eth: bnxt: always recalculate features after XDP clearing, fix null-deref
Jakub Kicinski
maintainerPATCH 1/1 eth: bnxt: always recalculate features after XDP clearing, fix null-deref
Recalculate features when XDP is detached.
Before:
# ip li set dev eth0 xdp obj xdp_dummy.bpf.o sec xdp # ip li set dev eth0 xdp off # ethtool -k eth0 | grep gro rx-gro-hw: off [requested on]
After:
# ip li set dev eth0 xdp obj xdp_dummy.bpf.o sec xdp # ip li set dev eth0 xdp off # ethtool -k eth0 | grep gro rx-gro-hw: on
The fact that HW-GRO doesn't get re-enabled automatically is just a minor annoyance. The real issue is that the features will randomly come back during another reconfiguration which just happens to invoke netdev_update_features(). The driver doesn't handle reconfiguring two things at a time very robustly.
Starting with commit 98ba1d931f61 ("bnxt_en: Fix RSS logic in __bnxt_reserve_rings()") we only reconfigure the RSS hash table if the "effective" number of Rx rings has changed. If HW-GRO is enabled "effective" number of rings is 2x what user sees. So if we are in the bad state, with HW-GRO re-enablement "pending" after XDP off, and we lower the rings by / 2 - the HW-GRO rings doing 2x and the ethtool -L doing / 2 may cancel each other out, and the:
if (old_rx_rings != bp->hw_resc.resv_rx_rings &&
condition in __bnxt_reserve_rings() will be false. The RSS map won't get updated, and we'll crash with:
BUG: kernel NULL pointer dereference, address: 0000000000000168
RIP: 0010:__bnxt_hwrm_vnic_set_rss+0x13a/0x1a0
bnxt_hwrm_vnic_rss_cfg_p5+0x47/0x180
__bnxt_setup_vnic_p5+0x58/0x110
bnxt_init_nic+0xb72/0xf50
__bnxt_open_nic+0x40d/0xab0
bnxt_open_nic+0x2b/0x60
ethtool_set_channels+0x18c/0x1d0As we try to access a freed ring.
The issue is present since XDP support was added, really, but prior to commit 98ba1d931f61 ("bnxt_en: Fix RSS logic in __bnxt_reserve_rings()") it wasn't causing major issues.
---
CC: [email protected]
CC: [email protected]
CC: [email protected]
CC: [email protected]
CC: [email protected]
CC: [email protected]
--- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 25 +++++++++++++++---- drivers/net/ethernet/broadcom/bnxt/bnxt.h | 2 +- drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c | 7 ------ 3 files changed, 21 insertions(+), 13 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +20 -5
@@ -4708,7 +4708,7 @@ void bnxt_set_ring_params(struct bnxt *bp)/* Changing allocation mode of RX rings. * TODO: Update when extending xdp_rxq_info to support allocation modes. */ -int bnxt_set_rx_skb_mode (struct bnxt *bp,bool page_mode) +static void __bnxt_set_rx_skb_mode (struct bnxt *bp,bool page_mode) {struct net_device *dev = bp->dev;
@@ -4729,15 +4729,30 @@ int bnxt_set_rx_skb_mode(struct bnxt *bp, bool page_mode) bp->rx_skb_func = bnxt_rx_page_skb; } bp->rx_dir =DMA_BIDIRECTIONAL ; -/* Disable LRO or GRO_HW */ -netdev_update_features (dev); }else { dev->max_mtu = bp->max_mtu; bp->flags &= ~BNXT_FLAG_RX_PAGE_MODE ; bp->rx_dir =DMA_FROM_DEVICE ; bp->rx_skb_func = bnxt_rx_skb; } -return 0 ; +} + +void bnxt_set_rx_skb_mode (struct bnxt *bp,bool page_mode) +{ +__bnxt_set_rx_skb_mode (bp, page_mode); + +if (!page_mode) { +int rx, tx; + +bnxt_get_max_rings (bp, &rx, &tx, true); +if (rx >1 ) { + bp->flags &= ~BNXT_FLAG_NO_AGG_RINGS ; + bp->dev->hw_features |=NETIF_F_LRO ; + } + } + +/* Update LRO and GRO_HW availability */ +netdev_update_features (bp->dev); }static void bnxt_free_vnic_attributes (struct bnxt *bp)
