| Message ID | 20190905011144.3513-1-kevin.laatz@intel.com |
|---|---|
| State | Awaiting Upstream |
| Headers | show |
| Series | [bpf-next] i40e: fix xdp handle calculations | expand |
On Thu, 5 Sep 2019 at 11:27, Kevin Laatz <kevin.laatz@intel.com> wrote: > > Currently, we don't add headroom to the handle in i40e_zca_free, > i40e_alloc_buffer_slow_zc and i40e_alloc_buffer_zc. The addition of the > headroom to the handle was removed in > commit 2f86c806a8a8 ("i40e: modify driver for handling offsets"), which > will break things when headroom is non-zero. This patch fixes this and uses > xsk_umem_adjust_offset to add it appropritely based on the mode being run. > > Fixes: 2f86c806a8a8 ("i40e: modify driver for handling offsets") > Reported-by: Bjorn Topel <bjorn.topel@intel.com> > Signed-off-by: Kevin Laatz <kevin.laatz@intel.com> Thanks Kevin! Acked-by: Björn Töpel <bjorn.topel@intel.com> > --- > drivers/net/ethernet/intel/i40e/i40e_xsk.c | 7 ++++--- > 1 file changed, 4 insertions(+), 3 deletions(-) > > diff --git a/drivers/net/ethernet/intel/i40e/i40e_xsk.c b/drivers/net/ethernet/intel/i40e/i40e_xsk.c > index eaca6162a6e6..0373bc6c7e61 100644 > --- a/drivers/net/ethernet/intel/i40e/i40e_xsk.c > +++ b/drivers/net/ethernet/intel/i40e/i40e_xsk.c > @@ -267,7 +267,7 @@ static bool i40e_alloc_buffer_zc(struct i40e_ring *rx_ring, > bi->addr = xdp_umem_get_data(umem, handle); > bi->addr += hr; > > - bi->handle = handle; > + bi->handle = xsk_umem_adjust_offset(umem, handle, umem->headroom); > > xsk_umem_discard_addr(umem); > return true; > @@ -304,7 +304,7 @@ static bool i40e_alloc_buffer_slow_zc(struct i40e_ring *rx_ring, > bi->addr = xdp_umem_get_data(umem, handle); > bi->addr += hr; > > - bi->handle = handle; > + bi->handle = xsk_umem_adjust_offset(umem, handle, umem->headroom); > > xsk_umem_discard_addr_rq(umem); > return true; > @@ -469,7 +469,8 @@ void i40e_zca_free(struct zero_copy_allocator *alloc, unsigned long handle) > bi->addr = xdp_umem_get_data(rx_ring->xsk_umem, handle); > bi->addr += hr; > > - bi->handle = (u64)handle; > + bi->handle = xsk_umem_adjust_offset(rx_ring->xsk_umem, (u64)handle, > + rx_ring->xsk_umem->headroom); > } > > /** > -- > 2.17.1 > > _______________________________________________ > Intel-wired-lan mailing list > Intel-wired-lan@osuosl.org > https://lists.osuosl.org/mailman/listinfo/intel-wired-lan
On 9/5/19 3:11 AM, Kevin Laatz wrote: > Currently, we don't add headroom to the handle in i40e_zca_free, > i40e_alloc_buffer_slow_zc and i40e_alloc_buffer_zc. The addition of the > headroom to the handle was removed in > commit 2f86c806a8a8 ("i40e: modify driver for handling offsets"), which > will break things when headroom is non-zero. This patch fixes this and uses > xsk_umem_adjust_offset to add it appropritely based on the mode being run. > > Fixes: 2f86c806a8a8 ("i40e: modify driver for handling offsets") > Reported-by: Bjorn Topel <bjorn.topel@intel.com> > Signed-off-by: Kevin Laatz <kevin.laatz@intel.com> Applied, thanks!
