| Message ID | 1440770776-27951-1-git-send-email-aik@ozlabs.ru |
|---|---|
| State | RFC, archived |
| Delegated to: | David Miller |
| Headers | show |
On Fri, Aug 28, 2015 at 7:06 AM, Alexey Kardashevskiy <aik@ozlabs.ru> wrote: > 68230242cdb breaks SRIOV on POWER8 system. I am not really suggesting > reverting the patch, rather asking for a fix. thanks for the detailed report, we will look into that. Just to be sure, when going back in time, what is the latest upstream version where this system/config works okay? is that 4.1 or later? > > To reproduce it: > > 1. boot latest upstream kernel (v4.2-rc8 sha1 4941b8f, ppc64le) > > 2. Run: > sudo rmmod mlx4_en mlx4_ib mlx4_core > sudo modprobe mlx4_core num_vfs=4 probe_vf=4 port_type_array=2,2 debug_level=1 > > 3. Run QEMU (just to give a complete picture): > /home/aik/qemu-system-ppc64 -enable-kvm -m 2048 -machine pseries \ > -nodefaults \ > -chardev stdio,id=id0,signal=off,mux=on \ > -device spapr-vty,id=id1,chardev=id0,reg=0x71000100 \ > -mon id=id2,chardev=id0,mode=readline -nographic -vga none \ > -initrd dhclient.cpio -kernel vml400bedbg \ > -device vfio-pci,id=id3,host=0003:03:00.1 > What guest is used does not matter at all. > > 4. Wait till guest boots and then run: > dhclient > This assigns IPs to both interfaces just fine. This is essential - > if interface was not brought up since guest started, the bug does not appear. > If interface was up and then down, this still causes the problem > (less likely though). > > 5. Run in the guest: shutdown -h 0 > Guest prints: > mlx4_en: eth0: Close port called > mlx4_en: eth1: Close port called > mlx4_core 0000:00:00.0: mlx4_shutdown was called > And then the host hangs. After 10-30 seconds the host console prints: > NMI watchdog: BUG: soft lockup - CPU#0 stuck for 22s! [qemu-system-ppc:5095] > OR > INFO: rcu_sched detected stalls on CPUs/tasks: > or some other random stuff but always related to some sort of lockup. > Backtraces are like these: > > [c000001e492a7ac0] [c000000000135b84] smp_call_function_many+0x2f4/0x3fable) > [c000001e492a7b40] [c000000000135db8] kick_all_cpus_sync+0x38/0x50 > [c000001e492a7b60] [c000000000048f38] pmdp_huge_get_and_clear+0x48/0x70 > [c000001e492a7b90] [c00000000023181c] change_huge_pmd+0xac/0x210 > [c000001e492a7bf0] [c0000000001fb9e8] change_protection+0x678/0x720 > [c000001e492a7d00] [c000000000217d38] change_prot_numa+0x28/0xa0 > [c000001e492a7d30] [c0000000000e0e40] task_numa_work+0x2a0/0x370 > [c000001e492a7db0] [c0000000000c5fb4] task_work_run+0xe4/0x160 > [c000001e492a7e00] [c0000000000169a4] do_notify_resume+0x84/0x90 > [c000001e492a7e30] [c0000000000098b8] ret_from_except_lite+0x64/0x68 > > OR > > [c000001def1b7280] [c000000ff941d368] 0xc000000ff941d368 (unreliable) > [c000001def1b7450] [c00000000001512c] __switch_to+0x1fc/0x350 > [c000001def1b7490] [c000001def1b74e0] 0xc000001def1b74e0 > [c000001def1b74e0] [c00000000011a50c] try_to_del_timer_sync+0x5c/0x90 > [c000001def1b7520] [c00000000011a590] del_timer_sync+0x50/0x70 > [c000001def1b7550] [c0000000009136fc] schedule_timeout+0x15c/0x2b0 > [c000001def1b7620] [c000000000910e6c] wait_for_common+0x12c/0x230 > [c000001def1b7660] [c0000000000fa22c] up+0x4c/0x80 > [c000001def1b76a0] [d000000016323e60] __mlx4_cmd+0x320/0x940 [mlx4_core] > [c000001def1b7760] [c000001def1b77a0] 0xc000001def1b77a0 > [c000001def1b77f0] [d0000000163528b4] mlx4_2RST_QP_wrapper+0x154/0x1e0 [mlx4_core] > [c000001def1b7860] [d000000016324934] mlx4_master_process_vhcr+0x1b4/0x6c0 [mlx4_core] > [c000001def1b7930] [d000000016324170] __mlx4_cmd+0x630/0x940 [mlx4_core] > [c000001def1b79f0] [d000000016346fec] __mlx4_qp_modify.constprop.8+0x1ec/0x350 [mlx4_core] > [c000001def1b7ac0] [d000000016292228] mlx4_ib_destroy_qp+0xd8/0x5d0 [mlx4_ib] > [c000001def1b7b60] [d000000013c7305c] ib_destroy_qp+0x1cc/0x290 [ib_core] > [c000001def1b7bb0] [d000000016284548] destroy_pv_resources.isra.14.part.15+0x48/0xf0 [mlx4_ib] > [c000001def1b7be0] [d000000016284d28] mlx4_ib_tunnels_update+0x168/0x170 [mlx4_ib] > [c000001def1b7c20] [d0000000162876e0] mlx4_ib_tunnels_update_work+0x30/0x50 [mlx4_ib] > [c000001def1b7c50] [c0000000000c0d34] process_one_work+0x194/0x490 > [c000001def1b7ce0] [c0000000000c11b0] worker_thread+0x180/0x5a0 > [c000001def1b7d80] [c0000000000c8a0c] kthread+0x10c/0x130 > [c000001def1b7e30] [c0000000000095a8] ret_from_kernel_thread+0x5c/0xb4 > > i.e. may or may not mention mlx4. > The issue may not happen on a first try but maximum on the second. so when you revert commit 68230242cdb on the host all works just fine? what guest driver are you running? This needs a fix, I don't think the right thing to do is just go and revert the commit, if the right fix misses 4.2 we will get it there through -stable -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 08/30/2015 04:28 PM, Or Gerlitz wrote: > On Fri, Aug 28, 2015 at 7:06 AM, Alexey Kardashevskiy <aik@ozlabs.ru> wrote: >> 68230242cdb breaks SRIOV on POWER8 system. I am not really suggesting >> reverting the patch, rather asking for a fix. > > thanks for the detailed report, we will look into that. > > Just to be sure, when going back in time, what is the latest upstream > version where > this system/config works okay? is that 4.1 or later? 4.1 is good, 4.2 is not. > >> >> To reproduce it: >> >> 1. boot latest upstream kernel (v4.2-rc8 sha1 4941b8f, ppc64le) >> >> 2. Run: >> sudo rmmod mlx4_en mlx4_ib mlx4_core >> sudo modprobe mlx4_core num_vfs=4 probe_vf=4 port_type_array=2,2 debug_level=1 >> >> 3. Run QEMU (just to give a complete picture): >> /home/aik/qemu-system-ppc64 -enable-kvm -m 2048 -machine pseries \ >> -nodefaults \ >> -chardev stdio,id=id0,signal=off,mux=on \ >> -device spapr-vty,id=id1,chardev=id0,reg=0x71000100 \ >> -mon id=id2,chardev=id0,mode=readline -nographic -vga none \ >> -initrd dhclient.cpio -kernel vml400bedbg \ >> -device vfio-pci,id=id3,host=0003:03:00.1 >> What guest is used does not matter at all. >> >> 4. Wait till guest boots and then run: >> dhclient >> This assigns IPs to both interfaces just fine. This is essential - >> if interface was not brought up since guest started, the bug does not appear. >> If interface was up and then