December 2023 - Kernel - mailweb.openeuler.org

[PATCH OLK-6.6 0/6] Add support for hisi HBM devices
by Zhang Zekun 12 Dec '23

12 Dec '23

Add support for HISI HBM device, and HBM ACLS repair. The patch set includes three functionalities. 1. Add support for hbm device hotplug. 2. provide extra information with hbm locality information 3. Add support for HBM acls repair. Zhang Zekun (6): ACPI: OSL: Export the symbol of acpi_hotplug_schedule soc: hisilicon: hisi_hbmdev: Add power domain control methods ACPI: memhotplug: export the state of each hotplug device soc: hisilicon: hisi_hbmdev: Provide extra memory topology information soc: hbmcache: Add support for online and offline the hbm cache soc: hisilicon: hisi_hbmdev: Add hbm acls repair and query methods drivers/acpi/acpi_memhotplug.c | 22 ++ drivers/acpi/internal.h | 1 - drivers/acpi/osl.c | 1 + drivers/base/container.c | 1 + drivers/soc/hisilicon/Kconfig | 36 +++ drivers/soc/hisilicon/Makefile | 3 + drivers/soc/hisilicon/hisi_hbmcache.c | 147 +++++++++ drivers/soc/hisilicon/hisi_hbmdev.c | 435 ++++++++++++++++++++++++++ drivers/soc/hisilicon/hisi_internal.h | 31 ++ include/linux/acpi.h | 2 + include/linux/memory_hotplug.h | 1 + 11 files changed, 679 insertions(+), 1 deletion(-) create mode 100644 drivers/soc/hisilicon/hisi_hbmcache.c create mode 100644 drivers/soc/hisilicon/hisi_hbmdev.c create mode 100644 drivers/soc/hisilicon/hisi_internal.h -- 2.17.1

2 7

[PATCH OLK-5.10] KVM: arm64: Translate logic cluster id to physical cluster id when updating lsudvmbm
by chenxiang 12 Dec '23

