| Message ID | 20230613133854.37832-1-ihuguet@redhat.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | e84a1e1e683f3558e30f437d7c99df35afb8b52c |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net] sfc: fix XDP queues mode with legacy IRQ | expand |
On Tue, Jun 13, 2023 at 03:38:54PM +0200, Íñigo Huguet wrote: > In systems without MSI-X capabilities, xdp_txq_queues_mode is calculated > in efx_allocate_msix_channels, but when enabling MSI-X fails, it was not > changed to a proper default value. This was leading to the driver > thinking that it has dedicated XDP queues, when it didn't. > > Fix it by setting xdp_txq_queues_mode to the correct value if the driver > fallbacks to MSI or legacy IRQ mode. The correct value is > EFX_XDP_TX_QUEUES_BORROWED because there are not XDP dedicated queues. Small typo: not should be no. > > The issue can be easily visible if the kernel is started with pci=nomsi, > then a call trace is shown. It is not shown only with sfc's modparam > interrupt_mode=2. Call trace example: > WARNING: CPU: 2 PID: 663 at drivers/net/ethernet/sfc/efx_channels.c:828 efx_set_xdp_channels+0x124/0x260 [sfc] > [...skip...] > Call Trace: > <TASK> > efx_set_channels+0x5c/0xc0 [sfc] > efx_probe_nic+0x9b/0x15a [sfc] > efx_probe_all+0x10/0x1a2 [sfc] > efx_pci_probe_main+0x12/0x156 [sfc] > efx_pci_probe_post_io+0x18/0x103 [sfc] > efx_pci_probe.cold+0x154/0x257 [sfc] > local_pci_probe+0x42/0x80 > > Fixes: 6215b608a8c4 ("sfc: last resort fallback for lack of xdp tx queues") > Reported-by: Yanghang Liu <yanghliu@redhat.com> > Signed-off-by: Íñigo Huguet <ihuguet@redhat.com> Acked-by: Martin Habets <habetsm.xilinx@gmail.com> > --- > drivers/net/ethernet/sfc/efx_channels.c | 2 ++ > drivers/net/ethernet/sfc/siena/efx_channels.c | 2 ++ > 2 files changed, 4 insertions(+) > > diff --git a/drivers/net/ethernet/sfc/efx_channels.c b/drivers/net/ethernet/sfc/efx_channels.c > index fcea3ea809d7..41b33a75333c 100644 > --- a/drivers/net/ethernet/sfc/efx_channels.c > +++ b/drivers/net/ethernet/sfc/efx_channels.c > @@ -301,6 +301,7 @@ int efx_probe_interrupts(struct efx_nic *efx) > efx->tx_channel_offset = 0; > efx->n_xdp_channels = 0; > efx->xdp_channel_offset = efx->n_channels; > + efx->xdp_txq_queues_mode = EFX_XDP_TX_QUEUES_BORROWED; > rc = pci_enable_msi(efx->pci_dev); > if (rc == 0) { > efx_get_channel(efx, 0)->irq = efx->pci_dev->irq; > @@ -322,6 +323,7 @@ int efx_probe_interrupts(struct efx_nic *efx) > efx->tx_channel_offset = efx_separate_tx_channels ? 1 : 0; > efx->n_xdp_channels = 0; > efx->xdp_channel_offset = efx->n_channels; > + efx->xdp_txq_queues_mode = EFX_XDP_TX_QUEUES_BORROWED; > efx->legacy_irq = efx->pci_dev->irq; > } > > diff --git a/drivers/net/ethernet/sfc/siena/efx_channels.c b/drivers/net/ethernet/sfc/siena/efx_channels.c > index 06ed74994e36..1776f7f8a7a9 100644 > --- a/drivers/net/ethernet/sfc/siena/efx_channels.c > +++ b/drivers/net/ethernet/sfc/siena/efx_channels.c > @@ -302,6 +302,7 @@ int efx_siena_probe_interrupts(struct efx_nic *efx) > efx->tx_channel_offset = 0; > efx->n_xdp_channels = 0; > efx->xdp_channel_offset = efx->n_channels; > + efx->xdp_txq_queues_mode = EFX_XDP_TX_QUEUES_BORROWED; > rc = pci_enable_msi(efx->pci_dev); > if (rc == 0) { > efx_get_channel(efx, 0)->irq = efx->pci_dev->irq; > @@ -323,6 +324,7 @@ int efx_siena_probe_interrupts(struct efx_nic *efx) > efx->tx_channel_offset = efx_siena_separate_tx_channels ? 1 : 0; > efx->n_xdp_channels = 0; > efx->xdp_channel_offset = efx->n_channels; > + efx->xdp_txq_queues_mode = EFX_XDP_TX_QUEUES_BORROWED; > efx->legacy_irq = efx->pci_dev->irq; > } > > -- > 2.40.1
Hello: This patch was applied to netdev/net.git (main) by David S. Miller <davem@davemloft.net>: On Tue, 13 Jun 2023 15:38:54 +0200 you wrote: > In systems without MSI-X capabilities, xdp_txq_queues_mode is calculated > in efx_allocate_msix_channels, but when enabling MSI-X fails, it was not > changed to a proper default value. This was leading to the driver > thinking that it has dedicated XDP queues, when it didn't. > > Fix it by setting xdp_txq_queues_mode to the correct value if the driver > fallbacks to MSI or legacy IRQ mode. The correct value is > EFX_XDP_TX_QUEUES_BORROWED because there are not XDP dedicated queues. > > [...] Here is the summary with links: - [net] sfc: fix XDP queues mode with legacy IRQ https://git.kernel.org/netdev/net/c/e84a1e1e683f You are awesome, thank you!
