- Kernel - mailweb.openeuler.org

[PATCH] net-netdevsim-fix-resource-leak
by liaichun 08 Jun '20

08 Jun '20

diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index f61d094..d5c01b8 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -241,8 +241,8 @@ static int __net_init nsim_fib_netns_init(struct net *net) void nsim_fib_exit(void) { - unregister_pernet_subsys(&nsim_fib_net_ops); unregister_fib_notifier(&nsim_fib_nb); + unregister_pernet_subsys(&nsim_fib_net_ops); } int nsim_fib_init(void) @@ -257,6 +257,7 @@ int nsim_fib_init(void) err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent); if (err < 0) { + unregister_pernet_subsys(&nsim_fib_net_ops); pr_err("Failed to register fib notifier\n"); goto err_out; } -- 1.8.3.1

1 0

[PATCH v2] scsi: hisi_sas: directly trigger error handle of scsi instead of waiting for timeout
by Luo Jiaxing 05 Jun '20

05 Jun '20

We used timeout mechanism of scsi layer to trigger some IO's error handle, this type of abnormal IO require driver to enter error handle to clear the residue in the hardware. But timeout mechanism caught error handle time to be longer, some threads need to wait for tens of seconds when IO failed and enter error handle, so we try to trigger error handling directly. Signed-off-by: Luo Jiaxing <luojiaxing(a)huawei.com> --- drivers/scsi/hisi_sas/hisi_sas_v1_hw.c | 4 +++- drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 4 +++- drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 4 +++- 3 files changed, 9 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c index 031091e..49ea3b7 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c @@ -1263,8 +1263,10 @@ static int slot_complete_v1_hw(struct hisi_hba *hisi_hba, !(cmplt_hdr_data & CMPLT_HDR_RSPNS_XFRD_MSK)) { slot_err_v1_hw(hisi_hba, task, slot); - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } diff --git a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c index 6568e20..8ebddf5 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c @@ -2406,8 +2406,10 @@ slot_complete_v2_hw(struct hisi_hba *hisi_hba, struct hisi_sas_slot *slot) error_info[0], error_info[1], error_info[2], error_info[3]); - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c index 5b80856..7fd067c 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c @@ -2408,8 +2408,10 @@ slot_complete_v3_hw(struct hisi_hba *hisi_hba, struct hisi_sas_slot *slot) dev_info(dev, "data underflow without sense, rsp_code:0x%x, rc:%d.\n", iu->resp_data[0], rc); } - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } -- 2.7.4

1 0

[PATCH v1 hulk-4.19-next] scsi: hisi_sas: directly trigger error handle of scsi instead of waiting for timeout
by Luo Jiaxing 05 Jun '20

05 Jun '20

driver inclusion category: bugfix bugzilla: NA DTS: NA ----------------------------------------------------------------------- We used timeout mechanism of scsi layer to trigger some IO's error handle, this type of abnormal IO require driver to enter error handle to clear the residue in the hardware. But timeout mechanism caught error handle time to be longer, some threads need to wait for tens of seconds when IO failed and enter error handle, so we try to trigger error handling directly. Signed-off-by: Luo Jiaxing <luojiaxing(a)huawei.com> --- drivers/scsi/hisi_sas/hisi_sas_v1_hw.c | 4 +++- drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 4 +++- drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 4 +++- 3 files changed, 9 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c index 031091e..49ea3b7 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c @@ -1263,8 +1263,10 @@ static int slot_complete_v1_hw(struct hisi_hba *hisi_hba, !(cmplt_hdr_data & CMPLT_HDR_RSPNS_XFRD_MSK)) { slot_err_v1_hw(hisi_hba, task, slot); - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } diff --git a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c index 6568e20..8ebddf5 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c @@ -2406,8 +2406,10 @@ slot_complete_v2_hw(struct hisi_hba *hisi_hba, struct hisi_sas_slot *slot) error_info[0], error_info[1], error_info[2], error_info[3]); - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c index 5b80856..7fd067c 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c @@ -2408,8 +2408,10 @@ slot_complete_v3_hw(struct hisi_hba *hisi_hba, struct hisi_sas_slot *slot) dev_info(dev, "data underflow without sense, rsp_code:0x%x, rc:%d.\n", iu->resp_data[0], rc); } - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } -- 2.7.4

1 0

【Meeting Notice】Kernel & testing sync between Linaro and openEuler Time: 2020-06-10 16:00-17:30
by Meeting Book 03 Jun '20

03 Jun '20

1 0

【Meeting Notice】Kernel & testing sync between Linaro and openEuler Time: 2020-06-10 16:00-17:30
by Meeting Book 03 Jun '20