12 Dec '23

From: Xiang Chen <chenxiang66(a)hisilicon.com> driver inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8HY1B ---------------------------------------------------------------- For dvmbm feature, MN requires physical cluster id while it is filled with logic cluster id right now. In some situations which physical cluster id is not equal to logic cluster id such as in PG boards, it will cause issues when enabling dvmbm. To avoid the issue, translate logic cluster id to physical cluster id when updating lsudvmbm. Signed-off-by: Xiang Chen <chenxiang66(a)hisilicon.com> Signed-off-by: Cai Jian <caijian11(a)h-partners.com> --- arch/arm64/kvm/hisilicon/hisi_virt.c | 166 +++++++++++++++++++++++++-- 1 file changed, 157 insertions(+), 9 deletions(-) diff --git a/arch/arm64/kvm/hisilicon/hisi_virt.c b/arch/arm64/kvm/hisilicon/hisi_virt.c index 10233b801896..1f4a8e44228e 100644 --- a/arch/arm64/kvm/hisilicon/hisi_virt.c +++ b/arch/arm64/kvm/hisilicon/hisi_virt.c @@ -236,12 +236,113 @@ static int kvm_dvmbm_get_dies_info(struct kvm *kvm, u64 *vm_aff3s, int size) return num; } +#define MAX_CLUSTERS_PER_DIE 8 +#define TB_MN_BASE 0x00C6067f0000 +#define TA_MN_BASE 0x0046067F0000 +#define CHIP_ADDR_OFFSET(_chip) (((((_chip) >> 3) & 0x1) * 0x80000000000) + \ + ((((_chip) >> 2) & 0x1) * (0x100000000000)) + \ + (((_chip) & 0x3) * 0x200000000000)) +#define MAX_PG_CFG_SKITS 4 +#define MAX_PG_CFG_DIES 2 +#define MN_ECO0_OFFSET 0xc00 +#define SOCKET_ID_MASK 0xf8 +#define SOCKET_ID_SHIFT 3 +#define DIE_ID_MASK 0x7 + +static u32 socket_num, die_num; + +static void kvm_get_socket_num(u32 *socket_num) +{ + int socket_id[MAX_PG_CFG_SKITS], cpu; + u32 num = 0; + + for_each_cpu(cpu, cpu_possible_mask) { + bool found = false; + u64 aff3, socket; + int i; + + aff3 = MPIDR_AFFINITY_LEVEL(cpu_logical_map(cpu), 3); + /* aff3[7:3]: socket ID */ + socket = (aff3 & SOCKET_ID_MASK) >> SOCKET_ID_SHIFT; + for (i = 0; i < num; i++) { + if (socket_id[i] == socket) { + found = true; + break; + } + } + if (!found) + socket_id[num++] = socket; + } + *socket_num = num; +} + +static void kvm_get_die_num(u32 *die_num) +{ + int die_id[MAX_PG_CFG_DIES], cpu; + u32 num = 0; + + for_each_cpu(cpu, cpu_possible_mask) { + bool found = false; + u64 aff3, die; + int i; + + aff3 = MPIDR_AFFINITY_LEVEL(cpu_logical_map(cpu), 3); + /* aff3[2:0]: die ID */ + die = aff3 & DIE_ID_MASK; + for (i = 0; i < num; i++) { + if (die_id[i] == die) { + found = true; + break; + } + } + if (!found) + die_id[num++] = die; + } + *die_num = num; +} + +static u32 g_pg_cfg[MAX_PG_CFG_SKITS * MAX_PG_CFG_DIES]; + +static u64 kvm_switch_aff2_logic_to_phy(unsigned long pg_cfg, u64 logic_aff2) +{ + u32 pg_num = 0, i, j; + u32 cluster_pg[MAX_CLUSTERS_PER_DIE]; + u32 pg_flag[MAX_CLUSTERS_PER_DIE]; + + for (i = 0; i < MAX_CLUSTERS_PER_DIE; i++) { + if (test_bit(i, &pg_cfg)) + pg_num++; + cluster_pg[i] = i; + pg_flag[i] = 0; + } + + for (i = 0; i < MAX_CLUSTERS_PER_DIE - pg_num; i++) { + if (test_bit(i, &pg_cfg)) { + for (j = 0; j < pg_num; j++) { + u32 cluster_bak = MAX_CLUSTERS_PER_DIE - pg_num + j; + + if (!test_bit(cluster_bak, &pg_cfg) && + !pg_flag[cluster_bak]) { + pg_flag[cluster_bak] = 