diff options
23 files changed, 250 insertions, 85 deletions
diff --git a/Documentation/bpf/btf.rst b/Documentation/bpf/btf.rst index ffc11afee569..e43c2fdafcd7 100644 --- a/Documentation/bpf/btf.rst +++ b/Documentation/bpf/btf.rst @@ -803,7 +803,7 @@ structure when .BTF.ext is generated. All ``bpf_core_relo`` structures within a single ``btf_ext_info_sec`` describe relocations applied to section named by ``btf_ext_info_sec->sec_name_off``. -See :ref:`Documentation/bpf/llvm_reloc <btf-co-re-relocations>` +See :ref:`Documentation/bpf/llvm_reloc.rst <btf-co-re-relocations>` for more information on CO-RE relocations. 4.2 .BTF_ids section diff --git a/Documentation/bpf/llvm_reloc.rst b/Documentation/bpf/llvm_reloc.rst index 73bf805000f2..44188e219d32 100644 --- a/Documentation/bpf/llvm_reloc.rst +++ b/Documentation/bpf/llvm_reloc.rst @@ -250,7 +250,7 @@ CO-RE Relocations From object file point of view CO-RE mechanism is implemented as a set of CO-RE specific relocation records. These relocation records are not related to ELF relocations and are encoded in .BTF.ext section. -See :ref:`Documentation/bpf/btf <BTF_Ext_Section>` for more +See :ref:`Documentation/bpf/btf.rst <BTF_Ext_Section>` for more information on .BTF.ext structure. CO-RE relocations are applied to BPF instructions to update immediate diff --git a/arch/s390/net/bpf_jit_comp.c b/arch/s390/net/bpf_jit_comp.c index 5e9371fbf3d5..de2fb12120d2 100644 --- a/arch/s390/net/bpf_jit_comp.c +++ b/arch/s390/net/bpf_jit_comp.c @@ -2088,6 +2088,7 @@ struct bpf_tramp_jit { */ int r14_off; /* Offset of saved %r14 */ int run_ctx_off; /* Offset of struct bpf_tramp_run_ctx */ + int tccnt_off; /* Offset of saved tailcall counter */ int do_fexit; /* do_fexit: label */ }; @@ -2258,12 +2259,16 @@ static int __arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, tjit->r14_off = alloc_stack(tjit, sizeof(u64)); tjit->run_ctx_off = alloc_stack(tjit, sizeof(struct bpf_tramp_run_ctx)); + tjit->tccnt_off = alloc_stack(tjit, sizeof(u64)); /* The caller has already reserved STACK_FRAME_OVERHEAD bytes. */ tjit->stack_size -= STACK_FRAME_OVERHEAD; tjit->orig_stack_args_off = tjit->stack_size + STACK_FRAME_OVERHEAD; /* aghi %r15,-stack_size */ EMIT4_IMM(0xa70b0000, REG_15, -tjit->stack_size); + /* mvc tccnt_off(4,%r15),stack_size+STK_OFF_TCCNT(%r15) */ + _EMIT6(0xd203f000 | tjit->tccnt_off, + 0xf000 | (tjit->stack_size + STK_OFF_TCCNT)); /* stmg %r2,%rN,fwd_reg_args_off(%r15) */ if (nr_reg_args) EMIT6_DISP_LH(0xeb000000, 0x0024, REG_2, @@ -2400,6 +2405,8 @@ static int __arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, (nr_stack_args * sizeof(u64) - 1) << 16 | tjit->stack_args_off, 0xf000 | tjit->orig_stack_args_off); + /* mvc STK_OFF_TCCNT(4,%r15),tccnt_off(%r15) */ + _EMIT6(0xd203f000 | STK_OFF_TCCNT, 0xf000 | tjit->tccnt_off); /* lgr %r1,%r8 */ EMIT4(0xb9040000, REG_1, REG_8); /* %r1() */ @@ -2456,6 +2463,9 @@ static int __arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, if (flags & (BPF_TRAMP_F_CALL_ORIG | BPF_TRAMP_F_RET_FENTRY_RET)) EMIT6_DISP_LH(0xe3000000, 0x0004, REG_2, REG_0, REG_15, tjit->retval_off); + /* mvc stack_size+STK_OFF_TCCNT(4,%r15),tccnt_off(%r15) */ + _EMIT6(0xd203f000 | (tjit->stack_size + STK_OFF_TCCNT), + 0xf000 | tjit->tccnt_off); /* aghi %r15,stack_size */ EMIT4_IMM(0xa70b0000, REG_15, tjit->stack_size); /* Emit an expoline for the following indirect jump. */ diff --git a/drivers/net/dsa/microchip/ksz_common.c b/drivers/net/dsa/microchip/ksz_common.c index 6673122266b7..42db7679c360 100644 --- a/drivers/net/dsa/microchip/ksz_common.c +++ b/drivers/net/dsa/microchip/ksz_common.c @@ -2335,13 +2335,27 @@ static u32 ksz_get_phy_flags(struct dsa_switch *ds, int port) { struct ksz_device *dev = ds->priv; - if (dev->chip_id == KSZ8830_CHIP_ID) { + switch (dev->chip_id) { + case KSZ8830_CHIP_ID: /* Silicon Errata Sheet (DS80000830A): * Port 1 does not work with LinkMD Cable-Testing. * Port 1 does not respond to received PAUSE control frames. */ if (!port) return MICREL_KSZ8_P1_ERRATA; + break; + case KSZ9477_CHIP_ID: + /* KSZ9477 Errata DS80000754C + * + * Module 4: Energy Efficient Ethernet (EEE) feature select must + * be manually disabled + * The EEE feature is enabled by default, but it is not fully + * operational. It must be manually disabled through register + * controls. If not disabled, the PHY ports can auto-negotiate + * to enable EEE, and this feature can cause link drops when + * linked to another device supporting EEE. + */ + return MICREL_NO_EEE; } return 0; diff --git a/drivers/net/ethernet/hisilicon/hns3/hnae3.h b/drivers/net/ethernet/hisilicon/hns3/hnae3.h index a4b43bcd2f0c..aaf1f42624a7 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hnae3.h +++ b/drivers/net/ethernet/hisilicon/hns3/hnae3.h @@ -814,6 +814,7 @@ struct hnae3_tc_info { u8 max_tc; /* Total number of TCs */ u8 num_tc; /* Total number of enabled TCs */ bool mqprio_active; + bool dcb_ets_active; }; #define HNAE3_MAX_DSCP 64 diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_debugfs.c b/drivers/net/ethernet/hisilicon/hns3/hns3_debugfs.c index f276b5ecb431..b8508533878b 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3_debugfs.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3_debugfs.c @@ -1045,6 +1045,7 @@ hns3_dbg_dev_specs(struct hnae3_handle *h, char *buf, int len, int *pos) struct hnae3_ae_dev *ae_dev = pci_get_drvdata(h->pdev); struct hnae3_dev_specs *dev_specs = &ae_dev->dev_specs; struct hnae3_knic_private_info *kinfo = &h->kinfo; + struct net_device *dev = kinfo->netdev; *pos += scnprintf(buf + *pos, len - *pos, "dev_spec:\n"); *pos += scnprintf(buf + *pos, len - *pos, "MAC entry num: %u\n", @@ -1087,6 +1088,9 @@ hns3_dbg_dev_specs(struct hnae3_handle *h, char *buf, int len, int *pos) dev_specs->mc_mac_size); *pos += scnprintf(buf + *pos, len - *pos, "MAC statistics number: %u\n", dev_specs->mac_stats_num); + *pos += scnprintf(buf + *pos, len - *pos, + "TX timeout threshold: %d seconds\n", + dev->watchdog_timeo / HZ); } static int hns3_dbg_dev_info(struct hnae3_handle *h, char *buf, int len) @@ -1411,9 +1415,9 @@ int hns3_dbg_init(struct hnae3_handle *handle) return 0; out: - mutex_destroy(&handle->dbgfs_lock); debugfs_remove_recursive(handle->hnae3_dbgfs); handle->hnae3_dbgfs = NULL; + mutex_destroy(&handle->dbgfs_lock); return ret; } @@ -1421,6 +1425,9 @@ void hns3_dbg_uninit(struct hnae3_handle *handle) { u32 i; + debugfs_remove_recursive(handle->hnae3_dbgfs); + handle->hnae3_dbgfs = NULL; + for (i = 0; i < ARRAY_SIZE(hns3_dbg_cmd); i++) if (handle->dbgfs_buf[i]) { kvfree(handle->dbgfs_buf[i]); @@ -1428,8 +1435,6 @@ void hns3_dbg_uninit(struct