENT_MASK		0x00FFFF
> >   
> [...]
> > @@ -45,6 +52,7 @@ int smc_ism_register_dmb(struct smc_link_group *lgr, int buf_size,
> >   void smc_ism_get_system_eid(u8 **eid);
> >   u16 smc_ism_get_chid(struct smcd_dev *dev);
> >   bool smc_ism_is_v2_capable(void);
> > +void smc_ism_set_v2_capable(void);
> >   int smc_ism_init(void);
> >   void smc_ism_exit(void);
> >   int smcd_nl_get_device(struct sk_buff *skb, struct netlink_callback *cb);