03 Jun '20

1 0

[PATCH 01/76] net: dsa: Do not make user port errors fatal
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Florian Fainelli <f.fainelli(a)gmail.com> commit 86f8b1c01a0a537a73d2996615133be63cdf75db upstream. Prior to 1d27732f411d ("net: dsa: setup and teardown ports"), we would not treat failures to set-up an user port as fatal, but after this commit we would, which is a regression for some systems where interfaces may be declared in the Device Tree, but the underlying hardware may not be present (pluggable daughter cards for instance). Fixes: 1d27732f411d ("net: dsa: setup and teardown ports") Signed-off-by: Florian Fainelli <f.fainelli(a)gmail.com> Reviewed-by: Andrew Lunn <andrew(a)lunn.ch> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- net/dsa/dsa2.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/dsa/dsa2.c b/net/dsa/dsa2.c index 46ae4de..b036c55 100644 --- a/net/dsa/dsa2.c +++ b/net/dsa/dsa2.c @@ -412,7 +412,7 @@ static int dsa_tree_setup_switches(struct dsa_switch_tree *dst) err = dsa_switch_setup(ds); if (err) - return err; + continue; for (port = 0; port < ds->num_ports; port++) { dp = &ds->ports[port]; -- 1.8.3

1 75

[PATCH 1/9] net/hinic: Fix the firmware compatibility bug in the MAC reuse scenario
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Chiqijun <chiqijun(a)huawei.com> driver inclusion category: bugfix bugzilla: 4472 ----------------------------------------------------------------------- 2.3.0.1 and later drivers and firmware add MAC reuse feature, when the driver is compatible with firmware version 2.3.0.0 and before, if the ip link command is used to configure the MAC and VLAN for the VF, it will cause the VF network anomaly. To solve this problem, the driver obtains the function capability from the firmware when loading, and determines whether to support MAC reuse for different processing. Signed-off-by: Chiqijun <chiqijun(a)huawei.com> Reviewed-by: Zengweiliang <zengweiliang.zengweiliang(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- drivers/net/ethernet/huawei/hinic/hinic_hwdev.h | 1 + .../ethernet/huawei/hinic/hinic_mgmt_interface.h | 11 ++++ drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c | 45 +++++++++++++++-- drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h | 4 ++ drivers/net/ethernet/huawei/hinic/hinic_nic_io.c | 10 +++- drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h | 2 + drivers/net/ethernet/huawei/hinic/hinic_sriov.c | 59 ++++++++++++++-------- 7 files changed, 106 insertions(+), 26 deletions(-) diff --git a/drivers/net/ethernet/huawei/hinic/hinic_hwdev.h b/drivers/net/ethernet/huawei/hinic/hinic_hwdev.h index a096f91..2a81867 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_hwdev.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_hwdev.h @@ -319,6 +319,7 @@ struct hinic_hwdev { struct hinic_board_info board_info; #define MGMT_VERSION_MAX_LEN 32 u8 mgmt_ver[MGMT_VERSION_MAX_LEN]; + u64 fw_support_func_flag; }; int hinic_init_comm_ch(struct hinic_hwdev *hwdev); diff --git a/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h b/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h index 3c3be6e..8673cf5 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h @@ -690,6 +690,15 @@ struct hinic_port_rt_cmd { u8 rsvd1[6]; }; +struct fw_support_func { + u8 status; + u8 version; + u8 rsvd0[6]; + + u64 flag; + u64 rsvd; +}; + struct hinic_vf_dcb_state { u8 status; u8 version; @@ -741,6 +750,8 @@ struct hinic_set_link_follow { int hinic_get_base_qpn(void *hwdev, u16 *global_qpn); +int hinic_get_fw_support_func(void *hwdev); + int hinic_vf_func_init(struct hinic_hwdev *hwdev); void hinic_vf_func_free(struct hinic_hwdev *hwdev); diff --git a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c index f9e16ec..7dfedbd 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c @@ -201,6 +201,34 @@ int hinic_get_base_qpn(void *hwdev, u16 *global_qpn) return 0; } +int hinic_get_fw_support_func(void *hwdev) +{ + struct fw_support_func support_flag = {0}; + struct hinic_hwdev *dev = hwdev; + u16 out_size = sizeof(support_flag); + int err; + + if (!hwdev) + return -EINVAL; + + err = hinic_msg_to_mgmt_sync(hwdev, HINIC_MOD_L2NIC, + HINIC_PORT_CMD_GET_FW_SUPPORT_FLAG, + &support_flag, sizeof(support_flag), + &support_flag, &out_size, 0); + if (support_flag.status == HINIC_MGMT_CMD_UNSUPPORTED) { + nic_info(dev->dev_hdl, "Current firmware doesn't support to get function capability\n"); + support_flag.flag = 0; + } else if (support_flag.status || err || !out_size) { + nic_err(dev->dev_hdl, "Failed to get function capability, err: %d, status: 0x%x, out size: 0x%x\n", + err, support_flag.status, out_size); + return -EFAULT; + } + + dev->fw_support_func_flag = support_flag.flag; + + return 0; +} + #define HINIC_ADD_VLAN_IN_MAC 0x8000 #define HINIC_VLAN_ID_MASK 0x7FFF @@ -352,6 +380,11 @@ int hinic_update_mac_vlan(void *hwdev, u16 old_vlan, u16 new_vlan, int vf_id) if (!vf_info->pf_set_mac) return 0; + if (!FW_SUPPORT_MAC_REUSE_FUNC(dev)) { + nic_info(dev->dev_hdl, "Current firmware doesn't support mac reuse\n"); + return 0; + } + func_id = hinic_glb_pf_vf_offset(dev) + (u16)vf_id; vlan_id = old_vlan; if (vlan_id) @@ -2562,9 +2595,14 @@ static int hinic_init_vf_config(struct hinic_hwdev *hwdev, u16 vf_id) vf_info = hwdev->nic_io->vf_infos + HW_VF_ID_TO_OS(vf_id); if (vf_info->pf_set_mac) { func_id = hinic_glb_pf_vf_offset(hwdev) + vf_id; - vlan_id = vf_info->pf_vlan; - if (vlan_id) - vlan_id |= HINIC_ADD_VLAN_IN_MAC; + if (FW_SUPPORT_MAC_REUSE_FUNC(hwdev)) { + vlan_id = vf_info->pf_vlan; + if (vlan_id) + vlan_id |= HINIC_ADD_VLAN_IN_MAC; + } else { + vlan_id = 0; + } + err = hinic_set_mac(hwdev, vf_info->vf_mac_addr, vlan_id, func_id); if (err) { @@ -4065,3 +4103,4 @@ int hinic_get_sfp_type(void *hwdev, u8 *data0, u8 *data1) return 0; } + diff --git a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h index 427d469..e0979df 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h @@ -19,6 +19,10 @@ #define OS_VF_ID_TO_HW(os_vf_id) ((os_vf_id) + 1) #define HW_VF_ID_TO_OS(hw_vf_id) ((hw_vf_id) - 1) +#define FW_SUPPORT_MAC_REUSE 0x1 +#define FW_SUPPORT_MAC_REUSE_FUNC(hwdev) \ + ((hwdev)->fw_support_func_flag & FW_SUPPORT_MAC_REUSE) + #define HINIC_VLAN_PRIORITY_SHIFT 13 #define HINIC_RSS_INDIR_SIZE 256 diff --git a/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c b/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c index 1cb4e26..f2fb0bc 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c @@ -831,9 +831,17 @@ int hinic_init_nic_hwdev(void *hwdev, u16 rx_buff_len) } /* VFs don't set port routine command report */ - if (hinic_func_type(dev) != TYPE_VF) + if (hinic_func_type(dev) != TYPE_VF) { + /* Get the fw support mac reuse flag */ + err = hinic_get_fw_support_func(hwdev); + if (err) { + nic_err(dev->dev_hdl, "Failed to get function capability\n"); + return err; + } + /* Inform mgmt to send sfp's information to driver */ err = hinic_set_port_routine_cmd_report(hwdev, true); + } return err; } diff --git a/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h b/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h index 46464b1f..d428f48 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h @@ -146,6 +146,8 @@ enum hinic_port_cmd { HINIC_PORT_CMD_SET_PORT_LINK_STATUS = 0x76, HINIC_PORT_CMD_SET_CGE_PAUSE_TIME_CFG = 0x77, + HINIC_PORT_CMD_GET_FW_SUPPORT_FLAG = 0x79, + HINIC_PORT_CMD_SET_PORT_REPORT = 0x7B, HINIC_PORT_CMD_LINK_STATUS_REPORT = 0xa0, diff --git a/drivers/net/ethernet/huawei/hinic/hinic_sriov.c b/drivers/net/ethernet/huawei/hinic/hinic_sriov.c index 37d0116..73ab2d7 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_sriov.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_sriov.c @@ -266,6 +266,42 @@ int hinic_ndo_set_vf_mac(struct net_device *netdev, int vf, u8 *mac) /*lint -save -e574 -e734*/ #ifdef IFLA_VF_MAX +static int set_hw_vf_vlan(struct hinic_sriov_info *sriov_info, + u16 cur_vlanprio, int vf, u16 vlan, u8 qos) +{ + int err = 0; + u16 old_vlan = cur_vlanprio & VLAN_VID_MASK; + + if (vlan || qos) { + if (cur_vlanprio) { + err = hinic_kill_vf_vlan(sriov_info->hwdev, + OS_VF_ID_TO_HW(vf)); + if (err) { + nic_err(&sriov_info->pdev->dev, "Failed to delete vf %d old vlan %d\n", + vf, old_vlan); + return err; + } + } + err = hinic_add_vf_vlan(sriov_info->hwdev, + OS_VF_ID_TO_HW(vf), vlan, qos); + if (err) { + nic_err(&sriov_info->pdev->dev, "Failed to add vf %d new vlan %d\n", + vf, vlan); + return err; + } + } else { + err = hinic_kill_vf_vlan(sriov_info->hwdev, OS_VF_ID_TO_HW(vf)); + if (err) { + nic_err(&sriov_info->pdev->dev, "Failed to delete vf %d vlan %d\n", + vf, old_vlan); + return err; + } + } + + return hinic_update_mac_vlan(sriov_info->hwdev, old_vlan, vlan, + OS_VF_ID_TO_HW(vf)); +} + #ifdef IFLA_VF_VLAN_INFO_MAX int hinic_ndo_set_vf_vlan(struct net_device *netdev, int vf, u16 vlan, u8 qos, __be16 vlan_proto) @@ -276,7 +312,6 @@ int hinic_ndo_set_vf_vlan(struct net_device *netdev, int vf, u16 vlan, u8 qos) struct hinic_nic_dev *adapter = netdev_priv(netdev); struct hinic_sriov_info *sriov_info; u16 vlanprio, cur_vlanprio; - int err = 0; if (!FUNC_SUPPORT_SET_VF_MAC_VLAN(adapter->hwdev)) { nicif_err(adapter, drv, netdev, @@ -298,27 +333,7 @@ int hinic_ndo_set_vf_vlan(struct net_device *netdev, int vf, u16 vlan, u8 qos) if (vlanprio == cur_vlanprio) return 0; - if (vlan || qos) { - if (cur_vlanprio) - err = hinic_kill_vf_vlan(sriov_info->hwdev, - OS_VF_ID_TO_HW(vf)); - if (err) - goto out; - err = hinic_add_vf_vlan(sriov_info->hwdev, OS_VF_ID_TO_HW(vf), - vlan, qos); - } else { - err = hinic_kill_vf_vlan(sriov_info->hwdev, OS_VF_ID_TO_HW(vf)); - } - - if (err) - return err; - - err = hinic_update_mac_vlan(sriov_info->hwdev, - cur_vlanprio & VLAN_VID_MASK, vlan, - OS_VF_ID_TO_HW(vf)); - -out: - return err; + return set_hw_vf_vlan(sriov_info, cur_vlanprio, vf, vlan, qos); } #endif -- 1.8.3