down, this still causes the problem >> (less likely though). >> >> 5. Run in the guest: shutdown -h 0 >> Guest prints: >> mlx4_en: eth0: Close port called >> mlx4_en: eth1: Close port called >> mlx4_core 0000:00:00.0: mlx4_shutdown was called >> And then the host hangs. After 10-30 seconds the host console prints: >> NMI watchdog: BUG: soft lockup - CPU#0 stuck for 22s! [qemu-system-ppc:5095] >> OR >> INFO: rcu_sched detected stalls on CPUs/tasks: >> or some other random stuff but always related to some sort of lockup. >> Backtraces are like these: >> >> [c000001e492a7ac0] [c000000000135b84] smp_call_function_many+0x2f4/0x3fable) >> [c000001e492a7b40] [c000000000135db8] kick_all_cpus_sync+0x38/0x50 >> [c000001e492a7b60] [c000000000048f38] pmdp_huge_get_and_clear+0x48/0x70 >> [c000001e492a7b90] [c00000000023181c] change_huge_pmd+0xac/0x210 >> [c000001e492a7bf0] [c0000000001fb9e8] change_protection+0x678/0x720 >> [c000001e492a7d00] [c000000000217d38] change_prot_numa+0x28/0xa0 >> [c000001e492a7d30] [c0000000000e0e40] task_numa_work+0x2a0/0x370 >> [c000001e492a7db0] [c0000000000c5fb4] task_work_run+0xe4/0x160 >> [c000001e492a7e00] [c0000000000169a4] do_notify_resume+0x84/0x90 >> [c000001e492a7e30] [c0000000000098b8] ret_from_except_lite+0x64/0x68 >> >> OR >> >> [c000001def1b7280] [c000000ff941d368] 0xc000000ff941d368 (unreliable) >> [c000001def1b7450] [c00000000001512c] __switch_to+0x1fc/0x350 >> [c000001def1b7490] [c000001def1b74e0] 0xc000001def1b74e0 >> [c000001def1b74e0] [c00000000011a50c] try_to_del_timer_sync+0x5c/0x90 >> [c000001def1b7520] [c00000000011a590] del_timer_sync+0x50/0x70 >> [c000001def1b7550] [c0000000009136fc] schedule_timeout+0x15c/0x2b0 >> [c000001def1b7620] [c000000000910e6c] wait_for_common+0x12c/0x230 >> [c000001def1b7660] [c0000000000fa22c] up+0x4c/0x80 >> [c000001def1b76a0] [d000000016323e60] __mlx4_cmd+0x320/0x940 [mlx4_core] >> [c000001def1b7760] [c000001def1b77a0] 0xc000001def1b77a0 >> [c000001def1b77f0] [d0000000163528b4] mlx4_2RST_QP_wrapper+0x154/0x1e0 [mlx4_core] >> [c000001def1b7860] [d000000016324934] mlx4_master_process_vhcr+0x1b4/0x6c0 [mlx4_core] >> [c000001def1b7930] [d000000016324170] __mlx4_cmd+0x630/0x940 [mlx4_core] >> [c000001def1b79f0] [d000000016346fec] __mlx4_qp_modify.constprop.8+0x1ec/0x350 [mlx4_core] >> [c000001def1b7ac0] [d000000016292228] mlx4_ib_destroy_qp+0xd8/0x5d0 [mlx4_ib] >> [c000001def1b7b60] [d000000013c7305c] ib_destroy_qp+0x1cc/0x290 [ib_core] >> [c000001def1b7bb0] [d000000016284548] destroy_pv_resources.isra.14.part.15+0x48/0xf0 [mlx4_ib] >> [c000001def1b7be0] [d000000016284d28] mlx4_ib_tunnels_update+0x168/0x170 [mlx4_ib] >> [c000001def1b7c20] [d0000000162876e0] mlx4_ib_tunnels_update_work+0x30/0x50 [mlx4_ib] >> [c000001def1b7c50] [c0000000000c0d34] process_one_work+0x194/0x490 >> [c000001def1b7ce0] [c0000000000c11b0] worker_thread+0x180/0x5a0 >> [c000001def1b7d80] [c0000000000c8a0c] kthread+0x10c/0x130 >> [c000001def1b7e30] [c0000000000095a8] ret_from_kernel_thread+0x5c/0xb4 >> >> i.e. may or may not mention mlx4. >> The issue may not happen on a first try but maximum