hnae3_handle *handle) } mutex_destroy(&handle->dbgfs_lock); - debugfs_remove_recursive(handle->hnae3_dbgfs); - handle->hnae3_dbgfs = NULL; } void hns3_dbg_register_debugfs(const char *debugfs_dir_name) diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c index eac2d0573241..b4895c7b3efd 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c @@ -2103,8 +2103,12 @@ static void hns3_tx_doorbell(struct hns3_enet_ring *ring, int num, */ if (test_bit(HNS3_NIC_STATE_TX_PUSH_ENABLE, &priv->state) && num && !ring->pending_buf && num <= HNS3_MAX_PUSH_BD_NUM && doorbell) { + /* This smp_store_release() pairs with smp_load_aquire() in + * hns3_nic_reclaim_desc(). Ensure that the BD valid bit + * is updated. + */ + smp_store_release(&ring->last_to_use, ring->next_to_use); hns3_tx_push_bd(ring, num); - WRITE_ONCE(ring->last_to_use, ring->next_to_use); return; } @@ -2115,6 +2119,11 @@ static void hns3_tx_doorbell(struct hns3_enet_ring *ring, int num, return; } + /* This smp_store_release() pairs with smp_load_aquire() in + * hns3_nic_reclaim_desc(). Ensure that the BD valid bit is updated. + */ + smp_store_release(&ring->last_to_use, ring->next_to_use); + if (ring->tqp->mem_base) hns3_tx_mem_doorbell(ring); else @@ -2122,7 +2131,6 @@ static void hns3_tx_doorbell(struct hns3_enet_ring *ring, int num, ring->tqp->io_base + HNS3_RING_TX_RING_TAIL_REG); ring->pending_buf = 0; - WRITE_ONCE(ring->last_to_use, ring->next_to_use); } static void hns3_tsyn(struct net_device *netdev, struct sk_buff *skb, @@ -3308,8 +3316,6 @@ static void hns3_set_default_feature(struct net_device *netdev) netdev->priv_flags |= IFF_UNICAST_FLT; - netdev->gso_partial_features |= NETIF_F_GSO_GRE_CSUM; - netdev->features |= NETIF_F_HW_VLAN_CTAG_FILTER | NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_RX | NETIF_F_RXCSUM | NETIF_F_SG | NETIF_F_GSO | @@ -3563,9 +3569,8 @@ static void hns3_reuse_buffer(struct hns3_enet_ring *ring, int i) static bool hns3_nic_reclaim_desc(struct hns3_enet_ring *ring, int *bytes, int *pkts, int budget) { - /* pair with ring->last_to_use update in hns3_tx_doorbell(), - * smp_store_release() is not used in hns3_tx_doorbell() because - * the doorbell operation already have the needed barrier operation. + /* This smp_load_acquire() pairs with smp_store_release() in + * hns3_tx_doorbell(). */ int ltu = smp_load_acquire(&ring->last_to_use); int ntc = ring->next_to_clean; diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c b/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c index 36858a72d771..682239f33082 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c @@ -773,7 +773,9 @@ static int hns3_get_link_ksettings(struct net_device *netdev, hns3_get_ksettings(h, cmd); break; case HNAE3_MEDIA_TYPE_FIBER: - if (module_type == HNAE3_MODULE_TYPE_CR) + if (module_type == HNAE3_MODULE_TYPE_UNKNOWN) + cmd->base.port = PORT_OTHER; + else if (module_type == HNAE3_MODULE_TYPE_CR) cmd->base.port = PORT_DA; else cmd->base.port = PORT_FIBRE; diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_dcb.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_dcb.c index fad5a5ff3cda..b98301e205f7 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_dcb.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_dcb.c @@ -259,7 +259,7 @@ static int hclge_ieee_setets(struct hnae3_handle *h, struct ieee_ets *ets) int ret; if (!