1 8

[PATCH] irqchip/gic-v3-its: Probe ITS page size for all GITS_BASERn registers
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Marc Zyngier <maz(a)kernel.org> mainline inclusion from mainline-v5.6-rc4 commit d5df9dc96eb7423d3f742b13d5e1e479ff795eaa category: bugfix bugzilla: NA CVE: NA --------------------------- The GICv3 ITS driver assumes that once it has latched on a page size for a given BASER register, it can use the same page size as the maximum page size for all subsequent BASER registers. Although it worked so far, nothing in the architecture guarantees this, and Nianyao Tang hit this problem on some undisclosed implementation. Let's bite the bullet and probe the the supported page size on all BASER registers before starting to populate the tables. This simplifies the setup a bit, at the expense of a few additional MMIO accesses. Signed-off-by: Marc Zyngier <maz(a)kernel.org> Reported-by: Nianyao Tang <tangnianyao(a)huawei.com> Tested-by: Nianyao Tang <tangnianyao(a)huawei.com> Link: https://lore.kernel.org/r/1584089195-63897-1-git-send-email-zhangshaokun@hi… Signed-off-by: Hongbo Yao <yaohongbo(a)huawei.com> conflicts: drivers/irqchip/irq-gic-v3-its.c Reviewed-by: Hanjun Guo <guohanjun(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- drivers/irqchip/irq-gic-v3-its.c | 101 ++++++++++++++++++++++++++------------- 1 file changed, 67 insertions(+), 34 deletions(-) diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c index 59c6c0d..bfedab2 100644 --- a/drivers/irqchip/irq-gic-v3-its.c +++ b/drivers/irqchip/irq-gic-v3-its.c @@ -17,6 +17,7 @@ #include <linux/acpi.h> #include <linux/acpi_iort.h> +#include <linux/bitfield.h> #include <linux/bitmap.h> #include <linux/cpu.h> #include <linux/crash_dump.h> @@ -1768,17 +1769,16 @@ static void its_write_baser(struct its_node *its, struct its_baser *baser, } static int its_setup_baser(struct its_node *its, struct its_baser *baser, - u64 cache, u64 shr, u32 psz, u32 order, - bool indirect) + u64 cache, u64 shr, u32 order, bool indirect) { u64 val = its_read_baser(its, baser); u64 esz = GITS_BASER_ENTRY_SIZE(val); u64 type = GITS_BASER_TYPE(val); u64 baser_phys, tmp; - u32 alloc_pages; + u32 alloc_pages, psz; void *base; -retry_alloc_baser: + psz = baser->psz; alloc_pages = (PAGE_ORDER_TO_SIZE(order) / psz); if (alloc_pages > GITS_BASER_PAGES_MAX) { pr_warn("ITS@%pa: %s too large, reduce ITS pages %u->%u\n", @@ -1851,25 +1851,6 @@ static int its_setup_baser(struct its_node *its, struct its_baser *baser, goto retry_baser; } - if ((val ^ tmp) & GITS_BASER_PAGE_SIZE_MASK) { - /* - * Page size didn't stick. Let's try a smaller - * size and retry. If we reach 4K, then - * something is horribly wrong... - */ - free_pages((unsigned long)base, order); - baser->base = NULL; - - switch (psz) { - case SZ_16K: - psz = SZ_4K; - goto retry_alloc_baser; - case SZ_64K: - psz = SZ_16K; - goto retry_alloc_baser; - } - } - if (val != tmp) { pr_err("ITS@%pa: %s doesn't stick: %llx %llx\n", &its->phys_base, its_base_type_string[type], @@ -1895,13 +1876,14 @@ static int its_setup_baser(struct its_node *its, struct its_baser *baser, static bool its_parse_indirect_baser(struct its_node *its, struct its_baser *baser, - u32 psz, u32 *order, u32 ids) + u32 *order, u32 ids) { u64 tmp = its_read_baser(its, baser); u64 type = GITS_BASER_TYPE(tmp); u64 esz = GITS_BASER_ENTRY_SIZE(tmp); u64 val = GITS_BASER_InnerShareable | GITS_BASER_RaWaWb; u32 new_order = *order; + u32 psz = baser->psz; bool indirect = false; /* No need to enable Indirection if memory requirement < (psz*2)bytes */ @@ -1960,11 +1942,58 @@ static void its_free_tables(struct its_node *its) } } +static int its_probe_baser_psz(struct its_node *its, struct its_baser *baser) +{ + u64 psz = SZ_64K; + + while (psz) { + u64 val, gpsz; + + val = its_read_baser(its, baser); + val &= ~GITS_BASER_PAGE_SIZE_MASK; + + switch (psz) { + case SZ_64K: + gpsz = GITS_BASER_PAGE_SIZE_64K; + break; + case SZ_16K: + gpsz = GITS_BASER_PAGE_SIZE_16K; + break; + case SZ_4K: + default: + gpsz = GITS_BASER_PAGE_SIZE_4K; + break; + } + + gpsz >>= GITS_BASER_PAGE_SIZE_SHIFT; + + val |= FIELD_PREP(GITS_BASER_PAGE_SIZE_MASK, gpsz); + its_write_baser(its, baser, val); + + if (FIELD_GET(GITS_BASER_PAGE_SIZE_MASK, baser->val) == gpsz) + break; + + switch (psz) { + case SZ_64K: + psz = SZ_16K; + break; + case SZ_16K: + psz = SZ_4K; + break; + case SZ_4K: + default: + return -1; + } + } + + baser->psz = psz; + return 0; +} + static int its_alloc_tables(struct its_node *its) { u64 shr = GITS_BASER_InnerShareable; u64 cache = GITS_BASER_RaWaWb; - u32 psz = SZ_64K; int err, i; if (its->flags & ITS_FLAGS_WORKAROUND_CAVIUM_22375) @@ -1975,34 +2004,38 @@ static int its_alloc_tables(struct its_node *its) struct its_baser *baser = its->tables + i; u64 val = its_read_baser(its, baser); u64 type = GITS_BASER_TYPE(val); - u32 order = get_order(psz); bool indirect = false; + u32 order; - switch (type) { - case GITS_BASER_TYPE_NONE: + if (type == GITS_BASER_TYPE_NONE) continue; + if (its_probe_baser_psz(its, baser)) { + its_free_tables(its); + return -ENXIO; + } + + order = get_order(baser->psz); + + switch (type) { case GITS_BASER_TYPE_DEVICE: - indirect = its_parse_indirect_baser(its, baser, - psz, &order, + indirect = its_parse_indirect_baser(its, baser, &order, its->device_ids); break; case GITS_BASER_TYPE_VCPU: - indirect = its_parse_indirect_baser(its, baser, - psz, &order, + indirect = its_parse_indirect_baser(its, baser, &order, ITS_MAX_VPEID_BITS); break; } - err = its_setup_baser(its, baser, cache, shr, psz, order, indirect); + err = its_setup_baser(its, baser, cache, shr, order, indirect); if (err < 0) { its_free_tables(its); return err; } /* Update settings which will be used for next BASERn */ - psz = baser->psz; cache = baser->val & GITS_BASER_CACHEABILITY_MASK; shr = baser->val & GITS_BASER_SHAREABILITY_MASK; } -- 1.8.3