On Thu, Sep 5, 2019 at 4:11 AM Kevin Laatz <kevin.laatz@intel.com> wrote: > > Currently, we don't add headroom to the handle in i40e_zca_free, > i40e_alloc_buffer_slow_zc and i40e_alloc_buffer_zc. The addition of the > headroom to the handle was removed in > commit 2f86c806a8a8 ("i40e: modify driver for handling offsets"), Hi, it looks to me that headroom is still broken after this commit. i40e_run_xdp_zc adds it a second time, i.e.: 1. xdp->handle already has the headroom added (after this patch). 2. bpf_prog_run_xdp(...); 3. u64 offset = umem->headroom; offset += xdp->data - xdp->data_hard_start; xdp->handle = xsk_umem_adjust_offset(umem, xdp->handle, offset); Step 3 adds the headroom one extra time. I didn't look at ixgbe, it might also need to be fixed. > which > will break things when headroom is non-zero. This patch fixes this and uses > xsk_umem_adjust_offset to add it appropritely based on the mode being run. > > Fixes: 2f86c806a8a8 ("i40e: modify driver for handling offsets") > Reported-by: Bjorn Topel <bjorn.topel@intel.com> > Signed-off-by: Kevin Laatz <kevin.laatz@intel.com> > Acked-by: Björn Töpel <bjorn.topel@intel.com> > --- > drivers/net/ethernet/intel/i40e/i40e_xsk.c | 7 ++++--- > 1 file changed, 4 insertions(+), 3 deletions(-) > > diff --git a/drivers/net/ethernet/intel/i40e/i40e_xsk.c b/drivers/net/ethernet/intel/i40e/i40e_xsk.c > index eaca6162a6e6..0373bc6c7e61 100644 > --- a/drivers/net/ethernet/intel/i40e/i40e_xsk.c > +++ b/drivers/net/ethernet/intel/i40e/i40e_xsk.c > @@ -267,7 +267,7 @@ static bool i40e_alloc_buffer_zc(struct i40e_ring *rx_ring, > bi->addr = xdp_umem_get_data(umem, handle); > bi->addr += hr; > > - bi->handle = handle; > + bi->handle = xsk_umem_adjust_offset(umem, handle, umem->headroom); > > xsk_umem_discard_addr(umem); > return true; > @@ -304,7 +304,7 @@ static bool i40e_alloc_buffer_slow_zc(struct i40e_ring *rx_ring, > bi->addr = xdp_umem_get_data(umem, handle); > bi->addr += hr; > > - bi->handle = handle; > + bi->handle = xsk_umem_adjust_offset(umem, handle, umem->headroom); > > xsk_umem_discard_addr_rq(umem); > return true; > @@ -469,7 +469,8 @@ void i40e_zca_free(struct zero_copy_allocator *alloc, unsigned long handle) > bi->addr = xdp_umem_get_data(rx_ring->xsk_umem, handle); > bi->addr += hr; > > - bi->handle = (u64)handle; > + bi->handle = xsk_umem_adjust_offset(rx_ring->xsk_umem, (u64)handle, > + rx_ring->xsk_umem->headroom); > } > > /**
On 2019-09-17 12:48, Maxim Mikityanskiy wrote: > Hi, it looks to me that headroom is still broken after this commit. > i40e_run_xdp_zc adds it a second time, i.e.: Indeed, but Ciara fixed that in this series [1]. Thanks for reviewing the patch! Cheers, Björn [1] https://lore.kernel.org/bpf/20190913103948.32053-1-ciara.loftus@intel.com/
diff --git a/drivers/net/ethernet/intel/i40e/i40e_xsk.c b/drivers/net/ethernet/intel/i40e/i40e_xsk.c index eaca6162a6e6..0373bc6c7e61 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_xsk.c +++ b/drivers/net/ethernet/intel/i40e/i40e_xsk.c @@ -267,7 +267,7 @@ static bool i40e_alloc_buffer_zc(struct i40e_ring *rx_ring, bi->addr = xdp_umem_get_data(umem, handle); bi->addr += hr; - bi->handle = handle; + bi->handle = xsk_umem_adjust_offset(umem, handle, umem->headroom); xsk_umem_discard_addr(umem); return true; @@ -304,7 +304,7 @@ static bool i40e_alloc_buffer_slow_zc(struct i40e_ring *rx_ring, bi->addr = xdp_umem_get_data(umem, handle); bi->addr += hr; - bi->handle = handle; + bi->handle = xsk_umem_adjust_offset(umem, handle, umem->headroom); xsk_umem_discard_addr_rq(umem); return true; @@ -469,7 +469,8 @@ void i40e_zca_free(struct zero_copy_allocator *alloc, unsigned long handle) bi->addr = xdp_umem_get_data(rx_ring->xsk_umem, handle); bi->addr += hr; - bi->handle = (u64)handle; + bi->handle = xsk_umem_adjust_offset(rx_ring->xsk_umem, (u64)handle, + rx_ring->xsk_umem->headroom); } /**
Currently, we don't add headroom to the handle in i40e_zca_free, i40e_alloc_buffer_slow_zc and i40e_alloc_buffer_zc. The addition of the headroom to the handle was removed in commit 2f86c806a8a8 ("i40e: modify driver for handling offsets"), which will break things when headroom is non-zero. This patch fixes this and uses xsk_umem_adjust_offset to add it appropritely based on the mode being run. Fixes: 2f86c806a8a8 ("i40e: modify driver for handling offsets") Reported-by: Bjorn Topel <bjorn.topel@intel.com> Signed-off-by: Kevin Laatz <kevin.laatz@intel.com> --- drivers/net/ethernet/intel/i40e/i40e_xsk.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)