On Thu, Jul 20, 2023 at 09:57:29AM +0100, John Garry wrote:
> On 20/07/2023 01:42, Niklas Cassel wrote:
> > From: Hannes Reinecke <hare@xxxxxxx>
> >
> > Is now a wrapper around kfree(), so call it directly.
> >
> > Signed-off-by: Hannes Reinecke <hare@xxxxxxx>
> > Signed-off-by: Niklas Cassel <niklas.cassel@xxxxxxx>
> > ---
> > drivers/ata/libata-sata.c | 18 ------------------
> > drivers/scsi/libsas/sas_ata.c | 2 +-
> > drivers/scsi/libsas/sas_discover.c | 2 +-
> > include/linux/libata.h | 1 -
> > 4 files changed, 2 insertions(+), 21 deletions(-)
> >
> > diff --git a/drivers/ata/libata-sata.c b/drivers/ata/libata-sata.c
> > index d3b595294eee..b5de0f40ea25 100644
> > --- a/drivers/ata/libata-sata.c
> > +++ b/drivers/ata/libata-sata.c
> > @@ -1177,10 +1177,6 @@ EXPORT_SYMBOL_GPL(ata_sas_sync_probe);
> > int ata_sas_port_init(struct ata_port *ap)
>
> This is a bit of a daft function now, considering it only does
> atomic_inc_return(&ata_print_id). Do we really need to export a symbol for
> that?
$ git grep ata_print_id
drivers/ata/libata-core.c:atomic_t ata_print_id = ATOMIC_INIT(0);
drivers/ata/libata-core.c: host->ports[i]->print_id = atomic_inc_return(&ata_print_id);
drivers/ata/libata-sata.c: ap->print_id = atomic_inc_return(&ata_print_id);
drivers/ata/libata.h:extern atomic_t ata_print_id;
It seems to be defined and used only in libata, while I agree that the function
is a bit silly, with my limited knowledge of how the linker works, moving it to
libsas seems a bit dangerous...
You can build libata as a module and libsas as built-in, and vice versa...
Also, since there are no direct users in libsas, I'd rather keep it in libata.
>
> > {
> > - int rc = ap->ops->port_start(ap);
>
> I am not sure how this change is really relevant to " Is
> (ata_sas_port_destroy()) now a wrapper around kfree(), so call it directly."
Agreed, I will move it to the previous patch, so we also avoid the null
pointer dereference in the previous patch.
>
> > -
> > - if (rc)
> > - return rc;
> > ap->print_id = atomic_inc_return(&ata_print_id);
> > return 0;
>
> always returns 0, so pretty pointless to return a value at all
Yes, that would be a small optimization, but I would consider such
a change outside the scope of this series.
Kind regards,
Niklas
>
> > }
> > @@ -1198,20 +1194,6 @@ void ata_sas_tport_delete(struct ata_port *ap)
> > }
> > EXPORT_SYMBOL_GPL(ata_sas_tport_delete);
> > -/**
> > - * ata_sas_port_destroy - Destroy a SATA port allocated by ata_sas_port_alloc
> > - * @ap: SATA port to destroy
> > - *
> > - */
> > -
> > -void ata_sas_port_destroy(struct ata_port *ap)
> > -{
> > - if (ap->ops->port_stop)
> > - ap->ops->port_stop(ap);
> > - kfree(ap);
> > -}
> > -EXPORT_SYMBOL_GPL(ata_sas_port_destroy);
> > -
> > /**
> > * ata_sas_slave_configure - Default slave_config routine for libata devices
> > * @sdev: SCSI device to configure
> > diff --git a/drivers/scsi/libsas/sas_ata.c b/drivers/scsi/libsas/sas_ata.c
> > index 7ead1f1be97f..a2eb9a2191c0 100644
> > --- a/drivers/scsi/libsas/sas_ata.c
> > +++ b/drivers/scsi/libsas/sas_ata.c
> > @@ -619,7 +619,7 @@ int sas_ata_init(struct domain_device *found_dev)
> > return 0;
> > destroy_port:
> > - ata_sas_port_destroy(ap);
> > + kfree(ap);
> > free_host:
> > ata_host_put(ata_host);
> > return rc;
> > diff --git a/drivers/scsi/libsas/sas_discover.c b/drivers/scsi/libsas/sas_discover.c
> > index 8c6afe724944..07e18cdb85c7 100644
> > --- a/drivers/scsi/libsas/sas_discover.c
> > +++ b/drivers/scsi/libsas/sas_discover.c
> > @@ -301,7 +301,7 @@ void sas_free_device(struct kref *kref)
> > if (dev_is_sata(dev) && dev->sata_dev.ap) {
> > ata_sas_tport_delete(dev->sata_dev.ap);
> > - ata_sas_port_destroy(dev->sata_dev.ap);
> > + kfree(dev->sata_dev.ap);
> > ata_host_put(dev->sata_dev.ata_host);
> > dev->sata_dev.ata_host = NULL;
> > dev->sata_dev.ap = NULL;
> > diff --git a/include/linux/libata.h b/include/linux/libata.h
> > index 9424c490ef0b..53cfb1a4b97a 100644
> > --- a/include/linux/libata.h
> > +++ b/include/linux/libata.h
> > @@ -1238,7 +1238,6 @@ extern int sata_link_debounce(struct ata_link *link,
> > extern int sata_link_scr_lpm(struct ata_link *link, enum ata_lpm_policy policy,
> > bool spm_wakeup);
> > extern int ata_slave_link_init(struct ata_port *ap);
> > -extern void ata_sas_port_destroy(struct ata_port *);
> > extern struct ata_port *ata_sas_port_alloc(struct ata_host *,
> > struct ata_port_info *, struct Scsi_Host *);
> > extern void ata_sas_async_probe(struct ata_port *ap);
>