on the second. > > so when you revert commit 68230242cdb on the host all works just fine? > what guest driver are you running? To be precise, I did checkout 68230242cdb, checked that it does not work, then reverted 68230242cdb right there and checked that it works. I did not try reverting later revisions yet. My guest kernel in this test has tag v4.0. I get the same effect with some 3.18 from Ubuntu 14.04 LTS so the guest kernel version does not make a difference afaict. > This needs a fix, I don't think the right thing to do is just go and > revert the commit, if the right fix misses 4.2 we will get it there > through -stable v4.2 was just released :)
diff --git a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c index 73db584..802eb2a 100644 --- a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c +++ b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c @@ -723,9 +723,6 @@ static void update_gid(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *inbox, } } -static int handle_counter(struct mlx4_dev *dev, struct mlx4_qp_context *qpc, - u8 slave, int port); - static int update_vport_qp_param(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *inbox, u8 slave, u32 qpn) @@ -741,10 +738,6 @@ static int update_vport_qp_param(struct mlx4_dev *dev, vp_oper = &priv->mfunc.master.vf_oper[slave].vport[port]; qp_type = (be32_to_cpu(qpc->flags) >> 16) & 0xff; - err = handle_counter(dev, qpc, slave, port); - if (err) - goto out; - if (MLX4_VGT != vp_oper->state.default_vlan) { /* the reserved QPs (special, proxy, tunnel) * do not operate over vlans @@ -889,83 +882,6 @@ static void put_res(struct mlx4_dev *dev, int slave, u64 res_id, spin_unlock_irq(mlx4_tlock(dev)); } -static int counter_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd, - u64 in_param, u64 *out_param, int port); - -static int handle_existing_counter(struct mlx4_dev *dev, u8 slave, int port, - int counter_index) -{ - struct res_common *r; - struct res_counter *counter; - int ret = 0; - - if (counter_index == MLX4_SINK_COUNTER_INDEX(dev)) - return ret; - - spin_lock_irq(mlx4_tlock(dev)); - r = find_res(dev, counter_index, RES_COUNTER); - if (!r || r->owner != slave) - ret = -EINVAL; - counter = container_of(r, struct res_counter, com); - if (!counter->port) - counter->port = port; - - spin_unlock_irq(mlx4_tlock(dev)); - return ret; -} - -static int handle_unexisting_counter(struct mlx4_dev *dev, - struct mlx4_qp_context *qpc, u8 slave, - int port) -{ - struct mlx4_priv *priv = mlx4_priv(dev); - struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker; - struct res_common *tmp; - struct res_counter *counter; - u64 counter_idx = MLX4_SINK_COUNTER_INDEX(dev); - int err = 0; - - spin_lock_irq(mlx4_tlock(dev)); - list_for_each_entry(tmp, - &tracker->slave_list[slave].res_list[RES_COUNTER], - list) { - counter = container_of(tmp, struct res_counter, com); - if (port == counter->port) { - qpc->pri_path.counter_index = counter->com.res_id; - spin_unlock_irq(mlx4_tlock(dev)); - return 0; - } - } - spin_unlock_irq(mlx4_tlock(dev)); - - /* No existing counter, need