(hdev->dcbx_cap & DCB_CAP_DCBX_VER_IEEE) || - hdev->flag & HCLGE_FLAG_MQPRIO_ENABLE) + h->kinfo.tc_info.mqprio_active) return -EINVAL; ret = hclge_ets_validate(hdev, ets, &num_tc, &map_changed); @@ -275,10 +275,7 @@ static int hclge_ieee_setets(struct hnae3_handle *h, struct ieee_ets *ets) } hclge_tm_schd_info_update(hdev, num_tc); - if (num_tc > 1) - hdev->flag |= HCLGE_FLAG_DCB_ENABLE; - else - hdev->flag &= ~HCLGE_FLAG_DCB_ENABLE; + h->kinfo.tc_info.dcb_ets_active = num_tc > 1; ret = hclge_ieee_ets_to_tm_info(hdev, ets); if (ret) @@ -487,7 +484,7 @@ static u8 hclge_getdcbx(struct hnae3_handle *h) struct hclge_vport *vport = hclge_get_vport(h); struct hclge_dev *hdev = vport->back; - if (hdev->flag & HCLGE_FLAG_MQPRIO_ENABLE) + if (h->kinfo.tc_info.mqprio_active) return 0; return hdev->dcbx_cap; @@ -611,7 +608,8 @@ static int hclge_setup_tc(struct hnae3_handle *h, if (!test_bit(HCLGE_STATE_NIC_REGISTERED, &hdev->state)) return -EBUSY; - if (hdev->flag & HCLGE_FLAG_DCB_ENABLE) + kinfo = &vport->nic.kinfo; + if (kinfo->tc_info.dcb_ets_active) return -EINVAL; ret = hclge_mqprio_qopt_check(hdev, mqprio_qopt); @@ -625,7 +623,6 @@ static int hclge_setup_tc(struct hnae3_handle *h, if (ret) return ret; - kinfo = &vport->nic.kinfo; memcpy(&old_tc_info, &kinfo->tc_info, sizeof(old_tc_info)); hclge_sync_mqprio_qopt(&kinfo->tc_info, mqprio_qopt); kinfo->tc_info.mqprio_active = tc > 0; @@ -634,13 +631,6 @@ static int hclge_setup_tc(struct hnae3_handle *h, if (ret) goto err_out; - hdev->flag &= ~HCLGE_FLAG_DCB_ENABLE; - - if (tc > 1) - hdev->flag |= HCLGE_FLAG_MQPRIO_ENABLE; - else - hdev->flag &= ~HCLGE_FLAG_MQPRIO_ENABLE; - return hclge_notify_init_up(hdev); err_out: diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c index f01a7a9ee02c..ff3f8f424ad9 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c @@ -1519,7 +1519,7 @@ static int hclge_dbg_fd_tcam_read(struct hclge_dev *hdev, bool sel_x, struct hclge_desc desc[3]; int pos = 0; int ret, i; - u32 *req; + __le32 *req; hclge_cmd_setup_basic_desc(&desc[0], HCLGE_OPC_FD_TCAM_OP, true); desc[0].flag |= cpu_to_le16(HCLGE_COMM_CMD_FLAG_NEXT); @@ -1544,22 +1544,22 @@ static int hclge_dbg_fd_tcam_read(struct hclge_dev *hdev, bool sel_x, tcam_msg.loc); /* tcam_data0 ~ tcam_data1 */ - req = (u32 *)req1->tcam_data; + req = (__le32 *)req1->tcam_data; for (i = 0; i < 2; i++) pos += scnprintf(tcam_buf + pos, HCLGE_DBG_TCAM_BUF_SIZE - pos, - "%08x\n", *req++); + "%08x\n", le32_to_cpu(*req++)); /* tcam_data2 ~ tcam_data7 */ - req = (u32 *)req2->tcam_data; + req = (__le32 *)req2->tcam_data; for (i = 0; i < 6; i++) pos += scnprintf(tcam_buf + pos, HCLGE_DBG_TCAM_BUF_SIZE - pos, - "%08x\n", *req++); + "%08x\n", le32_to_cpu(*req++)); /* tcam_data8 ~ tcam_data12 */ - req = (u32 *)req3->tcam_data; + req = (__le32 *)req3->tcam_data; for (i = 0; i < 5; i++) pos += scnprintf(tcam_buf + pos, HCLGE_DBG_TCAM_BUF_SIZE - pos, - "%08x\n", *req++); + "%08x\n", le32_to_cpu(*req++)); return ret; } diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c index 0f50dba6cc47..8ca368424436 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c @@ -11026,6 +11026,7 @@ static void hclge_get_mdix_mode(struct hnae3_handle *handle, static void hclge_info_show(struct hclge_dev *hdev) { + struct