diff --git a/drivers/net/ethernet/sfc/efx_channels.c b/drivers/net/ethernet/sfc/efx_channels.c index fcea3ea809d7..41b33a75333c 100644 --- a/drivers/net/ethernet/sfc/efx_channels.c +++ b/drivers/net/ethernet/sfc/efx_channels.c @@ -301,6 +301,7 @@ int efx_probe_interrupts(struct efx_nic *efx) efx->tx_channel_offset = 0; efx->n_xdp_channels = 0; efx->xdp_channel_offset = efx->n_channels; + efx->xdp_txq_queues_mode = EFX_XDP_TX_QUEUES_BORROWED; rc = pci_enable_msi(efx->pci_dev); if (rc == 0) { efx_get_channel(efx, 0)->irq = efx->pci_dev->irq; @@ -322,6 +323,7 @@ int efx_probe_interrupts(struct efx_nic *efx) efx->tx_channel_offset = efx_separate_tx_channels ? 1 : 0; efx->n_xdp_channels = 0; efx->xdp_channel_offset = efx->n_channels; + efx->xdp_txq_queues_mode = EFX_XDP_TX_QUEUES_BORROWED; efx->legacy_irq = efx->pci_dev->irq; } diff --git a/drivers/net/ethernet/sfc/siena/efx_channels.c b/drivers/net/ethernet/sfc/siena/efx_channels.c index 06ed74994e36..1776f7f8a7a9 100644 --- a/drivers/net/ethernet/sfc/siena/efx_channels.c +++ b/drivers/net/ethernet/sfc/siena/efx_channels.c @@ -302,6 +302,7 @@ int efx_siena_probe_interrupts(struct efx_nic *efx) efx->tx_channel_offset = 0; efx->n_xdp_channels = 0; efx->xdp_channel_offset = efx->n_channels; + efx->xdp_txq_queues_mode = EFX_XDP_TX_QUEUES_BORROWED; rc = pci_enable_msi(efx->pci_dev); if (rc == 0) { efx_get_channel(efx, 0)->irq = efx->pci_dev->irq; @@ -323,6 +324,7 @@ int efx_siena_probe_interrupts(struct efx_nic *efx) efx->tx_channel_offset = efx_siena_separate_tx_channels ? 1 : 0; efx->n_xdp_channels = 0; efx->xdp_channel_offset = efx->n_channels; + efx->xdp_txq_queues_mode = EFX_XDP_TX_QUEUES_BORROWED; efx->legacy_irq = efx->pci_dev->irq; }
In systems without MSI-X capabilities, xdp_txq_queues_mode is calculated in efx_allocate_msix_channels, but when enabling MSI-X fails, it was not changed to a proper default value. This was leading to the driver thinking that it has dedicated XDP queues, when it didn't. Fix it by setting xdp_txq_queues_mode to the correct value if the driver fallbacks to MSI or legacy IRQ mode. The correct value is EFX_XDP_TX_QUEUES_BORROWED because there are not XDP dedicated queues. The issue can be easily visible if the kernel is started with pci=nomsi, then a call trace is shown. It is not shown only with sfc's modparam interrupt_mode=2. Call trace example: WARNING: CPU: 2 PID: 663 at drivers/net/ethernet/sfc/efx_channels.c:828 efx_set_xdp_channels+0x124/0x260 [sfc] [...skip...] Call Trace: <TASK> efx_set_channels+0x5c/0xc0 [sfc] efx_probe_nic+0x9b/0x15a [sfc] efx_probe_all+0x10/0x1a2 [sfc] efx_pci_probe_main+0x12/0x156 [sfc] efx_pci_probe_post_io+0x18/0x103 [sfc] efx_pci_probe.cold+0x154/0x257 [sfc] local_pci_probe+0x42/0x80 Fixes: 6215b608a8c4 ("sfc: last resort fallback for lack of xdp tx queues") Reported-by: Yanghang Liu <yanghliu@redhat.com> Signed-off-by: Íñigo Huguet <ihuguet@redhat.com> --- drivers/net/ethernet/sfc/efx_channels.c | 2 ++ drivers/net/ethernet/sfc/siena/efx_channels.c | 2 ++ 2 files changed, 4 insertions(+)