1; + cluster_pg[i] = cluster_bak; + cluster_pg[cluster_bak] = i; + break; + } + } + } + } + + return cluster_pg[logic_aff2]; +} + static void kvm_update_vm_lsudvmbm(struct kvm *kvm) { - u64 mpidr, aff3, aff2, aff1; + u64 mpidr, aff3, aff2, aff1, phy_aff2; u64 vm_aff3s[DVMBM_MAX_DIES]; u64 val; int cpu, nr_dies; + unsigned long pg_cfg; + u32 socket_id; nr_dies = kvm_dvmbm_get_dies_info(kvm, vm_aff3s, DVMBM_MAX_DIES); if (nr_dies > 2) { @@ -258,8 +359,14 @@ static void kvm_update_vm_lsudvmbm(struct kvm *kvm) mpidr = cpu_logical_map(cpu); aff2 = MPIDR_AFFINITY_LEVEL(mpidr, 2); aff1 = MPIDR_AFFINITY_LEVEL(mpidr, 1); - - val |= 1ULL << (aff2 * 4 + aff1); + socket_id = (aff3 & SOCKET_ID_MASK) >> SOCKET_ID_SHIFT; + + if ((vm_aff3s[0] & DIE_ID_MASK) == 0x3) /* TB totem */ + pg_cfg = g_pg_cfg[0 + socket_id * die_num] & 0xff; + else /* TA totem */ + pg_cfg = g_pg_cfg[1 + socket_id * die_num] & 0xff; + phy_aff2 = kvm_switch_aff2_logic_to_phy(pg_cfg, aff2); + val |= 1ULL << (phy_aff2 * 4 + aff1); } goto out_update; @@ -268,7 +375,7 @@ static void kvm_update_vm_lsudvmbm(struct kvm *kvm) /* nr_dies == 2 */ val = DVMBM_RANGE_TWO_DIES << DVMBM_RANGE_SHIFT | DVMBM_GRAN_CLUSTER << DVMBM_GRAN_SHIFT | - vm_aff3s[0] << DVMBM_DIE1_SHIFT | + vm_aff3s[0] << DVMBM_DIE1_SHIFT | vm_aff3s[1] << DVMBM_DIE2_SHIFT; /* and fulfill bits [43:0] */ @@ -276,11 +383,23 @@ static void kvm_update_vm_lsudvmbm(struct kvm *kvm) mpidr = cpu_logical_map(cpu); aff3 = MPIDR_AFFINITY_LEVEL(mpidr, 3); aff2 = MPIDR_AFFINITY_LEVEL(mpidr, 2); - - if (aff3 == vm_aff3s[0]) - val |= 1ULL << (aff2 + DVMBM_DIE1_CLUSTER_SHIFT); - else - val |= 1ULL << (aff2 + DVMBM_DIE2_CLUSTER_SHIFT); + socket_id = (aff3 & SOCKET_ID_MASK) >> SOCKET_ID_SHIFT; + + if (aff3 == vm_aff3s[0]) { + if ((vm_aff3s[0] & DIE_ID_MASK) == 0x3) /* TB totem */ + pg_cfg = g_pg_cfg[0 + socket_id * die_num] & 0xff; + else /* TA totem */ + pg_cfg = g_pg_cfg[1 + socket_id * die_num] & 0xff; + phy_aff2 = kvm_switch_aff2_logic_to_phy(pg_cfg, aff2); + val |= 1ULL << (phy_aff2 + DVMBM_DIE1_CLUSTER_SHIFT); + } else { + if ((vm_aff3s[1] & DIE_ID_MASK) == 0x3) /* TB totem */ + pg_cfg = g_pg_cfg[0 + socket_id * die_num] & 0xff; + else /* TA totem */ + pg_cfg = g_pg_cfg[1 + socket_id * die_num] & 0xff; + phy_aff2 = kvm_switch_aff2_logic_to_phy(pg_cfg, aff2); + val |= 1ULL << (phy_aff2 + DVMBM_DIE2_CLUSTER_SHIFT); + } } out_update: @@ -357,6 +476,33 @@ void kvm_hisi_dvmbm_put(struct kvm_vcpu *vcpu) */ } +static void kvm_get_pg_cfg(u32 *pg_cfgs) +{ + void __iomem *mn_base; + u32 i, j; + u64 mn_phy_base; + u32 val; + + kvm_get_socket_num(&socket_num); + kvm_get_die_num(&die_num); + for (i = 0; i < socket_num; i++) { + for (j = 0; j < die_num; j++) { + mn_phy_base = (j == 0) ? TB_MN_BASE : TA_MN_BASE; + mn_phy_base += CHIP_ADDR_OFFSET(i); + mn_phy_base += MN_ECO0_OFFSET; + + mn_base = ioremap(mn_phy_base, 4); + if (!mn_base) { + kvm_info("MN base addr ioremap fail\n"); + return; + } + val = readl_relaxed(mn_base); + pg_cfgs[j + i * die_num] = val & 0xff; + iounmap(mn_base); + } + } +} + int kvm_hisi_init_dvmbm(struct kvm *kvm) { if (!kvm_dvmbm_support) @@ -367,6 +513,8 @@ int kvm_hisi_init_dvmbm(struct kvm *kvm) if (!kvm->arch.dvm_cpumask) return -ENOMEM; + kvm_get_pg_cfg(g_pg_cfg); + return 0; } -- 2.30.0