hnae3_handle *handle = &hdev->vport->nic; struct device *dev = &hdev->pdev->dev; dev_info(dev, "PF info begin:\n"); @@ -11042,9 +11043,9 @@ static void hclge_info_show(struct hclge_dev *hdev) dev_info(dev, "This is %s PF\n", hdev->flag & HCLGE_FLAG_MAIN ? "main" : "not main"); dev_info(dev, "DCB %s\n", - hdev->flag & HCLGE_FLAG_DCB_ENABLE ? "enable" : "disable"); + handle->kinfo.tc_info.dcb_ets_active ? "enable" : "disable"); dev_info(dev, "MQPRIO %s\n", - hdev->flag & HCLGE_FLAG_MQPRIO_ENABLE ? "enable" : "disable"); + handle->kinfo.tc_info.mqprio_active ? "enable" : "disable"); dev_info(dev, "Default tx spare buffer size: %u\n", hdev->tx_spare_buf_size); diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h index ec233ec57222..7bc2049b723d 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h @@ -919,8 +919,6 @@ struct hclge_dev { #define HCLGE_FLAG_MAIN BIT(0) #define HCLGE_FLAG_DCB_CAPABLE BIT(1) -#define HCLGE_FLAG_DCB_ENABLE BIT(2) -#define HCLGE_FLAG_MQPRIO_ENABLE BIT(3) u32 flag; u32 pkt_buf_size; /* Total pf buf size for tx/rx */ diff --git a/drivers/net/phy/micrel.c b/drivers/net/phy/micrel.c index b6d7981b2d1e..927d3d54658e 100644 --- a/drivers/net/phy/micrel.c +++ b/drivers/net/phy/micrel.c @@ -1800,9 +1800,6 @@ static const struct ksz9477_errata_write ksz9477_errata_writes[] = { /* Transmit waveform amplitude can be improved (1000BASE-T, 100BASE-TX, 10BASE-Te) */ {0x1c, 0x04, 0x00d0}, - /* Energy Efficient Ethernet (EEE) feature select must be manually disabled */ - {0x07, 0x3c, 0x0000}, - /* Register settings are required to meet data sheet supply current specifications */ {0x1c, 0x13, 0x6eff}, {0x1c, 0x14, 0xe6ff}, @@ -1847,6 +1844,12 @@ static int ksz9477_config_init(struct phy_device *phydev) return err; } + /* According to KSZ9477 Errata DS80000754C (Module 4) all EEE modes + * in this switch shall be regarded as broken. + */ + if (phydev->dev_flags & MICREL_NO_EEE) + phydev->eee_broken_modes = -1; + err = genphy_restart_aneg(phydev); if (err) return err; diff --git a/include/linux/micrel_phy.h b/include/linux/micrel_phy.h index 322d87255984..4e27ca7c49de 100644 --- a/include/linux/micrel_phy.h +++ b/include/linux/micrel_phy.h @@ -44,6 +44,7 @@ #define MICREL_PHY_50MHZ_CLK BIT(0) #define MICREL_PHY_FXEN BIT(1) #define MICREL_KSZ8_P1_ERRATA BIT(2) +#define MICREL_NO_EEE BIT(3) #define MICREL_KSZ9021_EXTREG_CTRL 0xB #define MICREL_KSZ9021_EXTREG_DATA_WRITE 0xC diff --git a/kernel/bpf/bpf_local_storage.c b/kernel/bpf/bpf_local_storage.c index b5149cfce7d4..146824cc9689 100644 --- a/kernel/bpf/bpf_local_storage.c +++ b/kernel/bpf/bpf_local_storage.c @@ -553,7 +553,7 @@ bpf_local_storage_update(void *owner, struct bpf_local_storage_map *smap, void *value, u64 map_flags, gfp_t gfp_flags) { struct bpf_local_storage_data *old_sdata = NULL; - struct bpf_local_storage_elem *selem = NULL; + struct bpf_local_storage_elem *alloc_selem, *selem = NULL; struct bpf_local_storage *local_storage; unsigned long flags; int err; @@ -607,11 +607,12 @@ bpf_local_storage_update(void *owner, struct bpf_local_storage_map *smap, } } - if (gfp_flags == GFP_KERNEL) { - selem = bpf_selem_alloc(smap, owner, value, true, gfp_flags); - if (!selem) - return ERR_PTR(-ENOMEM); - } + /* A lookup has just been done before and concluded a new selem is + * needed. The chance