| Message ID | 20231004083454.20143-1-tirthendu.sarkar@intel.com (mailing list archive) |
|---|---|
| State | Awaiting Upstream |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net] i40e: sync next_to_clean and next_to_process for programming status desc | expand |
On Wed, Oct 04, 2023 at 02:04:54PM +0530, Tirthendu Sarkar wrote: > When a programming status desc is encountered on the rx_ring, > next_to_process is bumped along with cleaned_count but next_to_clean is > not. This causes I40E_DESC_UNUSED() macro to misbehave resulting in > overwriting whole ring with new buffers. > > Update next_to_clean to point to next_to_process on seeing a programming > status desc if not in the middle of handling a multi-frag packet. Also, > bump cleaned_count only for such case as otherwise next_to_clean buffer > may be returned to hardware on reaching clean_threshold. > > Fixes: e9031f2da1ae ("i40e: introduce next_to_process to i40e_ring") > Suggested-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com> > Reported-by: hq.dev+kernel@msdfc.xyz > Reported by: Solomon Peachy <pizza@shaftnet.org> > Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217678 > Tested-by: hq.dev+kernel@msdfc.xyz Could you ask for a name of that someone? > Tested by: Indrek Järve <incx@dustbite.net> > Signed-off-by: Tirthendu Sarkar <tirthendu.sarkar@intel.com> Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com> > --- > drivers/net/ethernet/intel/i40e/i40e_txrx.c | 9 ++++++++- > 1 file changed, 8 insertions(+), 1 deletion(-) > > diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c > index 0b3a27f118fb..50c70a8e470a 100644 > --- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c > +++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c > @@ -2544,7 +2544,14 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget, > rx_buffer = i40e_rx_bi(rx_ring, ntp); > i40e_inc_ntp(rx_ring); > i40e_reuse_rx_page(rx_ring, rx_buffer); > - cleaned_count++; > + /* Update ntc and bump cleaned count if not in the > + * middle of mb packet. > + */ > + if (rx_ring->next_to_clean == ntp) { > + rx_ring->next_to_clean = > + rx_ring->next_to_process; > + cleaned_count++; > + } > continue; > } > > -- > 2.34.1 >
> -----Original Message----- > From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of Tirthendu Sarkar > Sent: Wednesday, October 4, 2023 2:05 PM > To: intel-wired-lan@lists.osuosl.org > Cc: Fijalkowski, Maciej <maciej.fijalkowski@intel.com>; netdev@vger.kernel.org; Brandeburg, Jesse <jesse.brandeburg@intel.com>; Nguyen, Anthony L <anthony.l.nguyen@intel.com>; bpf@vger.kernel.org; Karlsson, Magnus <magnus.karlsson@intel.com> > Subject: [Intel-wired-lan] [PATCH net] i40e: sync next_to_clean and next_to_process for programming status desc > > When a programming status desc is encountered on the rx_ring, next_to_process is bumped along with cleaned_count but next_to_clean is not. This causes I40E_DESC_UNUSED() macro to misbehave resulting in overwriting whole ring with new buffers. > > Update next_to_clean to point to next_to_process on seeing a programming status desc if not in the middle of handling a multi-frag packet. Also, bump cleaned_count only for such case as otherwise next_to_clean buffer may be returned to hardware on reaching clean_threshold. > > Fixes: e9031f2da1ae ("i40e: introduce next_to_process to i40e_ring") > Suggested-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com> > Reported-by: hq.dev+kernel@msdfc.xyz > Reported by: Solomon Peachy <pizza@shaftnet.org> > Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217678 > Tested-by: hq.dev+kernel@msdfc.xyz > Tested by: Indrek Järve <incx@dustbite.net> > Signed-off-by: Tirthendu Sarkar <tirthendu.sarkar@intel.com> > --- > drivers/net/ethernet/intel/i40e/i40e_txrx.c | 9 ++++++++- > 1 file changed, 8 insertions(+), 1 deletion(-) > Tested-by: Arpana Arland <arpanax.arland@intel.com> (A Contingent worker at Intel)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c index 0b3a27f118fb..50c70a8e470a 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c +++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c @@ -2544,7 +2544,14 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget, rx_buffer = i40e_rx_bi(rx_ring, ntp); i40e_inc_ntp(rx_ring); i40e_reuse_rx_page(rx_ring, rx_buffer); - cleaned_count++; + /* Update ntc and bump cleaned count if not in the + * middle of mb packet. + */ + if (rx_ring->next_to_clean == ntp) { + rx_ring->next_to_clean = + rx_ring->next_to_process; + cleaned_count++; + } continue; }
When a programming status desc is encountered on the rx_ring, next_to_process is bumped along with cleaned_count but next_to_clean is not. This causes I40E_DESC_UNUSED() macro to misbehave resulting in overwriting whole ring with new buffers. Update next_to_clean to point to next_to_process on seeing a programming status desc if not in the middle of handling a multi-frag packet. Also, bump cleaned_count only for such case as otherwise next_to_clean buffer may be returned to hardware on reaching clean_threshold. Fixes: e9031f2da1ae ("i40e: introduce next_to_process to i40e_ring") Suggested-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com> Reported-by: hq.dev+kernel@msdfc.xyz Reported by: Solomon Peachy <pizza@shaftnet.org> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217678 Tested-by: hq.dev+kernel@msdfc.xyz Tested by: Indrek Järve <incx@dustbite.net> Signed-off-by: Tirthendu Sarkar <tirthendu.sarkar@intel.com> --- drivers/net/ethernet/intel/i40e/i40e_txrx.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)