3 3

[PATCH OLK-6.6 v4 0/7] support userswap feature
by Peng Zhang 12 Dec '23

12 Dec '23

From: ZhangPeng <zhangpeng362(a)huawei.com> This patch series support userswap feature including registration, unregistration, swap-out and swap-in. Userswap feature depends on CONFIG_USERSWAP and can be enabled by cmdline enable_userswap. We tested the concurrent scenario of multi-threaded page fault and multi-threaded swap-in in the uswap demo;and the remapping in the swap-out phase and the copy-free function in the swap-in phase were ok. During the test, related debugging functions including CONFIG_DEBUG_VM, lockdep, slub debug, kasan and kmemleak are enabled. ChangeLog: v3->v4: - define UFFDIO_REGISTER_MODE_USWAP ((__u64)1<<10) rather than ((__u64)1<<3) v2->v3: - update patch 7 commit v1->v2: - enable CONFIG_USERSWAP for openeuler_defconfig ZhangPeng (7): mm/userswap: add VM_USWAP and SWP_USERSWAP_ENTRY mm/userswap: add enable_userswap boot option mm/userswap: introduce MREMAP_USWAP_SET_PTE mm/userswap: support userswap via userfaultfd mm/userswap: introduce UFFDIO_COPY_MODE_DIRECT_MAP mm/userswap: provide cpu info in userfault msg mm/userswap: openeuler_defconfig: enable userswap arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/proc/task_mmu.c | 3 + fs/userfaultfd.c | 43 +- include/linux/mm.h | 8 + include/linux/swap.h | 14 +- include/linux/swapops.h | 16 + include/linux/userfaultfd_k.h | 5 + include/linux/userswap.h | 60 +++ include/trace/events/mmflags.h | 7 + include/uapi/linux/mman.h | 1 + include/uapi/linux/userfaultfd.h | 3 + mm/Kconfig | 10 + mm/Makefile | 1 + mm/memory.c | 8 + mm/mremap.c | 8 +- mm/userfaultfd.c | 12 + mm/userswap.c | 552 +++++++++++++++++++++++++ 18 files changed, 747 insertions(+), 6 deletions(-) create mode 100644 include/linux/userswap.h create mode 100644 mm/userswap.c -- 2.25.1