of an unnecessary alloc is unlikely. + */ + alloc_selem = selem = bpf_selem_alloc(smap, owner, value, true, gfp_flags); + if (!alloc_selem) + return ERR_PTR(-ENOMEM); raw_spin_lock_irqsave(&local_storage->lock, flags); @@ -623,13 +624,13 @@ bpf_local_storage_update(void *owner, struct bpf_local_storage_map *smap, * simple. */ err = -EAGAIN; - goto unlock_err; + goto unlock; } old_sdata = bpf_local_storage_lookup(local_storage, smap, false); err = check_flags(old_sdata, map_flags); if (err) - goto unlock_err; + goto unlock; if (old_sdata && (map_flags & BPF_F_LOCK)) { copy_map_value_locked(&smap->map, old_sdata->data, value, @@ -638,23 +639,7 @@ bpf_local_storage_update(void *owner, struct bpf_local_storage_map *smap, goto unlock; } - if (gfp_flags != GFP_KERNEL) { - /* local_storage->lock is held. Hence, we are sure - * we can unlink and uncharge the old_sdata successfully - * later. Hence, instead of charging the new selem now - * and then uncharge the old selem later (which may cause - * a potential but unnecessary charge failure), avoid taking - * a charge at all here (the "!old_sdata" check) and the - * old_sdata will not be uncharged later during - * bpf_selem_unlink_storage_nolock(). - */ - selem = bpf_selem_alloc(smap, owner, value, !old_sdata, gfp_flags); - if (!selem) { - err = -ENOMEM; - goto unlock_err; - } - } - + alloc_selem = NULL; /* First, link the new selem to the map */ bpf_selem_link_map(smap, selem); @@ -665,20 +650,16 @@ bpf_local_storage_update(void *owner, struct bpf_local_storage_map *smap, if (old_sdata) { bpf_selem_unlink_map(SELEM(old_sdata)); bpf_selem_unlink_storage_nolock(local_storage, SELEM(old_sdata), - false, false); + true, false); } unlock: raw_spin_unlock_irqrestore(&local_storage->lock, flags); - return SDATA(selem); - -unlock_err: - raw_spin_unlock_irqrestore(&local_storage->lock, flags); - if (selem) { + if (alloc_selem) { mem_uncharge(smap, owner, smap->elem_size); - bpf_selem_free(selem, smap, true); + bpf_selem_free(alloc_selem, smap, true); } - return ERR_PTR(err); + return err ? ERR_PTR(err) : SDATA(selem); } static u16 bpf_local_storage_cache_idx_get(struct bpf_local_storage_cache *cache) @@ -779,7 +760,7 @@ void bpf_local_storage_destroy(struct bpf_local_storage *local_storage) * of the loop will set the free_cgroup_storage to true. */ free_storage = bpf_selem_unlink_storage_nolock( - local_storage, selem, false, true); + local_storage, selem, true, true); } raw_spin_unlock_irqrestore(&local_storage->lock, flags); diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c index ebeb0695305a..eb01c31ed591 100644 --- a/kernel/bpf/syscall.c +++ b/kernel/bpf/syscall.c @@ -5502,9 +5502,9 @@ int kern_sys_bpf(int cmd, union bpf_attr *attr, unsigned int size) } run_ctx.bpf_cookie = 0; - run_ctx.saved_run_ctx = NULL; if (!__bpf_prog_enter_sleepable_recur(prog, &run_ctx)) { /* recursion detected */ + __bpf_prog_exit_sleepable_recur(prog, 0, &run_ctx); bpf_prog_put(prog); return -EBUSY; } diff --git a/kernel/bpf/trampoline.c b/kernel/bpf/trampoline.c index 78acf28d4873..53ff50cac61e 100644 --- a/kernel/bpf/trampoline.c +++ b/kernel/bpf/trampoline.c @@ -926,13 +926,12 @@ u64 notrace __bpf_prog_enter_sleepable_recur(struct bpf_prog *prog, migrate_disable(); might_fault(); + run_ctx->saved_run_ctx = bpf_set_run_ctx(&run_ctx->run_ctx); + if (unlikely(this_cpu_inc_return(*(prog->active)) != 1)) { bpf_prog_inc_misses_counter(prog); return 0; } - - run_ctx->saved_run_ctx = bpf_set_run_ctx(&run_ctx->run_ctx); - return bpf_prog_start_time(); } diff --git a/net/core/skmsg.c b/net/core/skmsg.c index a0659fc29bcc..6c31eefbd777 100644 --- a/net/core/skmsg.c +++ b/net/core/skmsg.c @@ -612,12 +612,18 @@ static int sk_psock_skb_ingress_self(struct sk_psock *psock, struct sk_buff *skb static int sk_psock_handle_skb(struct sk_psock *psock, struct sk_buff *skb, u32 off, u32 len, bool ingress) { + int err = 0; + if (!ingress) { if (!sock_writeable(psock->sk)) return -EAGAIN; return skb_send_sock(psock->sk, skb, off, len); } - return sk_psock_skb_ingress(psock, skb, off, len); + skb_get(skb); + err = sk_psock_skb_ingress(psock, skb, off, len); + if (err < 0) + kfree_skb(skb); + return err; } static void sk_psock_skb_state(struct sk_psock *psock, @@ -685,9 +691,7 @@ static void sk_psock_backlog(struct work_struct *work) } while (len); skb = skb_dequeue(&psock->ingress_skb); - if (!ingress) { - kfree_skb(skb); - } + kfree_skb(skb); } end: mutex_unlock(&psock->work_mutex); diff --git a/tools/testing/selftests/bpf/prog_tests/sk_storage_omem_uncharge.c b/tools/testing/selftests/bpf/prog_tests/sk_storage_omem_uncharge.c new file mode 100644 index 000000000000..f35852d245e3 --- /dev/null +++ b/tools/testing/selftests/bpf/prog_tests/sk_storage_omem_uncharge.c @@ -0,0 +1,56 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) 2023 Facebook */ +#include <test_progs.h> +#include <bpf/libbpf.h> +#include <sys/types.h> +#include <sys/socket.h> +#include "sk_storage_omem_uncharge.skel.h" + +void test_sk_storage_omem_uncharge(void) +{ + struct sk_storage_omem_uncharge *skel; + int sk_fd = -1, map_fd, err, value; + socklen_t optlen; + + skel = sk_storage_omem_uncharge__open_and_load(); + if (!ASSERT_OK_PTR(skel, "skel open_and_load")) + return; + map_fd = bpf_map__fd(skel->maps.sk_storage); + + /* A standalone socket not binding to addr:port, + * so nentns is not needed. + */ + sk_fd = socket(AF_INET6, SOCK_STREAM, 0); + if (!ASSERT_GE(sk_fd, 0, "socket")) + goto done; + + optlen = sizeof(skel->bss->cookie); + err = getsockopt(sk_fd, SOL_SOCKET, SO_COOKIE, &skel->bss->cookie, &optlen); + if (!ASSERT_OK(err, "getsockopt(SO_COOKIE)")) + goto done; + + value = 0; + err = bpf_map_update_elem(map_fd, &sk_fd, &value, 0); + if (!ASSERT_OK(err, "bpf_map_update_elem(value=0)")) + goto done; + + value = 0xdeadbeef; + err = bpf_map_update_elem(map_fd, &sk_fd, &value, 0); + if (!ASSERT_OK(err, "bpf_map_update_elem(value=0xdeadbeef)")) + goto done; + + err = sk_storage_omem_uncharge__attach(skel); + if (!ASSERT_OK(err, "attach")) + goto done; + + close(sk_fd); + sk_fd = -1; + + ASSERT_EQ(skel->bss->cookie_found, 2, "cookie_found"); + ASSERT_EQ(skel->bss->omem, 0, "omem"); + +done: + sk_storage_omem_uncharge__destroy(skel); + if (sk_fd != -1) + close(sk_fd); +} diff --git a/tools/testing/selftests/bpf/prog_tests/sockmap_helpers.h b/tools/testing/selftests/bpf/prog_tests/sockmap_helpers.h index d12665490a90..36d829a65aa4 100644 --- a/tools/testing/selftests/bpf/prog_tests/sockmap_helpers.h +++ b/tools/testing/selftests/bpf/prog_tests/sockmap_helpers.h @@ -179,6 +179,32 @@ __ret; \ }) +static inline int poll_connect(int