> -----Original Message-----
> From: Wei Fang <wei.fang@xxxxxxx>
> Sent: Thursday, February 20, 2025 7:07 AM
[...]
> Subject: RE: [PATCH v2 net 9/9] net: enetc: fix the off-by-one issue in
> enetc_map_tx_tso_buffs()
>
> > > -----Original Message-----
> > > From: Wei Fang <wei.fang@xxxxxxx>
> > > Sent: Wednesday, February 19, 2025 7:43 AM
> > [...]
> > > Subject: [PATCH v2 net 9/9] net: enetc: fix the off-by-one issue in
> > > enetc_map_tx_tso_buffs()
> > >
> > > There is an off-by-one issue for the err_chained_bd path, it will
> > > free one more tx_swbd than expected. But there is no such issue for
> > > the err_map_data path. To fix this off-by-one issue and make the two
> > > error handling consistent, the loop condition of error handling is
> > > modified and the 'count++' operation is moved before
> enetc_map_tx_tso_data().
> > >
> > > Fixes: fb8629e2cbfc ("net: enetc: add support for software TSO")
> > > Cc: stable@xxxxxxxxxxxxxxx
> > > Signed-off-by: Wei Fang <wei.fang@xxxxxxx>
Reviewed-by: Claudiu Manoil <claudiu.manoil@xxxxxxx>
> > > ---
> > > drivers/net/ethernet/freescale/enetc/enetc.c | 6 +++---
> > > 1 file changed, 3 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c
> > > b/drivers/net/ethernet/freescale/enetc/enetc.c
> > > index 9a24d1176479..fe3967268a19 100644
> > > --- a/drivers/net/ethernet/freescale/enetc/enetc.c
> > > +++ b/drivers/net/ethernet/freescale/enetc/enetc.c
> > > @@ -832,6 +832,7 @@ static int enetc_map_tx_tso_buffs(struct
> > > enetc_bdr *tx_ring, struct sk_buff *skb
> > > txbd = ENETC_TXBD(*tx_ring, i);
> > > tx_swbd = &tx_ring->tx_swbd[i];
> > > prefetchw(txbd);
> > > + count++;
> > >
> > > /* Compute the checksum over this segment of data
> and
> > > * add it to the csum already computed (over the L4
> @@ -848,7
> > > +849,6 @@ static int enetc_map_tx_tso_buffs(struct enetc_bdr
> > > *tx_ring, struct sk_buff *skb
> > > goto err_map_data;
> > >
> > > data_len -= size;
> > > - count++;
> >
> > Hi Wei,
> >
> > My issue is that:
> > enetc_map_tx_tso_hdr() not only updates the current tx_swbd (so 1
> > count++ needed), but in case of extension flag it advances 'tx_swbd'
> > and 'i' with another position so and extra 'count++' would be needed
> > in that case.
> >
>
> I think the patch 2 (net: enetc: correct the tx_swbd statistics) is to resolve your
> issue.
Ok, I missed that. Thanks.