2 8

[PATCH OLK-6.6] amba-pl011: Fix no irq issue due to no IRQ domain found
by Lin Yujun 12 Dec '23

12 Dec '23

From: Lin Ruizhe <linruizhe(a)huawei.com> hulk inclusion category: bugfix bugzilla: 176552 https://gitee.com/openeuler/kernel/issues/I4DDEL ------------------------------------------------- If pl011 interrupt is connected to MBIGEN interrupt controller, because the mbigen initialization is too late, which will lead to no IRQ due to no IRQ domain found, logs is shown below, "irq: no irq domain found for uart0 !" When dev->irq[0] is zero, try to get IRQ by of_irq_get() again, and return -EPROBE_DEFER if the IRQ domain is not yet created. Using deferred probing mechanism to fix the issue. Signed-off-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Lin ruizhe <linruizhe(a)huawei.com> Signed-off-by: He Ying <heying24(a)huawei.com> Signed-off-by: Chen Jun <chenjun102(a)huawei.com> Signed-off-by: Zheng Zengkai <zhengzengkai(a)huawei.com> conflicts: drivers/tty/serial/amba-pl011.c Signed-off-by: Lin Yujun <linyujun809(a)huawei.com> --- drivers/tty/serial/amba-pl011.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) diff --git a/drivers/tty/serial/amba-pl011.c b/drivers/tty/serial/amba-pl011.c index 3dc9b0fcab1c..648f99d4774c 100644 --- a/drivers/tty/serial/amba-pl011.c +++ b/drivers/tty/serial/amba-pl011.c @@ -2815,10 +2815,20 @@ static int pl011_probe(struct amba_device *dev, const struct amba_id *id) uap->vendor = vendor; uap->fifosize = vendor->get_fifosize(dev); uap->port.iotype = vendor->access_32b ? UPIO_MEM32 : UPIO_MEM; - uap->port.irq = dev->irq[0]; uap->port.ops = &amba_pl011_pops; uap->port.rs485_config = pl011_rs485_config; uap->port.rs485_supported = pl011_rs485_supported; + + /* if no irq domain found, irq number is 0, try again */ + if (!dev->irq[0] && dev->dev.of_node) { + ret = of_irq_get(dev->dev.of_node, 0); + if (ret < 0) + return ret; + dev->irq[0] = ret; + } + + uap->port.irq = dev->irq[0]; + snprintf(uap->type, sizeof(uap->type), "PL011 rev%u", amba_rev(dev)); if (device_property_read_u32(&dev->dev, "reg-io-width", &val) == 0) { -- 2.34.1

3 2

[PATCH OLK-6.6] x86/ioapic: add an interface for driver to sync hardware data into memory
by Lin Yujun 12 Dec '23