@@ -16214,7 +16229,7 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)if (bp->max_fltr <BNXT_MAX_FLTR ) bp->max_fltr =BNXT_MAX_FLTR ;bnxt_init_l2_fltr_tbl (bp); -bnxt_set_rx_skb_mode (bp, false); +__bnxt_set_rx_skb_mode (bp, false);bnxt_set_tpa_flags (bp);bnxt_set_ring_params (bp);bnxt_rdma_aux_device_init (bp);
drivers/net/ethernet/broadcom/bnxt/bnxt.h +1 -1
@@ -2846,7 +2846,7 @@ u32 bnxt_fw_health_readl(struct bnxt *bp, int reg_idx);bool bnxt_bs_trace_avail (struct bnxt *bp,u16 type);void bnxt_set_tpa_flags (struct bnxt *bp);void bnxt_set_ring_params (struct bnxt *); -int bnxt_set_rx_skb_mode (struct bnxt *bp,bool page_mode); +void bnxt_set_rx_skb_mode (struct bnxt *bp,bool page_mode);void bnxt_insert_usr_fltr (struct bnxt *bp,struct bnxt_filter_base *fltr);void bnxt_del_one_usr_fltr (struct bnxt *bp,struct bnxt_filter_base *fltr); int bnxt_hwrm_func_drv_rgtr(struct bnxt *bp, unsigned long *bmap,
drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c +0 -7
@@ -422,15 +422,8 @@ static int bnxt_xdp_set(struct bnxt *bp, struct bpf_prog *prog)bnxt_set_rx_skb_mode (bp, true);xdp_features_set_redirect_target (dev, true); }else { -int rx, tx; -xdp_features_clear_redirect_target (dev);bnxt_set_rx_skb_mode (bp, false); -bnxt_get_max_rings (bp, &rx, &tx, true); -if (rx >1 ) { - bp->flags &= ~BNXT_FLAG_NO_AGG_RINGS ; - bp->dev->hw_features |=NETIF_F_LRO ; - } } bp->tx_nr_rings_xdp = tx_xdp; bp->tx_nr_rings = bp->tx_nr_rings_per_tc * tc + tx_xdp;
Signature · 1 line
2.47.1
Quoted 9 lines · Jakub Kicinski
Recalculate features when XDP is detached.
Before:
# ip li set dev eth0 xdp obj xdp_dummy.bpf.o sec xdp # ip li set dev eth0 xdp off # ethtool -k eth0 | grep gro rx-gro-hw: off [requested on]
Quoted 8 lines
After:
# ip li set dev eth0 xdp obj xdp_dummy.bpf.o sec xdp # ip li set dev eth0 xdp off # ethtool -k eth0 | grep gro rx-gro-hw: onThe fact that HW-GRO doesn't get re-enabled automatically is just a minor annoyance.
I knew about the annoyance, but didn't know about this possible crash.
Quoted 30 lines
The real issue is that the features will randomly
come back during another reconfiguration which just happens to invokenetdev_update_features(). The driver doesn't handle reconfiguring two things at a time very robustly.Starting with commit
98ba1d931f61("bnxt_en: Fix RSS logic in__bnxt_reserve_rings()") we only reconfigure the RSS hash table if the "effective" number of Rx rings has changed. If HW-GRO is enabled "effective" number of rings is 2x what user sees. So if we are in the bad state, with HW-GRO re-enablement "pending" after XDP off, and we lower the rings by / 2 - the HW-GRO rings doing 2x and the ethtool -L doing / 2 may cancel each other out, and the:if (
old_rx_rings!=bp->hw_resc.resv_rx_rings&&condition in
__bnxt_reserve_rings()will be false. The RSS map won't get updated, and we'll crash with:BUG: kernel NULL pointer dereference, address: 0000000000000168 RIP: 0010:__bnxt_hwrm_vnic_set_rss+0x13a/0x1a0 bnxt_hwrm_vnic_rss_cfg_p5+0x47/0x180 __bnxt_setup_vnic_p5+0x58/0x110 bnxt_init_nic+0xb72/0xf50 __bnxt_open_nic+0x40d/0xab0 bnxt_open_nic+0x2b/0x60 ethtool_set_channels+0x18c/0x1d0As we try to access a freed ring.
