| Message ID | 20220715040949.3807070-1-quic_kshivnan@quicinc.com (mailing list archive) |
|---|---|
| State | Superseded, archived |
| Headers | show |
| Series | [v3] PM: QoS: Add check to make sure CPU freq is non-negative | expand |
On Fri, Jul 15, 2022 at 6:10 AM Shivnandan Kumar <quic_kshivnan@quicinc.com> wrote: > > CPU frequency should never be negative. > If some client driver calls freq_qos_update_request with > negative value which will be very high in absolute terms, > then qos driver set max CPU freq at fmax as it consider > it's absolute value but it will add plist node with negative > priority. plist node has priority from INT_MIN (highest) > to INT_MAX(lowest). Once priority is set as negative, > another client will not be able to reduce CPU frequency. > Adding check to make sure CPU freq is non-negative will > fix this problem. > > Signed-off-by: Shivnandan Kumar <quic_kshivnan@quicinc.com> > --- > v2->v3 > -changed commit text > v1->v2 > -addressed comments from Rafael > -changed commit text accordingly > kernel/power/qos.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/kernel/power/qos.c b/kernel/power/qos.c > index ec7e1e85923e..27e6596f287a 100644 > --- a/kernel/power/qos.c > +++ b/kernel/power/qos.c > @@ -531,7 +531,7 @@ int freq_qos_add_request(struct freq_constraints *qos, > { > int ret; > > - if (IS_ERR_OR_NULL(qos) || !req) > + if (IS_ERR_OR_NULL(qos) || !req || value < FREQ_QOS_MIN_DEFAULT_VALUE) Why not just put 0 in there directly instead of FREQ_QOS_MIN_DEFAULT_VALUE? As is, it is somewhat confusing (and same below). > return -EINVAL; > > if (WARN(freq_qos_request_active(req), > @@ -563,7 +563,7 @@ EXPORT_SYMBOL_GPL(freq_qos_add_request); > */ > int freq_qos_update_request(struct freq_qos_request *req, s32 new_value) > { > - if (!req) > + if (!req || new_value < FREQ_QOS_MIN_DEFAULT_VALUE) > return -EINVAL; > > if (WARN(!freq_qos_request_active(req), > -- > 2.25.1 >
diff --git a/kernel/power/qos.c b/kernel/power/qos.c index ec7e1e85923e..27e6596f287a 100644 --- a/kernel/power/qos.c +++ b/kernel/power/qos.c @@ -531,7 +531,7 @@ int freq_qos_add_request(struct freq_constraints *qos, { int ret; - if (IS_ERR_OR_NULL(qos) || !req) + if (IS_ERR_OR_NULL(qos) || !req || value < FREQ_QOS_MIN_DEFAULT_VALUE) return -EINVAL; if (WARN(freq_qos_request_active(req), @@ -563,7 +563,7 @@ EXPORT_SYMBOL_GPL(freq_qos_add_request); */ int freq_qos_update_request(struct freq_qos_request *req, s32 new_value) { - if (!req) + if (!req || new_value < FREQ_QOS_MIN_DEFAULT_VALUE) return -EINVAL; if (WARN(!freq_qos_request_active(req),
CPU frequency should never be negative. If some client driver calls freq_qos_update_request with negative value which will be very high in absolute terms, then qos driver set max CPU freq at fmax as it consider it's absolute value but it will add plist node with negative priority. plist node has priority from INT_MIN (highest) to INT_MAX(lowest). Once priority is set as negative, another client will not be able to reduce CPU frequency. Adding check to make sure CPU freq is non-negative will fix this problem. Signed-off-by: Shivnandan Kumar <quic_kshivnan@quicinc.com> --- v2->v3 -changed commit text v1->v2 -addressed comments from Rafael -changed commit text accordingly kernel/power/qos.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)