12 Dec '23

From: Hongbo Yao <yaohongbo(a)huawei.com> hulk inclusion category: bugfix bugzilla: 20890, https://gitee.com/openeuler/kernel/issues/I5ZV1W -------------------------------- The default interrupt trigger mode of Pci driver is Level, however, some pci drivers want to change irq trigger mode to edge by writing hardware register directly. In current kernel, this kind of driver needs to sync entry data which has already been reservd in memory. This patch provides an interface for the driver to sync entry data in memory. Signed-off-by: Hongbo Yao <yaohongbo(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: zhangyi (F) <yi.zhang(a)huawei.com> Signed-off-by: Xiongfeng Wang <wangxiongfeng2(a)huawei.com> Reviewed-by: Zhang Jianhua <chris.zjh(a)huawei.com> Reviewed-by: Liao Chang <liaochang1(a)huawei.com> Signed-off-by: Zheng Zengkai <zhengzengkai(a)huawei.com> Signed-off-by: Lin Yujun <linyujun809(a)huawei.com> --- arch/x86/include/asm/io_apic.h | 1 + arch/x86/kernel/apic/io_apic.c | 25 +++++++++++++++++++++++++ 2 files changed, 26 insertions(+) diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h index 51c782600e02..435ae723a238 100644 --- a/arch/x86/include/asm/io_apic.h +++ b/arch/x86/include/asm/io_apic.h @@ -174,6 +174,7 @@ extern void clear_IO_APIC(void); extern void restore_boot_irq_mode(void); extern int IO_APIC_get_PCI_irq_vector(int bus, int devfn, int pin); extern void print_IO_APICs(void); +extern void ioapic_sync_hardware_data(int irq); #else /* !CONFIG_X86_IO_APIC */ #define IO_APIC_IRQ(x) 0 diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 00da6cf6b07d..d256e0b57e94 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1202,6 +1202,31 @@ int IO_APIC_get_PCI_irq_vector(int bus, int slot, int pin) } EXPORT_SYMBOL(IO_APIC_get_PCI_irq_vector); +/* + * The function is only provided for Drivers to keep the data value in + * memory consistent with the value of the register, and these drivers + * must use it carefully. + * + * The correct step should be: + * Change register ---> ioapic_sync_hardware_data ---> request_irq + */ +void ioapic_sync_hardware_data(int irq) +{ + struct IO_APIC_route_entry entry; + struct mp_chip_data *data; + struct irq_pin_list *pin_list; + + data = irq_get_chip_data(irq); + for_each_irq_pin(pin_list, data->irq_2_pin) { + entry = ioapic_read_entry(pin_list->apic, pin_list->pin); + data->entry.trigger = entry.trigger; + data->entry.polarity = entry.polarity; + pr_debug("irq[%d] update trigger[%d] polarity[%d]\n", + irq, entry.trigger, entry.polarity); + } +} +EXPORT_SYMBOL(ioapic_sync_hardware_data); + static struct irq_chip ioapic_chip, ioapic_ir_chip; static void __init setup_IO_APIC_irqs(void) -- 2.34.1

2 1

[PATCH OLK-6.6 0/9] files cgroup featrue
by chenridong 12 Dec '23

12 Dec '23

Binder Makin (1): cgroups: Resource controller for open files Hou Tao (1): cgroup/files: use task_get_css() to get a valid css during dup_fd() Lu Jialin (3): enable CONFIG_CGROUP_FILES in openeuler_defconfig for x86 and arm64 fs: fix files.usage bug when move tasks fs/filescontrol.c: fix warning:large integer implicitly truncated to unsigned type Yang Yingliang (1): cgroup/files: support boot parameter to control if disable files cgroup Yu Kuai (1): fs/filescontrol: add a switch to enable / disable accounting of open fds Zhang Xiaoxu (2): files_cgroup: fix error pointer when kvm_vm_worker_thread files_cgroup: Fix soft lockup when refcnt overflow. .../admin-guide/kernel-parameters.txt | 7 +- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/Makefile | 1 + fs/file.c | 63 +++- fs/filescontrol.c | 312 ++++++++++++++++++ include/linux/cgroup-defs.h | 8 +- include/linux/cgroup_subsys.h | 6 + include/linux/fdtable.h | 1 + include/linux/filescontrol.h | 40 +++ init/Kconfig | 10 + 11 files changed, 442 insertions(+), 8 deletions(-) create mode 100644 fs/filescontrol.c create mode 100644 include/linux/filescontrol.h -- 2.34.1

2 10

[PATCH OLK-6.6 0/9] files cgroup enble
by chenridong 11 Dec '23

11 Dec '23

Binder Makin (1): cgroups: Resource controller for open files Hou Tao (1): cgroup/files: use task_get_css() to get a valid css during dup_fd() Lu Jialin (3): enable CONFIG_CGROUP_FILES in openeuler_defconfig for x86 and arm64 fs: fix files.usage bug when move tasks fs/filescontrol.c: fix warning:large integer implicitly truncated to unsigned type Yang Yingliang (1): cgroup/files: support boot parameter to control if disable files cgroup Yu Kuai (1): fs/filescontrol: add a switch to enable / disable accounting of open fds Zhang Xiaoxu (2): files_cgroup: fix error pointer when kvm_vm_worker_thread files_cgroup: Fix soft lockup when refcnt overflow. .../admin-guide/kernel-parameters.txt | 7 +- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/Makefile | 1 + fs/file.c | 63 +++- fs/filescontrol.c | 312 ++++++++++++++++++ include/linux/cgroup-defs.h | 8 +- include/linux/cgroup_subsys.h | 6 + include/linux/fdtable.h | 1 + include/linux/filescontrol.h | 40 +++ init/Kconfig | 10 + 11 files changed, 442 insertions(+), 8 deletions(-) create mode 100644 fs/filescontrol.c create mode 100644 include/linux/filescontrol.h -- 2.34.1