Quoted 7 lines
The issue is present since XDP support was added, really, but prior to commit
98ba1d931f61("bnxt_en: Fix RSS logic in__bnxt_reserve_rings()") it wasn't causing major issues.Fixes 1054aee82321 ("bnxt_en: Use NETIF_F_GRO_HW.")Fixes 98ba1d931f61 ("bnxt_en: Fix RSS logic in __bnxt_reserve_rings()")Signed-off-by Jakub Kicinski <[email protected]>
Thanks for the patch.
Quoted 156 lines · Jakub Kicinski
Recalculate features when XDP is detached.
Before:
# ip li set dev eth0 xdp obj xdp_dummy.bpf.o sec xdp # ip li set dev eth0 xdp off # ethtool -k eth0 | grep gro rx-gro-hw: off [requested on]After:
# ip li set dev eth0 xdp obj xdp_dummy.bpf.o sec xdp # ip li set dev eth0 xdp off # ethtool -k eth0 | grep gro rx-gro-hw: onThe fact that HW-GRO doesn't get re-enabled automatically is just a minor annoyance. The real issue is that the features will randomly come back during another reconfiguration which just happens to invoke
netdev_update_features(). The driver doesn't handle reconfiguring two things at a time very robustly.Starting with commit
98ba1d931f61("bnxt_en: Fix RSS logic in__bnxt_reserve_rings()") we only reconfigure the RSS hash table if the "effective" number of Rx rings has changed. If HW-GRO is enabled "effective" number of rings is 2x what user sees. So if we are in the bad state, with HW-GRO re-enablement "pending" after XDP off, and we lower the rings by / 2 - the HW-GRO rings doing 2x and the ethtool -L doing / 2 may cancel each other out, and the:if (
old_rx_rings!=bp->hw_resc.resv_rx_rings&&condition in
__bnxt_reserve_rings()will be false. The RSS map won't get updated, and we'll crash with:BUG: kernel NULL pointer dereference, address: 0000000000000168 RIP: 0010:__bnxt_hwrm_vnic_set_rss+0x13a/0x1a0 bnxt_hwrm_vnic_rss_cfg_p5+0x47/0x180 __bnxt_setup_vnic_p5+0x58/0x110 bnxt_init_nic+0xb72/0xf50 __bnxt_open_nic+0x40d/0xab0 bnxt_open_nic+0x2b/0x60 ethtool_set_channels+0x18c/0x1d0As we try to access a freed ring.
The issue is present since XDP support was added, really, but prior to commit
98ba1d931f61("bnxt_en: Fix RSS logic in__bnxt_reserve_rings()") it wasn't causing major issues.Fixes 1054aee82321 ("bnxt_en: Use NETIF_F_GRO_HW.")Fixes 98ba1d931f61 ("bnxt_en: Fix RSS logic in __bnxt_reserve_rings()")Signed-off-by Jakub Kicinski <[email protected]>---
CC: [email protected]
CC: [email protected]
CC: [email protected]
CC: [email protected]
CC: [email protected]
CC: [email protected]--- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 25 +++++++++++++++---- drivers/net/ethernet/broadcom/bnxt/bnxt.h | 2 +- drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c | 7 ------ 3 files changed, 21 insertions(+), 13 deletions(-)drivers/net/ethernet/broadcom/bnxt/bnxt.c +20
