| Message ID | 20230321133609.49591-2-nbd@nbd.name (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net-next,1/2] net: ethernet: mtk_eth_soc: add code for offloading flows from wlan devices | expand |
On Tue, Mar 21, 2023 at 02:36:09PM +0100, Felix Fietkau wrote: > When a device is roaming between interfaces and a new flow entry is created, > we should assume that its output device is more up to date than whatever > entry existed already. As per patch 1/2. checkpatch complains that the patch description has lines more than 75 characters long. That aside, this change looks good to me. But I'm wondering if it is fixing a bug. Or just improving something suboptimal (form a user experience POV). > > Signed-off-by: Felix Fietkau <nbd@nbd.name> > --- > drivers/net/ethernet/mediatek/mtk_ppe.c | 14 ++++++++++++-- > 1 file changed, 12 insertions(+), 2 deletions(-) > > diff --git a/drivers/net/ethernet/mediatek/mtk_ppe.c b/drivers/net/ethernet/mediatek/mtk_ppe.c > index 6883eb34cd8b..5e150007bb7d 100644 > --- a/drivers/net/ethernet/mediatek/mtk_ppe.c > +++ b/drivers/net/ethernet/mediatek/mtk_ppe.c > @@ -580,10 +580,20 @@ void mtk_foe_entry_clear(struct mtk_ppe *ppe, struct mtk_flow_entry *entry) > static int > mtk_foe_entry_commit_l2(struct mtk_ppe *ppe, struct mtk_flow_entry *entry) > { > + struct mtk_flow_entry *prev; > + > entry->type = MTK_FLOW_TYPE_L2; > > - return rhashtable_insert_fast(&ppe->l2_flows, &entry->l2_node, > - mtk_flow_l2_ht_params); > + prev = rhashtable_lookup_get_insert_fast(&ppe->l2_flows, &entry->l2_node, > + mtk_flow_l2_ht_params); > + if (likely(!prev)) > + return 0; > + > + if (IS_ERR(prev)) > + return PTR_ERR(prev); > + > + return rhashtable_replace_fast(&ppe->l2_flows, &prev->l2_node, > + &entry->l2_node, mtk_flow_l2_ht_params); > } > > int mtk_foe_entry_commit(struct mtk_ppe *ppe, struct mtk_flow_entry *entry) > -- > 2.39.0 >
On 22.03.23 15:00, Simon Horman wrote: > On Tue, Mar 21, 2023 at 02:36:09PM +0100, Felix Fietkau wrote: >> When a device is roaming between interfaces and a new flow entry is created, >> we should assume that its output device is more up to date than whatever >> entry existed already. > > As per patch 1/2. checkpatch complains that the patch description > has lines more than 75 characters long. Will do. > That aside, this change looks good to me. > But I'm wondering if it is fixing a bug. > Or just improving something suboptimal (form a user experience POV). In my opinion, this is just an improvement, not a bugfix. Thanks, - Felix
On Wed, Mar 22, 2023 at 04:09:25PM +0100, Felix Fietkau wrote: > On 22.03.23 15:00, Simon Horman wrote: > > On Tue, Mar 21, 2023 at 02:36:09PM +0100, Felix Fietkau wrote: > > > When a device is roaming between interfaces and a new flow entry is created, > > > we should assume that its output device is more up to date than whatever > > > entry existed already. > > > > As per patch 1/2. checkpatch complains that the patch description > > has lines more than 75 characters long. > Will do. > > > That aside, this change looks good to me. > > But I'm wondering if it is fixing a bug. > > Or just improving something suboptimal (form a user experience POV). > In my opinion, this is just an improvement, not a bugfix. Sure, thanks for the clarification.
diff --git a/drivers/net/ethernet/mediatek/mtk_ppe.c b/drivers/net/ethernet/mediatek/mtk_ppe.c index 6883eb34cd8b..5e150007bb7d 100644 --- a/drivers/net/ethernet/mediatek/mtk_ppe.c +++ b/drivers/net/ethernet/mediatek/mtk_ppe.c @@ -580,10 +580,20 @@ void mtk_foe_entry_clear(struct mtk_ppe *ppe, struct mtk_flow_entry *entry) static int mtk_foe_entry_commit_l2(struct mtk_ppe *ppe, struct mtk_flow_entry *entry) { + struct mtk_flow_entry *prev; + entry->type = MTK_FLOW_TYPE_L2; - return rhashtable_insert_fast(&ppe->l2_flows, &entry->l2_node, - mtk_flow_l2_ht_params); + prev = rhashtable_lookup_get_insert_fast(&ppe->l2_flows, &entry->l2_node, + mtk_flow_l2_ht_params); + if (likely(!prev)) + return 0; + + if (IS_ERR(prev)) + return PTR_ERR(prev); + + return rhashtable_replace_fast(&ppe->l2_flows, &prev->l2_node, + &entry->l2_node, mtk_flow_l2_ht_params); } int mtk_foe_entry_commit(struct mtk_ppe *ppe, struct mtk_flow_entry *entry)
When a device is roaming between interfaces and a new flow entry is created, we should assume that its output device is more up to date than whatever entry existed already. Signed-off-by: Felix Fietkau <nbd@nbd.name> --- drivers/net/ethernet/mediatek/mtk_ppe.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-)