2 10

[PATCH openEuler-1.0-LTS] tun: avoid double free in tun_free_netdev
by Ziyang Xuan 11 Dec '23

11 Dec '23

From: George Kennedy <george.kennedy(a)oracle.com> mainline inclusion from mainline-v5.16-rc7 commit 158b515f703e75e7d68289bf4d98c664e1d632df category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I8NHAP CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… ------------------------------- Avoid double free in tun_free_netdev() by moving the dev->tstats and tun->security allocs to a new ndo_init routine (tun_net_init()) that will be called by register_netdevice(). ndo_init is paired with the desctructor (tun_free_netdev()), so if there's an error in register_netdevice() the destructor will handle the frees. BUG: KASAN: double-free or invalid-free in selinux_tun_dev_free_security+0x1a/0x20 security/selinux/hooks.c:5605 CPU: 0 PID: 25750 Comm: syz-executor416 Not tainted 5.16.0-rc2-syzk #1 Hardware name: Red Hat KVM, BIOS Call Trace: <TASK> __dump_stack lib/dump_stack.c:88 [inline] dump_stack_lvl+0x89/0xb5 lib/dump_stack.c:106 print_address_description.constprop.9+0x28/0x160 mm/kasan/report.c:247 kasan_report_invalid_free+0x55/0x80 mm/kasan/report.c:372 ____kasan_slab_free mm/kasan/common.c:346 [inline] __kasan_slab_free+0x107/0x120 mm/kasan/common.c:374 kasan_slab_free include/linux/kasan.h:235 [inline] slab_free_hook mm/slub.c:1723 [inline] slab_free_freelist_hook mm/slub.c:1749 [inline] slab_free mm/slub.c:3513 [inline] kfree+0xac/0x2d0 mm/slub.c:4561 selinux_tun_dev_free_security+0x1a/0x20 security/selinux/hooks.c:5605 security_tun_dev_free_security+0x4f/0x90 security/security.c:2342 tun_free_netdev+0xe6/0x150 drivers/net/tun.c:2215 netdev_run_todo+0x4df/0x840 net/core/dev.c:10627 rtnl_unlock+0x13/0x20 net/core/rtnetlink.c:112 __tun_chr_ioctl+0x80c/0x2870 drivers/net/tun.c:3302 tun_chr_ioctl+0x2f/0x40 drivers/net/tun.c:3311 vfs_ioctl fs/ioctl.c:51 [inline] __do_sys_ioctl fs/ioctl.c:874 [inline] __se_sys_ioctl fs/ioctl.c:860 [inline] __x64_sys_ioctl+0x19d/0x220 fs/ioctl.c:860 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x3a/0x80 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x44/0xae Reported-by: syzkaller <syzkaller(a)googlegroups.com> Signed-off-by: George Kennedy <george.kennedy(a)oracle.com> Suggested-by: Jakub Kicinski <kuba(a)kernel.org> Link: https://lore.kernel.org/r/1639679132-19884-1-git-send-email-george.kennedy@… Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Conflicts: drivers/net/tun.c Signed-off-by: Ziyang Xuan <william.xuanziyang(a)huawei.com> --- drivers/net/tun.c | 111 +++++++++++++++++++++++++--------------------- 1 file changed, 60 insertions(+), 51 deletions(-) diff --git a/drivers/net/tun.c b/drivers/net/tun.c index a6b233a01f35..e9b8a68b350a 100644 --- a/drivers/net/tun.c +++ b/drivers/net/tun.c @@ -248,6 +248,9 @@ struct tun_struct { struct tun_prog __rcu *steering_prog; struct tun_prog __rcu *filter_prog; struct ethtool_link_ksettings link_ksettings; + /* init args */ + struct file *file; + struct ifreq *ifr; }; struct veth { @@ -273,6 +276,9 @@ void *tun_ptr_to_xdp(void *ptr) } EXPORT_SYMBOL(tun_ptr_to_xdp); +static void tun_flow_init(struct tun_struct *tun); +static void tun_flow_uninit(struct tun_struct *tun); + static int tun_napi_receive(struct napi_struct *napi, int budget) { struct tun_file *tfile = container_of(napi, struct tun_file, napi); @@ -1030,6 +1036,49 @@ static int check_filter(struct tap_filter *filter, const struct sk_buff *skb) static const struct ethtool_ops tun_ethtool_ops; +static int tun_net_init(struct net_device *dev) +{ + struct tun_struct *tun = netdev_priv(dev); + struct ifreq *ifr = tun->ifr; + int err; + + tun->pcpu_stats = netdev_alloc_pcpu_stats(struct tun_pcpu_stats); + if (!tun->pcpu_stats) + return -ENOMEM; + + spin_lock_init(&tun->lock); + + err = security_tun_dev_alloc_security(&tun->security); + if (err < 0) { + free_percpu(tun->pcpu_stats); + return err; + } + + tun_flow_init(tun); + + dev->hw_features = NETIF_F_SG | NETIF_F_FRAGLIST | + TUN_USER_FEATURES | NETIF_F_HW_VLAN_CTAG_TX | + NETIF_F_HW_VLAN_STAG_TX; + dev->features = dev->hw_features | NETIF_F_LLTX; + dev->vlan_features = dev->features & + ~(NETIF_F_HW_VLAN_CTAG_TX | + NETIF_F_HW_VLAN_STAG_TX); + + tun->flags = (tun->flags & ~TUN_FEATURES) | + (ifr->ifr_flags & TUN_FEATURES); + + INIT_LIST_HEAD(&tun->disabled); + err = tun_attach(tun, tun->file, false, ifr->ifr_flags & IFF_NAPI, + ifr->ifr_flags & IFF_NAPI_FRAGS, false); + if (err < 0) { + tun_flow_uninit(tun); + security_tun_dev_free_security(tun->security); + free_percpu(tun->pcpu_stats); + return err; + } + return 0; +} + /* Net device detach from fd. */ static void tun_net_uninit(struct net_device *dev) { @@ -1260,6 +1309,7 @@ static int tun_xdp(struct net_device *dev, struct netdev_bpf *xdp) } static const struct net_device_ops tun_netdev_ops = { + .ndo_init = tun_net_init, .ndo_uninit = tun_net_uninit, .ndo_open = tun_net_open, .ndo_stop = tun_net_close, @@ -1339,6 +1389,7 @@ static int tun_xdp_tx(struct net_device *dev, struct xdp_buff *xdp) } static const struct net_device_ops tap_netdev_ops = { + .ndo_init = tun_net_init, .ndo_uninit = tun_net_uninit, .ndo_open = tun_net_open, .ndo_stop = tun_net_close, @@ -1378,7 +1429,7 @@ static void tun_flow_uninit(struct tun_struct *tun) #define MAX_MTU 65535 /* Initialize net device. */ -static void tun_net_init(struct net_device *dev) +static void tun_net_initialize(struct net_device *dev) { struct tun_struct *tun = netdev_priv(dev); @@ -2633,9 +2684,6 @@ static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr) if (!dev) return -ENOMEM; - err = dev_get_valid_name(net, dev, name); - if (err < 0) - goto err_free_dev; dev_net_set(dev, net); dev->rtnl_link_ops = &tun_link_ops; @@ -2654,42 +2702,17 @@ static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr) tun->rx_batched = 0; RCU_INIT_POINTER(tun->steering_prog, NULL); - tun->pcpu_stats = netdev_alloc_pcpu_stats(struct tun_pcpu_stats); - if (!tun->pcpu_stats) { - err = -ENOMEM; - goto err_free_dev; - } - - spin_lock_init(&tun->lock); - - err = security_tun_dev_alloc_security(&tun->security); - if (err < 0) - goto err_free_stat; - - tun_net_init(dev); - tun_flow_init(tun); - - dev->hw_features = NETIF_F_SG | NETIF_F_FRAGLIST | - TUN_USER_FEATURES | NETIF_F_HW_VLAN_CTAG_TX | - NETIF_F_HW_VLAN_STAG_TX; - dev->features = dev->hw_features | NETIF_F_LLTX; - dev->vlan_features = dev->features & - ~(NETIF_F_HW_VLAN_CTAG_TX | - NETIF_F_HW_VLAN_STAG_TX); + tun->ifr = ifr; + tun->file = file; - tun->flags = (tun->flags & ~TUN_FEATURES) | - (ifr->ifr_flags & TUN_FEATURES); - - INIT_LIST_HEAD(&tun->disabled); - err = tun_attach(tun, file, false, ifr->ifr_flags & IFF_NAPI, - ifr->ifr_flags & IFF_NAPI_FRAGS, false); - if (err < 0) - goto err_free_flow; + tun_net_initialize(dev); err = register_netdevice(tun->dev); - if (err < 0) - goto err_detach; - /* free_netdev() won't check refcnt, to aovid race + if (err < 0) { + free_netdev(dev); + return err; + } + /* free_netdev() won't check refcnt, to avoid race * with dev_put() we need publish tun after registration. */ rcu_assign_pointer(tfile->tun, tun); @@ -2707,20 +2730,6 @@ static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr) strcpy(ifr->ifr_name, tun->dev->name); return 0; - -err_detach: - tun_detach_all(dev); - /* register_netdevice() already called tun_free_netdev() */ - goto err_free_dev; - -err_free_flow: - tun_flow_uninit(tun); - security_tun_dev_free_security(tun->security); -err_free_stat: - free_percpu(tun->pcpu_stats); -err_free_dev: - free_netdev(dev); - return err; } static void tun_get_iff(struct net *net, struct tun_struct *tun, -- 2.25.1

