igc: Fix BUG: scheduling while atomic: kworker/u64:0/9/0x00000002
Subsystems (2)
PATCH 1/1 igc: Fix BUG: scheduling while atomic: kworker/u64:0/9/0x00000002
Replace usleep_range() method with udelay() method to allow atomic contects in low-level MDIO access functions.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=2014971
--- drivers/net/ethernet/intel/igc/igc_phy.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/ethernet/intel/igc/igc_phy.c +2 -2
@@ -581,7 +581,7 @@ static s32 igc_read_phy_reg_mdic(struct igc_hw *hw, u32 offset, u16 *data)* the lower time out */ for (i =0 ; i <IGC_GEN_POLL_TIMEOUT ; i++) { -usleep_range (500 ,1000 ); +udelay (50 ); mdic =rd32 (IGC_MDIC );if (mdic &IGC_MDIC_READY )break ;
@@ -638,7 +638,7 @@ static s32 igc_write_phy_reg_mdic(struct igc_hw *hw, u32 offset, u16 data)* the lower time out */ for (i =0 ; i <IGC_GEN_POLL_TIMEOUT ; i++) { -usleep_range (500 ,1000 ); +udelay (50 ); mdic =rd32 (IGC_MDIC );if (mdic &IGC_MDIC_READY )break ;
Signature · 1 line
2.30.2
PATCH 1/1 igc: Fix BUG: scheduling while atomic: kworker/u64:0/9/0x00000002
Quoted 1 line · Sasha Neftin
Replace
usleep_range()method withudelay()method to allow atomic contects
s/contects/contexts
Quoted 3 lines
in low-level MDIO access functions.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=2014971
I don't have access to that bugzilla, so can I ask to include the splat in the commit message itself?
Quoted 37 lines
Fixes 5586838fe9ce ("igc: Add code for PHY support")Reported-by Corinna Vinschen <[email protected]>Suggested-by Dima Ruinskiy <[email protected]>Signed-off-by Sasha Neftin <[email protected]>--- drivers/net/ethernet/intel/igc/igc_phy.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)drivers/net/ethernet/intel/igc/igc_phy.c +2
-2@@ -581,7 +581,7 @@ static s32 igc_read_phy_reg_mdic(struct igc_hw *hw, u32 offset, u16 *data)* the lower time out */ for (i =0 ; i <IGC_GEN_POLL_TIMEOUT ; i++) { -usleep_range (500 ,1000 ); +udelay (50 ); mdic =rd32 (IGC_MDIC );if (mdic &IGC_MDIC_READY )break ;@@ -638,7 +638,7 @@ static s32 igc_write_phy_reg_mdic(struct igc_hw *hw, u32 offset, u16 data)* the lower time out */ for (i =0 ; i <IGC_GEN_POLL_TIMEOUT ; i++) { -usleep_range (500 ,1000 ); +udelay (50 ); mdic =rd32 (IGC_MDIC );if (mdic &IGC_MDIC_READY )break ;Signature · 6 lines
2.30.2 _______________________________________________ Intel-wired-lan mailing list Intel-wired-lan at osuosl.org https://lists.osuosl.org/mailman/listinfo/intel-wired-lan
PATCH 1/1 igc: Fix BUG: scheduling while atomic: kworker/u64:0/9/0x00000002
In reply to Sasha NeftinHi Sasha,
Quoted 2 lines · Sasha Neftin
Replace
usleep_range()method withudelay()method to allow atomic contects in low-level MDIO access functions.