to allocate a new counter */ - err = counter_alloc_res(dev, slave, RES_OP_RESERVE, 0, 0, &counter_idx, - port); - if (err == -ENOENT) { - err = 0; - } else if (err && err != -ENOSPC) { - mlx4_err(dev, "%s: failed to create new counter for slave %d err %d\n", - __func__, slave, err); - } else { - qpc->pri_path.counter_index = counter_idx; - mlx4_dbg(dev, "%s: alloc new counter for slave %d index %d\n", - __func__, slave, qpc->pri_path.counter_index); - err = 0; - } - - return err; -} - -static int handle_counter(struct mlx4_dev *dev, struct mlx4_qp_context *qpc, - u8 slave, int port) -{ - if (qpc->pri_path.counter_index != MLX4_SINK_COUNTER_INDEX(dev)) - return handle_existing_counter(dev, slave, port, - qpc->pri_path.counter_index); - - return handle_unexisting_counter(dev, qpc, slave, port); -} - static struct res_common *alloc_qp_tr(int id) { struct res_qp *ret; @@ -2109,7 +2025,7 @@ static int vlan_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd, } static int counter_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd, - u64 in_param, u64 *out_param, int port) + u64 in_param, u64 *out_param) { u32 index; int err; @@ -2127,7 +2043,7 @@ static int counter_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd, return err; } - err = add_res_range(dev, slave, index, 1, RES_COUNTER, port); + err = add_res_range(dev, slave, index, 1, RES_COUNTER, 0); if (err) { __mlx4_counter_free(dev, index); mlx4_release_resource(dev, slave, RES_COUNTER, 1, 0); @@ -2209,7 +2125,7 @@ int mlx4_ALLOC_RES_wrapper(struct mlx4_dev *dev, int slave, case RES_COUNTER: err = counter_alloc_res(dev, slave, vhcr->op_modifier, alop, - vhcr->in_param, &vhcr->out_param, 0); + vhcr->in_param, &vhcr->out_param); break; case RES_XRCD:
68230242cdb breaks SRIOV on POWER8 system. I am not really suggesting reverting the patch, rather asking for a fix. To reproduce it: 1. boot latest upstream kernel (v4.2-rc8 sha1 4941b8f, ppc64le) 2. Run: sudo rmmod mlx4_en mlx4_ib mlx4_core sudo modprobe mlx4_core num_vfs=4 probe_vf=4 port_type_array=2,2 debug_level=1 3. Run QEMU (just to give a complete picture): /home/aik/qemu-system-ppc64 -enable-kvm -m 2048 -machine pseries \ -nodefaults \ -chardev stdio,id=id0,signal=off,mux=on \ -device spapr-vty,id=id1,chardev=id0,reg=0x71000100 \ -mon id=id2,chardev=id0,mode=readline -nographic -vga none \ -initrd dhclient.cpio -kernel vml400bedbg \ -device vfio-pci,id=id3,host=0003:03:00.1 What guest is used does not matter at all. 4. Wait till guest boots and then run: dhclient This assigns IPs to both interfaces just fine. This is essential - if interface was not brought up since guest started, the bug does not appear. If interface was up and then down, this still causes the problem (less likely though). 