| Message ID | 1532702862-22751-3-git-send-email-arbab@linux.ibm.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | npu2: Add support for relaxed-ordering mode | expand |
| Context | Check | Description |
|---|---|---|
| snowpatch_ozlabs/apply_patch | success | master/apply_patch Successfully applied |
Looks good to me. Reviewed-by: Alistair Popple <alistair@popple.id.au> On Friday, 27 July 2018 9:47:38 AM AEST Reza Arbab wrote: > Add an iterator to walk over every PHB. > > Signed-off-by: Alistair Popple <alistair@popple.id.au> > Signed-off-by: Reza Arbab <arbab@linux.ibm.com> > --- > core/pci.c | 12 ++++++++++++ > include/pci.h | 1 + > 2 files changed, 13 insertions(+) > > diff --git a/core/pci.c b/core/pci.c > index 6ba6eda..4faf88e 100644 > --- a/core/pci.c > +++ b/core/pci.c > @@ -1104,6 +1104,18 @@ struct phb *pci_get_phb(uint64_t phb_id) > return phbs[phb_id]; > } > > +/* Iterator to walk over every PHB */ > +struct phb *pci_next_phb(struct phb *last) > +{ > + uint64_t phb_id = last ? last->opal_id + 1 : 0; > + > + for (; phb_id < ARRAY_SIZE(phbs); phb_id++) > + if (phbs[phb_id]) > + return phbs[phb_id]; > + > + return NULL; > +} > + > static const char *pci_class_name(uint32_t class_code) > { > uint8_t class = class_code >> 16; > diff --git a/include/pci.h b/include/pci.h > index c6aa664..19fe9c0 100644 > --- a/include/pci.h > +++ b/include/pci.h > @@ -481,6 +481,7 @@ extern struct pci_cfg_reg_filter *pci_add_cfg_reg_filter(struct pci_device *pd, > extern int64_t pci_register_phb(struct phb *phb, int opal_id); > extern int64_t pci_unregister_phb(struct phb *phb); > extern struct phb *pci_get_phb(uint64_t phb_id); > +extern struct phb *pci_next_phb(struct phb *last); > > static inline struct phb *__pci_next_phb_idx(uint64_t *phb_id) { > struct phb *phb = NULL; >
On Sat, Jul 28, 2018 at 12:47 AM, Reza Arbab <arbab@linux.ibm.com> wrote: > Add an iterator to walk over every PHB. > > Signed-off-by: Alistair Popple <alistair@popple.id.au> > Signed-off-by: Reza Arbab <arbab@linux.ibm.com> > --- > core/pci.c | 12 ++++++++++++ > include/pci.h | 1 + > 2 files changed, 13 insertions(+) > > diff --git a/core/pci.c b/core/pci.c > index 6ba6eda..4faf88e 100644 > --- a/core/pci.c > +++ b/core/pci.c > @@ -1104,6 +1104,18 @@ struct phb *pci_get_phb(uint64_t phb_id) > return phbs[phb_id]; > } > > +/* Iterator to walk over every PHB */ > +struct phb *pci_next_phb(struct phb *last) > +{ > + uint64_t phb_id = last ? last->opal_id + 1 : 0; > + > + for (; phb_id < ARRAY_SIZE(phbs); phb_id++) > + if (phbs[phb_id]) > + return phbs[phb_id]; > + > + return NULL; > +} So __pci_next_phb_idx() is kind of weird so I don't mind replacing it with this, but you should probably do the replacement too ;) > + > static const char *pci_class_name(uint32_t class_code) > { > uint8_t class = class_code >> 16; > diff --git a/include/pci.h b/include/pci.h > index c6aa664..19fe9c0 100644 > --- a/include/pci.h > +++ b/include/pci.h > @@ -481,6 +481,7 @@ extern struct pci_cfg_reg_filter *pci_add_cfg_reg_filter(struct pci_device *pd, > extern int64_t pci_register_phb(struct phb *phb, int opal_id); > extern int64_t pci_unregister_phb(struct phb *phb); > extern struct phb *pci_get_phb(uint64_t phb_id); > +extern struct phb *pci_next_phb(struct phb *last); > > static inline struct phb *__pci_next_phb_idx(uint64_t *phb_id) { > struct phb *phb = NULL; > -- > 1.8.3.1 > > _______________________________________________ > Skiboot mailing list > Skiboot@lists.ozlabs.org > https://lists.ozlabs.org/listinfo/skiboot
On Mon, Jul 30, 2018 at 3:51 PM, Alistair Popple <alistair@popple.id.au> wrote: > Looks good to me. > > Reviewed-by: Alistair Popple <alistair@popple.id.au> hmm > > On Friday, 27 July 2018 9:47:38 AM AEST Reza Arbab wrote: >> Add an iterator to walk over every PHB. >> >> Signed-off-by: Alistair Popple <alistair@popple.id.au> >> Signed-off-by: Reza Arbab <arbab@linux.ibm.com> Hmmmmmm Sounds like someone is gaming the system to get CRAPs. I'm on to you Popple. Oliver
On Mon, Jul 30, 2018 at 04:31:43PM +1000, Oliver wrote: >On Sat, Jul 28, 2018 at 12:47 AM, Reza Arbab <arbab@linux.ibm.com> wrote: >> diff --git a/core/pci.c b/core/pci.c >> index 6ba6eda..4faf88e 100644 >> --- a/core/pci.c >> +++ b/core/pci.c >> @@ -1104,6 +1104,18 @@ struct phb *pci_get_phb(uint64_t phb_id) >> return phbs[phb_id]; >> } >> >> +/* Iterator to walk over every PHB */ >> +struct phb *pci_next_phb(struct phb *last) >> +{ >> + uint64_t phb_id = last ? last->opal_id + 1 : 0; >> + >> + for (; phb_id < ARRAY_SIZE(phbs); phb_id++) >> + if (phbs[phb_id]) >> + return phbs[phb_id]; >> + >> + return NULL; >> +} > >So __pci_next_phb_idx() is kind of weird so I don't mind replacing it >with this, but you should probably do the replacement too ;) Now that I look, the rest of the set doesn't even actually use pci_next_phb(), so I'm going to drop this patch entirely. Man... who wrote this crap?
diff --git a/core/pci.c b/core/pci.c index 6ba6eda..4faf88e 100644 --- a/core/pci.c +++ b/core/pci.c @@ -1104,6 +1104,18 @@ struct phb *pci_get_phb(uint64_t phb_id) return phbs[phb_id]; } +/* Iterator to walk over every PHB */ +struct phb *pci_next_phb(struct phb *last) +{ + uint64_t phb_id = last ? last->opal_id + 1 : 0; + + for (; phb_id < ARRAY_SIZE(phbs); phb_id++) + if (phbs[phb_id]) + return phbs[phb_id]; + + return NULL; +} + static const char *pci_class_name(uint32_t class_code) { uint8_t class = class_code >> 16; diff --git a/include/pci.h b/include/pci.h index c6aa664..19fe9c0 100644 --- a/include/pci.h +++ b/include/pci.h @@ -481,6 +481,7 @@ extern struct pci_cfg_reg_filter *pci_add_cfg_reg_filter(struct pci_device *pd, extern int64_t pci_register_phb(struct phb *phb, int opal_id); extern int64_t pci_unregister_phb(struct phb *phb); extern struct phb *pci_get_phb(uint64_t phb_id); +extern struct phb *pci_next_phb(struct phb *last); static inline struct phb *__pci_next_phb_idx(uint64_t *phb_id) { struct phb *phb = NULL;