The commit message is a bit on the short end. It should probably quote the reproducer from the below Bugzilla, rather than just pointing to the Bugzilla, i.e.,
$ modprobe -r bonding $ modprobe -v bonding max_bonds=1 mode=1 miimon=100 use_carrier=0 $ ip link set bond0 up $ ifenslave bond0 eth0 eth1
leads to the following [shortened] kernel log:
[ 120.198086] Voluntary context switch within RCU read-side critical section! [ 120.198093] WARNING: CPU: 3 PID: 10 at kernel/rcu/tree_plugin.h:318 rcu_note_context_switch+0x4ea/0x580 [ 120.207310] bond0: (slave ens5f0): Enslaving as a backup interface with a down link [ 120.229488] [...] [ 120.754723] Workqueue: bond0 bond_mii_monitor [bonding] [ 120.779106] RIP: 0010:rcu_note_context_switch+0x4ea/0x580 [ 120.803880] Code: 08 49 89 8e 40 08 00 00 4d 89 be 48 08 00 00 4d 89 07 e9 98 fe ff ff 48 c7 c7 48 83 94 b6 c6 05 74 3b e1 01 01 e8 35 fd 8a 00 <0f> 0b e9 67 fb ff ff 45 85 f6 0f 84 9e fe ff ff 4c 89 85 a8 00 00 [ 120.890778] RSP: 0018:ffffa2f5c01bbbc8 EFLAGS: 00010082 [ 120.915218] RAX: 0000000000000000 RBX: ffff938d6f8f1740 RCX: 0000000000000027 [ 120.948713] RDX: 0000000000000027 RSI: ffff938d6f8dfcc0 RDI: ffff938d6f8dfcc8 [ 120.982368] RBP: ffffa2f5c01bbc50 R08: 0000000000000000 R09: c0000000ffff7fff [ 121.016065] R10: 0000000000000001 R11: ffffa2f5c01bb9e0 R12: ffff938a022a9980 [ 121.049504] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000001 [ 121.082870] FS: 0000000000000000(0000) GS:ffff938d6f8c0000(0000) knlGS:0000000000000000 [ 121.121993] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 121.151532] CR2: 0000556901ea4f28 CR3: 0000000482810001 CR4: 00000000001706e0 [ 121.186337] Call Trace: [ 121.197543] <TASK> [ 121.207444] __schedule+0x9a/0x7b0 [ 121.223316] ? hrtimer_start_range_ns+0x11d/0x300 [ 121.245293] schedule+0x41/0xc0 [ 121.259860] schedule_hrtimeout_range_clock+0x83/0x100 [ 121.283548] ? hrtimer_init_sleeper+0x80/0x80 [ 121.303830] usleep_range_state+0x5b/0x80 [ 121.322747] igc_read_phy_reg_gpy+0x90/0x180 [igc] [ 121.345249] igc_phy_has_link+0x7d/0x170 [igc] [ 121.365980] igc_check_for_copper_link+0x60/0xd0 [igc] [ 121.390260] igc_has_link+0x2a/0x80 [igc] [ 121.409093] igc_ethtool_get_link+0x24/0x30 [igc] [ 121.431248] bond_check_dev_link.isra.49+0x5a/0x120 [bonding] [ 121.458271] ? pick_next_task_fair+0x3e/0x3b0 [ 121.478749] ? put_prev_entity+0x22/0xe0 [ 121.497154] ? put_prev_task_fair+0x1b/0x30 [ 121.516903] bond_mii_monitor+0x107/0x540 [bonding]
[...]
Quoted 32 lines
Fixes 5586838fe9ce ("igc: Add code for PHY support")Reported-by Corinna Vinschen <[email protected]>Suggested-by Dima Ruinskiy <[email protected]>Signed-off-by Sasha Neftin <[email protected]>--- drivers/net/ethernet/intel/igc/igc_phy.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)drivers/net/ethernet/intel/igc/igc_phy.c +2
-2@@ -581,7 +581,7 @@ static s32 igc_read_phy_reg_mdic(struct igc_hw *hw, u32 offset, u16 *data)* the lower time out */ for (i =0 ; i <IGC_GEN_POLL_TIMEOUT ; i++) { -usleep_range (500 ,1000 ); +udelay (50 ); mdic =rd32 (IGC_MDIC );if (mdic &IGC_MDIC_READY )break ;@@ -638,7 +638,7 @@ static s32 igc_write_phy_reg_mdic(struct igc_hw *hw, u32 offset, u16 data)* the lower time out */ for (i =0 ; i <IGC_GEN_POLL_TIMEOUT ; i++) { -usleep_range (500 ,1000 ); +udelay (50 ); mdic =rd32 (IGC_MDIC );if (mdic &IGC_MDIC_READY )break ;Signature · 1 line
2.30.2
Thanks,
Corinna
Neftin
PATCH 1/1 igc: Fix BUG: scheduling while atomic: kworker/u64:0/9/0x00000002
Quoted 9 lines · Corinna Vinschen
Hi Sasha,
Quoted 2 lines · Sasha Neftin
Replace
usleep_range()method withudelay()method to allow atomic contects in low-level MDIO access functions.The commit message is a bit on the short end. It should probably quote the reproducer from the below Bugzilla, rather than just pointing to the Bugzilla, i.e.,
Thanks Corinna and Maciej - I will resend v2.