5. Run in the guest: shutdown -h 0 Guest prints: mlx4_en: eth0: Close port called mlx4_en: eth1: Close port called mlx4_core 0000:00:00.0: mlx4_shutdown was called And then the host hangs. After 10-30 seconds the host console prints: NMI watchdog: BUG: soft lockup - CPU#0 stuck for 22s! [qemu-system-ppc:5095] OR INFO: rcu_sched detected stalls on CPUs/tasks: or some other random stuff but always related to some sort of lockup. Backtraces are like these: [c000001e492a7ac0] [c000000000135b84] smp_call_function_many+0x2f4/0x3fable) [c000001e492a7b40] [c000000000135db8] kick_all_cpus_sync+0x38/0x50 [c000001e492a7b60] [c000000000048f38] pmdp_huge_get_and_clear+0x48/0x70 [c000001e492a7b90] [c00000000023181c] change_huge_pmd+0xac/0x210 [c000001e492a7bf0] [c0000000001fb9e8] change_protection+0x678/0x720 [c000001e492a7d00] [c000000000217d38] change_prot_numa+0x28/0xa0 [c000001e492a7d30] [c0000000000e0e40] task_numa_work+0x2a0/0x370 [c000001e492a7db0] [c0000000000c5fb4] task_work_run+0xe4/0x160 [c000001e492a7e00] [c0000000000169a4] do_notify_resume+0x84/0x90 [c000001e492a7e30] [c0000000000098b8] ret_from_except_lite+0x64/0x68 OR [c000001def1b7280] [c000000ff941d368] 0xc000000ff941d368 (unreliable) [c000001def1b7450] [c00000000001512c] __switch_to+0x1fc/0x350 [c000001def1b7490] [c000001def1b74e0] 0xc000001def1b74e0 [c000001def1b74e0] [c00000000011a50c] try_to_del_timer_sync+0x5c/0x90 [c000001def1b7520] [c00000000011a590] del_timer_sync+0x50/0x70 [c000001def1b7550] [c0000000009136fc] schedule_timeout+0x15c/0x2b0 [c000001def1b7620] [c000000000910e6c] wait_for_common+0x12c/0x230 [c000001def1b7660] [c0000000000fa22c] up+0x4c/0x80 [c000001def1b76a0] [d000000016323e60] __mlx4_cmd+0x320/0x940 [mlx4_core] [c000001def1b7760] [c000001def1b77a0] 0xc000001def1b77a0 [c000001def1b77f0] [d0000000163528b4] mlx4_2RST_QP_wrapper+0x154/0x1e0 [mlx4_core] [c000001def1b7860] [d000000016324934] mlx4_master_process_vhcr+0x1b4/0x6c0 [mlx4_core] [c000001def1b7930] [d000000016324170] __mlx4_cmd+0x630/0x940 [mlx4_core] [c000001def1b79f0] [d000000016346fec] __mlx4_qp_modify.constprop.8+0x1ec/0x350 [mlx4_core] [c000001def1b7ac0] [d000000016292228] mlx4_ib_destroy_qp+0xd8/0x5d0 [mlx4_ib] [c000001def1b7b60] [d000000013c7305c] ib_destroy_qp+0x1cc/0x290 [ib_core] [c000001def1b7bb0] [d000000016284548] destroy_pv_resources.isra.14.part.15+0x48/0xf0 [mlx4_ib] [c000001def1b7be0] [d000000016284d28] mlx4_ib_tunnels_update+0x168/0x170 [mlx4_ib] [c000001def1b7c20] [d0000000162876e0] mlx4_ib_tunnels_update_work+0x30/0x50 [mlx4_ib] [c000001def1b7c50] [c0000000000c0d34] process_one_work+0x194/0x490 [c000001def1b7ce0] [c0000000000c11b0] worker_thread+0x180/0x5a0 [c000001def1b7d80] [c0000000000c8a0c] kthread+0x10c/0x130 [c000001def1b7e30] [c0000000000095a8] ret_from_kernel_thread+0x5c/0xb4 i.e. may or may not mention mlx4. The issue may not happen on a first try but maximum on the second. This is the function I am passing to the guest: 0003:03:00.1 Ethernet controller: Mellanox Technologies MT27500/MT27520 Family [ConnectX-3/ConnectX-3 Pro Virtual Function] Subsystem: IBM Device 61b0 Flags: bus master, fast devsel, latency 0 [virtual] Memory at 3c1080000000 (64-bit, prefetchable) [size=128M] Capabilities: <access denied> Kernel driver in use: mlx4_core And ideas? Some patches to try? Thanks! --- .../net/ethernet/mellanox/mlx4/resource_tracker.c | 90 +--------------------- 1 file changed, 3 insertions(+), 87 deletions(-)