1 0

[PATCH 1/4] vfio/mdev: Improve the create/remove sequence
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Parav Pandit <parav(a)mellanox.com> mainline inclusion from mainline-5.2 commit 522ecce08ab20b57342d65b05601818e0f95fb2c category: bugfix bugzilla: NA CVE: NA ------------------------------------------------- This patch addresses below two issues and prepares the code to address 3rd issue listed below. 1. mdev device is placed on the mdev bus before it is created in the vendor driver. Once a device is placed on the mdev bus without creating its supporting underlying vendor device, mdev driver's probe() gets triggered. However there isn't a stable mdev available to work on. create_store() mdev_create_device() device_register() ... vfio_mdev_probe() [...] parent->ops->create() vfio_ap_mdev_create() mdev_set_drvdata(mdev, matrix_mdev); /* Valid pointer set above */ Due to this way of initialization, mdev driver who wants to use the mdev, doesn't have a valid mdev to work on. 2. Current creation sequence is, parent->ops_create() groups_register() Remove sequence is, parent->ops->remove() groups_unregister() However, remove sequence should be exact mirror of creation sequence. Once this is achieved, all users of the mdev will be terminated first before removing underlying vendor device. (Follow standard linux driver model). At that point vendor's remove() ops shouldn't fail because taking the device off the bus should terminate any usage. 3. When remove operation fails, mdev sysfs removal attempts to add the file back on already removed device. Following call trace [1] is observed. [1] call trace: kernel: WARNING: CPU: 2 PID: 9348 at fs/sysfs/file.c:327 sysfs_create_file_ns+0x7f/0x90 kernel: CPU: 2 PID: 9348 Comm: bash Kdump: loaded Not tainted 5.1.0-rc6-vdevbus+ #6 kernel: Hardware name: Supermicro SYS-6028U-TR4+/X10DRU-i+, BIOS 2.0b 08/09/2016 kernel: RIP: 0010:sysfs_create_file_ns+0x7f/0x90 kernel: Call Trace: kernel: remove_store+0xdc/0x100 [mdev] kernel: kernfs_fop_write+0x113/0x1a0 kernel: vfs_write+0xad/0x1b0 kernel: ksys_write+0x5a/0xe0 kernel: do_syscall_64+0x5a/0x210 kernel: entry_SYSCALL_64_after_hwframe+0x49/0xbe Therefore, mdev core is improved in following ways. 1. Split the device registration/deregistration sequence so that some things can be done between initialization of the device and hooking it up to the bus respectively after deregistering it from the bus but before giving up our final reference. In particular, this means invoking the ->create() and ->remove() callbacks in those new windows. This gives the vendor driver an initialized mdev device to work with during creation. At the same time, a bus driver who wish to bind to mdev driver also gets initialized mdev device. This follows standard Linux kernel bus and device model. 2. During remove flow, first remove the device from the bus. This ensures that any bus specific devices are removed. Once device is taken off the mdev bus, invoke remove() of mdev from the vendor driver. 3. The driver core device model provides way to register and auto unregister the device sysfs attribute groups at dev->groups. Make use of dev->groups to let core create the groups and eliminate code to avoid explicit groups creation and removal. To ensure, that new sequence is solid, a below stack dump of a process is taken who attempts to remove the device while device is in use by vfio driver and user application. This stack dump validates that vfio driver guards against such device removal when device is in use. cat /proc/21962/stack [<0>] vfio_del_group_dev+0x216/0x3c0 [vfio] [<0>] mdev_remove+0x21/0x40 [mdev] [<0>] device_release_driver_internal+0xe8/0x1b0 [<0>] bus_remove_device+0xf9/0x170 [<0>] device_del+0x168/0x350 [<0>] mdev_device_remove_common+0x1d/0x50 [mdev] [<0>] mdev_device_remove+0x8c/0xd0 [mdev] [<0>] remove_store+0x71/0x90 [mdev] [<0>] kernfs_fop_write+0x113/0x1a0 [<0>] vfs_write+0xad/0x1b0 [<0>] ksys_write+0x5a/0xe0 [<0>] do_syscall_64+0x5a/0x210 [<0>] entry_SYSCALL_64_after_hwframe+0x49/0xbe [<0>] 0xffffffffffffffff This prepares the code to eliminate calling device_create_file() in subsequent patch. Reviewed-by: Cornelia Huck <cohuck(a)redhat.com> Signed-off-by: Parav Pandit <parav(a)mellanox.com> Signed-off-by: Alex Williamson <alex.williamson(a)redhat.com> Conflicts: drivers/vfio/mdev/mdev_core.c -mdev_device_remove_ops() and -mdev_device_remove_cb() are removed anyway, although they are not identical compared within the original patch. The type of uuid is still uuid_le, and it has not been changed to guid_t. drivers/vfio/mdev/mdev_private.h The type of uuid is still uuid_le, and it has not been changed to guid_t. Signed-off-by: Xiaoyang Xu <xuxiaoyang2(a)huawei.com> Reviewed-by: Yingtai Xie <xieyingtai(a)huawei.com> Reviewed-by: Hanjun Guo <guohanjun(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- drivers/vfio/mdev/mdev_core.c | 94 +++++++++++----------------------------- drivers/vfio/mdev/mdev_private.h | 2 +- drivers/vfio/mdev/mdev_sysfs.c | 2 +- 3 files changed, 27 insertions(+), 71 deletions(-) diff --git a/drivers/vfio/mdev/mdev_core.c b/drivers/vfio/mdev/mdev_core.c index e052f62..9705a4a 100644 --- a/drivers/vfio/mdev/mdev_core.c +++ b/drivers/vfio/mdev/mdev_core.c @@ -103,55 +103,10 @@ static inline void mdev_put_parent(struct mdev_parent *parent) kref_put(&parent->ref, mdev_release_parent); } -static int mdev_device_create_ops(struct kobject *kobj, - struct mdev_device *mdev) -{ - struct mdev_parent *parent = mdev->parent; - int ret; - - ret = parent->ops->create(kobj, mdev); - if (ret) - return ret; - - ret = sysfs_create_groups(&mdev->dev.kobj, - parent->ops->mdev_attr_groups); - if (ret) - parent->ops->remove(mdev); - - return ret; -} - -/* - * mdev_device_remove_ops gets called from sysfs's 'remove' and when parent - * device is being unregistered from mdev device framework. - * - 'force_remove' is set to 'false' when called from sysfs's 'remove' which - * indicates that if the mdev device is active, used by VMM or userspace - * application, vendor driver could return error then don't remove the device. - * - 'force_remove' is set to 'true' when called from mdev_unregister_device() - * which indicate that parent device is being removed from mdev device - * framework so remove mdev device forcefully. - */ -static int mdev_device_remove_ops(struct mdev_device *mdev, bool force_remove) -{ - struct mdev_parent *parent = mdev->parent; - int ret; - - /* - * Vendor driver can return error if VMM or userspace application is - * using this mdev device. - */ - ret = parent->ops->remove(mdev); - if (ret && !force_remove) - return -EBUSY; - - sysfs_remove_groups(&mdev->dev.kobj, parent->ops->mdev_attr_groups); - return 0; -} - static int mdev_device_remove_cb(struct device *dev, void *data) { if (dev_is_mdev(dev)) - mdev_device_remove(dev, true); + mdev_device_remove(dev); return 0; } @@ -311,41 +266,43 @@ int mdev_device_create(struct kobject *kobj, struct device *dev, uuid_le uuid) mdev->parent = parent; kref_init(&mdev->ref); + device_initialize(&mdev->dev); mdev->dev.parent = dev; mdev->dev.bus = &mdev_bus_type; mdev->dev.release = mdev_device_release; dev_set_name(&mdev->dev, "%pUl", uuid.b); + mdev->dev.groups = parent->ops->mdev_attr_groups; + mdev->type_kobj = kobj; - ret = device_register(&mdev->dev); - if (ret) { - put_device(&mdev->dev); - goto mdev_fail; - } + ret = parent->ops->create(kobj, mdev); + if (ret) + goto ops_create_fail; - ret = mdev_device_create_ops(kobj, mdev); + ret = device_add(&mdev->dev); if (ret) - goto create_fail; + goto add_fail; ret = mdev_create_sysfs_files(&mdev->dev, type); - if (ret) { - mdev_device_remove_ops(mdev, true); - goto create_fail; - } + if (ret) + goto sysfs_fail; - mdev->type_kobj = kobj; mdev->active = true; dev_dbg(&mdev->dev, "MDEV: created\n"); return 0; -create_fail: - device_unregister(&mdev->dev); +sysfs_fail: + device_del(&mdev->dev); +add_fail: + parent->ops->remove(mdev); +ops_create_fail: + put_device(&mdev->dev); mdev_fail: mdev_put_parent(parent); return ret; } -int mdev_device_remove(struct device *dev, bool force_remove) +int mdev_device_remove(struct device *dev) { struct mdev_device *mdev, *tmp; struct mdev_parent *parent; @@ -374,16 +331,15 @@ int mdev_device_remove(struct device *dev, bool force_remove) mutex_unlock(&mdev_list_lock); type = to_mdev_type(mdev->type_kobj); + mdev_remove_sysfs_files(dev, type); + device_del(&mdev->dev); parent = mdev->parent; + ret = parent->ops->remove(mdev); + if (ret) + dev_err(&mdev->dev, "Remove failed: err=%d\n", ret); - ret = mdev_device_remove_ops(mdev, force_remove); - if (ret) { - mdev->active = true; - return ret; - } - - mdev_remove_sysfs_files(dev, type); - device_unregister(dev); + /* Balances with device_initialize() */ + put_device(&mdev->dev); mdev_put_parent(parent); return 0; diff --git a/drivers/vfio/mdev/mdev_private.h b/drivers/vfio/mdev/mdev_private.h index b5819b7..9c65d8e 100644 --- a/drivers/vfio/mdev/mdev_private.h +++ b/drivers/vfio/mdev/mdev_private.h @@ -59,6 +59,6 @@ struct mdev_type { void mdev_remove_sysfs_files(struct device *dev, struct mdev_type *type); int mdev_device_create(struct kobject *kobj, struct device *dev, uuid_le uuid); -int mdev_device_remove(struct device *dev, bool force_remove); +int mdev_device_remove(struct device *dev); #endif /* MDEV_PRIVATE_H */ diff --git a/drivers/vfio/mdev/mdev_sysfs.c b/drivers/vfio/mdev/mdev_sysfs.c index e7770b5..1286366 100644 --- a/drivers/vfio/mdev/mdev_sysfs.c +++ b/drivers/vfio/mdev/mdev_sysfs.c @@ -236,7 +236,7 @@ static ssize_t remove_store(struct device *dev, struct device_attribute *attr, if (val && device_remove_file_self(dev, attr)) { int ret; - ret = mdev_device_remove(dev, false); + ret = mdev_device_remove(dev); if (ret) { device_create_file(dev, attr); return ret; -- 1.8.3