fd, unsigned int timeout_sec) +{ + struct timeval timeout = { .tv_sec = timeout_sec }; + fd_set wfds; + int r, eval; + socklen_t esize = sizeof(eval); + + FD_ZERO(&wfds); + FD_SET(fd, &wfds); + + r = select(fd + 1, NULL, &wfds, NULL, &timeout); + if (r == 0) + errno = ETIME; + if (r != 1) + return -1; + + if (getsockopt(fd, SOL_SOCKET, SO_ERROR, &eval, &esize) < 0) + return -1; + if (eval != 0) { + errno = eval; + return -1; + } + + return 0; +} + static inline int poll_read(int fd, unsigned int timeout_sec) { struct timeval timeout = { .tv_sec = timeout_sec }; diff --git a/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c b/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c index 5674a9d0cacf..8df8cbb447f1 100644 --- a/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c +++ b/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c @@ -1452,11 +1452,18 @@ static int vsock_socketpair_connectible(int sotype, int *v0, int *v1) if (p < 0) goto close_cli; + if (poll_connect(c, IO_TIMEOUT_SEC) < 0) { + FAIL_ERRNO("poll_connect"); + goto close_acc; + } + *v0 = p; *v1 = c; return 0; +close_acc: + close(p); close_cli: close(c); close_srv: diff --git a/tools/testing/selftests/bpf/progs/bpf_tracing_net.h b/tools/testing/selftests/bpf/progs/bpf_tracing_net.h index cfed4df490f3..0b793a102791 100644 --- a/tools/testing/selftests/bpf/progs/bpf_tracing_net.h +++ b/tools/testing/selftests/bpf/progs/bpf_tracing_net.h @@ -88,6 +88,7 @@ #define sk_v6_rcv_saddr __sk_common.skc_v6_rcv_saddr #define sk_flags __sk_common.skc_flags #define sk_reuse __sk_common.skc_reuse +#define sk_cookie __sk_common.skc_cookie #define s6_addr32 in6_u.u6_addr32 diff --git a/tools/testing/selftests/bpf/progs/sk_storage_omem_uncharge.c b/tools/testing/selftests/bpf/progs/sk_storage_omem_uncharge.c new file mode 100644 index 000000000000..3e745793b27a --- /dev/null +++ b/tools/testing/selftests/bpf/progs/sk_storage_omem_uncharge.c @@ -0,0 +1,61 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) 2023 Facebook */ +#include "vmlinux.h" +#include "bpf_tracing_net.h" +#include <bpf/bpf_helpers.h> +#include <bpf/bpf_tracing.h> +#include <bpf/bpf_core_read.h> + +void *local_storage_ptr = NULL; +void *sk_ptr = NULL; +int cookie_found = 0; +__u64 cookie = 0; +__u32 omem = 0; + +void *bpf_rdonly_cast(void *, __u32) __ksym; + +struct { + __uint(type, BPF_MAP_TYPE_SK_STORAGE); + __uint(map_flags, BPF_F_NO_PREALLOC); + __type(key, int); + __type(value, int); +} sk_storage SEC(".maps"); + +SEC("fexit/bpf_local_storage_destroy") +int BPF_PROG(bpf_local_storage_destroy, struct bpf_local_storage *local_storage) +{ + struct sock *sk; + + if (local_storage_ptr != local_storage) + return 0; + + sk = bpf_rdonly_cast(sk_ptr, bpf_core_type_id_kernel(struct sock)); + if (sk->sk_cookie.counter != cookie) + return 0; + + cookie_found++; + omem = sk->sk_omem_alloc.counter; + local_storage_ptr = NULL; + + return 0; +} + +SEC("fentry/inet6_sock_destruct") +int BPF_PROG(inet6_sock_destruct, struct sock *sk) +{ + int *value; + + if (!cookie || sk->sk_cookie.counter != cookie) + return 0; + + value = bpf_sk_storage_get(&sk_storage, sk, 0, 0); + if (value && *value == 0xdeadbeef) { + cookie_found++; + sk_ptr = sk; + local_storage_ptr = sk->sk_bpf_storage; + } + + return 0; +} + +char _license[] SEC("license") = "GPL"; |