-5@@ -4708,7 +4708,7 @@ void bnxt_set_ring_params(struct bnxt *bp)/* Changing allocation mode of RX rings. * TODO: Update when extending xdp_rxq_info to support allocation modes. */ -int bnxt_set_rx_skb_mode (struct bnxt *bp,bool page_mode) +static void __bnxt_set_rx_skb_mode (struct bnxt *bp,bool page_mode) {struct net_device *dev = bp->dev;@@ -4729,15 +4729,30 @@ int bnxt_set_rx_skb_mode(struct bnxt *bp, bool page_mode) bp->rx_skb_func = bnxt_rx_page_skb; } bp->rx_dir =DMA_BIDIRECTIONAL ; -/* Disable LRO or GRO_HW */ -netdev_update_features (dev); }else { dev->max_mtu = bp->max_mtu; bp->flags &= ~BNXT_FLAG_RX_PAGE_MODE ; bp->rx_dir =DMA_FROM_DEVICE ; bp->rx_skb_func = bnxt_rx_skb; } -return 0 ; +} + +void bnxt_set_rx_skb_mode (struct bnxt *bp,bool page_mode) +{ +__bnxt_set_rx_skb_mode (bp, page_mode); + +if (!page_mode) { +int rx, tx; + +bnxt_get_max_rings (bp, &rx, &tx, true); +if (rx >1 ) { + bp->flags &= ~BNXT_FLAG_NO_AGG_RINGS ; + bp->dev->hw_features |=NETIF_F_LRO ; + } + } + +/* Update LRO and GRO_HW availability */ +netdev_update_features (bp->dev); }static void bnxt_free_vnic_attributes (struct bnxt *bp)@@ -16214,7 +16229,7 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)if (bp->max_fltr <BNXT_MAX_FLTR ) bp->max_fltr =BNXT_MAX_FLTR ;bnxt_init_l2_fltr_tbl (bp); -bnxt_set_rx_skb_mode (bp, false); +__bnxt_set_rx_skb_mode (bp, false);bnxt_set_tpa_flags (bp);bnxt_set_ring_params (bp);bnxt_rdma_aux_device_init (bp);drivers/net/ethernet/broadcom/bnxt/bnxt.h +1
-1@@ -2846,7 +2846,7 @@ u32 bnxt_fw_health_readl(struct bnxt *bp, int reg_idx);bool bnxt_bs_trace_avail (struct bnxt *bp,u16 type);void bnxt_set_tpa_flags (struct bnxt *bp);void bnxt_set_ring_params (struct bnxt *); -int bnxt_set_rx_skb_mode (struct bnxt *bp,bool page_mode); +void bnxt_set_rx_skb_mode (struct bnxt *bp,bool page_mode);void bnxt_insert_usr_fltr (struct bnxt *bp,struct bnxt_filter_base *fltr);void bnxt_del_one_usr_fltr (struct bnxt *bp,struct bnxt_filter_base *fltr); int bnxt_hwrm_func_drv_rgtr(struct bnxt *bp, unsigned long *bmap,drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c +0
-8@@ -422,15 +422,8 @@ static int bnxt_xdp_set(struct bnxt *bp, struct bpf_prog *prog)bnxt_set_rx_skb_mode (bp, true);xdp_features_set_redirect_target (dev, true); }else { -int rx, tx; -xdp_features_clear_redirect_target (dev);bnxt_set_rx_skb_mode (bp, false); -bnxt_get_max_rings (bp, &rx, &tx, true); -if (rx >1 ) { - bp->flags &= ~BNXT_FLAG_NO_AGG_RINGS ; - bp->dev->hw_features |=NETIF_F_LRO ; - } } bp->tx_nr_rings_xdp = tx_xdp; bp->tx_nr_rings = bp->tx_nr_rings_per_tc * tc + tx_xdp; --
Quoted 3 lines
2.47.1
This patch was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:
Hello:
This patch was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:
Quoted 9 lines · On Wed, 8 Jan 2025 20:30:57 -0800 you wrote:
Recalculate features when XDP is detached.
Before:
# ip li set dev eth0 xdp obj xdp_dummy.bpf.o sec xdp # ip li set dev eth0 xdp off # ethtool -k eth0 | grep gro rx-gro-hw: off [requested on][...]
Here is the summary with links:
- [net] eth: bnxt: always recalculate features after XDP clearing, fix null-deref
https://git.kernel.org/netdev/net/c/f0aa6a37a3db
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html