1 3

[PATCH 1/2] arm64/mpam: Fix unreset resources when mkdir ctrl group or umount resctrl
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Wang ShaoBo <bobo.shaobowang(a)huawei.com> hulk inclusion category: bugfix bugzilla: 34588 CVE: NA ----------------------------------------------- There are two problems related to schemata: 1) When rmdir a group and then mkdir a new group under resctrl root directory, the new group still inherits the schemata configuration from old. e.g. > mount -t resctrl resctrl /sys/fs/resctrl > cd /sys/fs/resctrl > mkdir p1 && cd p1 > echo 'L3:0=7f' > schemata > cd .. && rmdir p1 && mkdir p1 && cd p1 > cat schemata L3:0=7f;1=7fff;2=7fff;3=7fff MB:0=100;1=100;2=100;3=100 2) It still exists when umount /sys/fs/resctrl and remount. e.g. > mount -t resctrl resctrl /sys/fs/resctrl > cd /sys/fs/resctrl > echo 'L3:0=7f' > schemata > umount /sys/fs/resctrl > mount -t resctrl resctrl /sys/fs/resctrl > cat schemata L3:0=7f;1=7fff;2=7fff;3=7fff MB:0=100;1=100;2=100;3=100 Firstly we make each resctrl resource obtains their corresponding default configuration. NOTE we use zero to initialize L3 default value instead of max cpbm bits, as zero configurarion equals to maximum configuration for L3 MSCs. And we use max-percentage masks of max bandwidth to generate maximum configuration for MB. Then we reset resources' configuration settings to default value and back MSCs to default state, when mkdir or umount happended. Fixes: caf75b6b2540 ("resctrlfs: mpam: init struct for mpam") Fixes: 916dd9321e3c ("resctrlfs: init support resctrlfs") Signed-off-by: Wang ShaoBo <bobo.shaobowang(a)huawei.com> Reviewed-by: Xie XiuQi <xiexiuqi(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- arch/arm64/include/asm/mpam_resource.h | 4 ++++ arch/arm64/include/asm/resctrl.h | 2 ++ arch/arm64/kernel/mpam.c | 17 +++++++++++++++++ arch/arm64/kernel/mpam_ctrlmon.c | 28 ++++++++++++++++++++++++++++ fs/resctrlfs.c | 5 +++++ 5 files changed, 56 insertions(+) diff --git a/arch/arm64/include/asm/mpam_resource.h b/arch/arm64/include/asm/mpam_resource.h index ab90596..1a6904c 100644 --- a/arch/arm64/include/asm/mpam_resource.h +++ b/arch/arm64/include/asm/mpam_resource.h @@ -89,6 +89,10 @@ /* [FIXME] hard code for hardlim */ #define MBW_MAX_SET(v) (MBW_MAX_HARDLIM|((v) << (16 - BWA_WD))) #define MBW_MAX_GET(v) (((v) & MBW_MAX_MASK) >> (16 - BWA_WD)) + +/* hard code for mbw_max max-percentage's cresponding masks */ +#define MBA_MAX_WD 63u + /* * emulate the mpam nodes * These should be reported by ACPI MPAM Table. diff --git a/arch/arm64/include/asm/resctrl.h b/arch/arm64/include/asm/resctrl.h index 0a0a12b..fb5fa6c 100644 --- a/arch/arm64/include/asm/resctrl.h +++ b/arch/arm64/include/asm/resctrl.h @@ -65,4 +65,6 @@ int mkdir_mondata_all(struct kernfs_node *parent_kn, mongroup_create_dir(struct kernfs_node *parent_kn, struct resctrl_group *prgrp, char *name, struct kernfs_node **dest_kn); +int rdtgroup_init_alloc(struct rdtgroup *rdtgrp); + #endif /* _ASM_ARM64_RESCTRL_H */ diff --git a/arch/arm64/kernel/mpam.c b/arch/arm64/kernel/mpam.c index 202e49a..120795e 100644 --- a/arch/arm64/kernel/mpam.c +++ b/arch/arm64/kernel/mpam.c @@ -86,6 +86,9 @@ static inline void mpam_node_assign_val(struct mpam_node *n, n->addr = hwpage_address; n->component_id = component_id; n->cpus_list = "0"; + + if (n->type == MPAM_RESOURCE_MC) + n->default_ctrl = MBA_MAX_WD; } #define MPAM_NODE_NAME_SIZE (10) @@ -544,6 +547,20 @@ void post_resctrl_mount(void) static int reset_all_ctrls(struct resctrl_resource *r) { + struct raw_resctrl_resource *rr; + struct rdt_domain *d; + int partid; + + rr = (struct raw_resctrl_resource *)r->res; + for (partid = 0; partid < rr->num_partid; partid++) { + list_for_each_entry(d, &r->domains, list) { + d->new_ctrl = rr->default_ctrl; + d->ctrl_val[partid] = rr->default_ctrl; + d->have_new_ctrl = true; + rr->msr_update(d, partid); + } + } + return 0; } diff --git a/arch/arm64/kernel/mpam_ctrlmon.c b/arch/arm64/kernel/mpam_ctrlmon.c index b9f9495..2b94efc 100644 --- a/arch/arm64/kernel/mpam_ctrlmon.c +++ b/arch/arm64/kernel/mpam_ctrlmon.c @@ -585,3 +585,31 @@ int resctrl_mkdir_ctrlmon_mondata(struct kernfs_node *parent_kn, } return ret; } + +/* Initialize the RDT group's allocations. */ +int rdtgroup_init_alloc(struct rdtgroup *rdtgrp) +{ + struct resctrl_resource *r; + struct raw_resctrl_resource *rr; + struct rdt_domain *d; + int ret; + + for_each_resctrl_resource(r) { + if (!r->alloc_enabled) + continue; + + rr = (struct raw_resctrl_resource *)r->res; + list_for_each_entry(d, &r->domains, list) { + d->new_ctrl = rr->default_ctrl; + d->have_new_ctrl = true; + } + + ret = update_domains(r, rdtgrp); + if (ret < 0) { + rdt_last_cmd_puts("Failed to initialize allocations\n"); + return ret; + } + } + + return 0; +} diff --git a/fs/resctrlfs.c b/fs/resctrlfs.c index 09e56384..8d463b8 100644 --- a/fs/resctrlfs.c +++ b/fs/resctrlfs.c @@ -670,6 +670,11 @@ static int resctrl_group_mkdir_ctrl_mon(struct kernfs_node *parent_kn, ret = 0; rdtgrp->closid = closid; + + ret = rdtgroup_init_alloc(rdtgrp); + if (ret < 0) + goto out_id_free; + list_add(&rdtgrp->resctrl_group_list, &resctrl_all_groups); if (resctrl_mon_capable) { -- 1.8.3

1 1