Quoted 83 lines
$ modprobe -r bonding $ modprobe -v bonding max_bonds=1 mode=1 miimon=100 use_carrier=0 $ ip link set bond0 up $ ifenslave bond0 eth0 eth1leads to the following [shortened] kernel log:
[ 120.198086] Voluntary context switch within RCU read-side critical section! [ 120.198093] WARNING: CPU: 3 PID: 10 at kernel/rcu/tree_plugin.h:318 rcu_note_context_switch+0x4ea/0x580 [ 120.207310] bond0: (slave ens5f0): Enslaving as a backup interface with a down link [ 120.229488] [...] [ 120.754723] Workqueue: bond0 bond_mii_monitor [bonding] [ 120.779106] RIP: 0010:rcu_note_context_switch+0x4ea/0x580 [ 120.803880] Code: 08 49 89 8e 40 08 00 00 4d 89 be 48 08 00 00 4d 89 07 e9 98 fe ff ff 48 c7 c7 48 83 94 b6 c6 05 74 3b e1 01 01 e8 35 fd 8a 00 <0f> 0b e9 67 fb ff ff 45 85 f6 0f 84 9e fe ff ff 4c 89 85 a8 00 00 [ 120.890778] RSP: 0018:ffffa2f5c01bbbc8 EFLAGS: 00010082 [ 120.915218] RAX: 0000000000000000 RBX: ffff938d6f8f1740 RCX: 0000000000000027 [ 120.948713] RDX: 0000000000000027 RSI: ffff938d6f8dfcc0 RDI: ffff938d6f8dfcc8 [ 120.982368] RBP: ffffa2f5c01bbc50 R08: 0000000000000000 R09: c0000000ffff7fff [ 121.016065] R10: 0000000000000001 R11: ffffa2f5c01bb9e0 R12: ffff938a022a9980 [ 121.049504] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000001 [ 121.082870] FS: 0000000000000000(0000) GS:ffff938d6f8c0000(0000) knlGS:0000000000000000 [ 121.121993] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 121.151532] CR2: 0000556901ea4f28 CR3: 0000000482810001 CR4: 00000000001706e0 [ 121.186337] Call Trace: [ 121.197543] <TASK> [ 121.207444] __schedule+0x9a/0x7b0 [ 121.223316] ? hrtimer_start_range_ns+0x11d/0x300 [ 121.245293] schedule+0x41/0xc0 [ 121.259860] schedule_hrtimeout_range_clock+0x83/0x100 [ 121.283548] ? hrtimer_init_sleeper+0x80/0x80 [ 121.303830] usleep_range_state+0x5b/0x80 [ 121.322747] igc_read_phy_reg_gpy+0x90/0x180 [igc] [ 121.345249] igc_phy_has_link+0x7d/0x170 [igc] [ 121.365980] igc_check_for_copper_link+0x60/0xd0 [igc] [ 121.390260] igc_has_link+0x2a/0x80 [igc] [ 121.409093] igc_ethtool_get_link+0x24/0x30 [igc] [ 121.431248] bond_check_dev_link.isra.49+0x5a/0x120 [bonding] [ 121.458271] ? pick_next_task_fair+0x3e/0x3b0 [ 121.478749] ? put_prev_entity+0x22/0xe0 [ 121.497154] ? put_prev_task_fair+0x1b/0x30 [ 121.516903] bond_mii_monitor+0x107/0x540 [bonding][...]
Quoted 32 lines
Fixes 5586838fe9ce ("igc: Add code for PHY support")Reported-by Corinna Vinschen <[email protected]>Suggested-by Dima Ruinskiy <[email protected]>Signed-off-by Sasha Neftin <[email protected]>--- drivers/net/ethernet/intel/igc/igc_phy.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)drivers/net/ethernet/intel/igc/igc_phy.c +2
-2@@ -581,7 +581,7 @@ static s32 igc_read_phy_reg_mdic(struct igc_hw *hw, u32 offset, u16 *data)* the lower time out */ for (i =0 ; i <IGC_GEN_POLL_TIMEOUT ; i++) { -usleep_range (500 ,1000 ); +udelay (50 ); mdic =rd32 (IGC_MDIC );if (mdic &IGC_MDIC_READY )break ;@@ -638,7 +638,7 @@ static s32 igc_write_phy_reg_mdic(struct igc_hw *hw, u32 offset, u16 data)* the lower time out */ for (i =0 ; i <IGC_GEN_POLL_TIMEOUT ; i++) { -usleep_range (500 ,1000 ); +udelay (50 ); mdic =rd32 (IGC_MDIC );if (mdic &IGC_MDIC_READY )break ;Signature · 1 line
2.30.2Tested-by Corinna Vinschen <[email protected]>Thanks,
Corinna
PATCH 1/1 igc: Fix BUG: scheduling while atomic: kworker/u64:0/9/0x00000002
In reply to Sasha NeftinQuoted 11 lines · Sasha Neftin
Replace
usleep_range()method withudelay()method to allow atomic contects in low-level MDIO access functions.Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=2014971
Fixes 5586838fe9ce ("igc: Add code for PHY support")Reported-by Corinna Vinschen <[email protected]>Suggested-by Dima Ruinskiy <[email protected]>Signed-off-by Sasha Neftin <[email protected]>--- drivers/net/ethernet/intel/igc/igc_phy.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)