2 1

[PATCH OLK-5.10 v3] smart_grid: cpufreq: clear offline and isolated CPU in warm CPUs
by Yipeng Zou 11 Dec '23

11 Dec '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8MFO6 CVE: NA ---------------------------------------- When there has some CPU has been offline or isolated, the warm zone CPUs will not clear these CPUs. Because of the warm zone were saluted by a complement of hot zone. So, we need to clear offline and isolated CPUs in every time's warm zone CPUs update. Fixes: 6a00891859d8 ("smart_grid: cpufreq: introduce smart_grid cpufreq control") Signed-off-by: Yipeng Zou <zouyipeng(a)huawei.com> Tested-by: jifa <jifa(a)huawei.com> --- kernel/sched/grid/qos.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/kernel/sched/grid/qos.c b/kernel/sched/grid/qos.c index f8c5deffd3ea..7ee3687347ce 100644 --- a/kernel/sched/grid/qos.c +++ b/kernel/sched/grid/qos.c @@ -179,6 +179,7 @@ int sched_grid_zone_update(bool is_locked) raw_spin_lock_irqsave(&sg_zone.lock, flags); cpumask_clear(&sg_zone.cpus[SMART_GRID_ZONE_HOT]); + cpumask_clear(&sg_zone.cpus[SMART_GRID_ZONE_WARM]); list_for_each(pos, &sg_zone.af_list_head) { af_pos = list_entry(pos, struct auto_affinity, af_list); @@ -190,10 +191,14 @@ int sched_grid_zone_update(bool is_locked) cpumask_or(&sg_zone.cpus[SMART_GRID_ZONE_HOT], &sg_zone.cpus[SMART_GRID_ZONE_HOT], af_pos->ad.domains[af_pos->ad.curr_level]); + /* Update warm zone CPUs to max level first */ + cpumask_or(&sg_zone.cpus[SMART_GRID_ZONE_WARM], &sg_zone.cpus[SMART_GRID_ZONE_WARM], + af_pos->ad.domains[af_pos->ad.dcount - 1]); } - cpumask_complement(&sg_zone.cpus[SMART_GRID_ZONE_WARM], - &sg_zone.cpus[SMART_GRID_ZONE_HOT]); + /* Then reset warm zone CPUs without hot zone CPUs */ + cpumask_andnot(&sg_zone.cpus[SMART_GRID_ZONE_WARM], &sg_zone.cpus[SMART_GRID_ZONE_WARM], + &sg_zone.cpus[SMART_GRID_ZONE_HOT]); if (!is_locked) raw_spin_unlock_irqrestore(&sg_zone.lock, flags); -- 2.34.1

2 1

[PATCH OLK-6.6 0/9] files cgroup enable
by chenridong 11 Dec '23

11 Dec '23

Binder Makin (1): cgroups: Resource controller for open files Hou Tao (1): cgroup/files: use task_get_css() to get a valid css during dup_fd() Lu Jialin (3): enable CONFIG_CGROUP_FILES in openeuler_defconfig for x86 and arm64 fs: fix files.usage bug when move tasks fs/filescontrol.c: fix warning:large integer implicitly truncated to unsigned type Yang Yingliang (1): cgroup/files: support boot parameter to control if disable files cgroup Yu Kuai (1): fs/filescontrol: add a switch to enable / disable accounting of open fds Zhang Xiaoxu (2): files_cgroup: fix error pointer when kvm_vm_worker_thread files_cgroup: Fix soft lockup when refcnt overflow. .../admin-guide/kernel-parameters.txt | 7 +- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/Makefile | 1 + fs/file.c | 63 +++- fs/filescontrol.c | 312 ++++++++++++++++++ include/linux/cgroup-defs.h | 8 +- include/linux/cgroup_subsys.h | 6 + include/linux/fdtable.h | 1 + include/linux/filescontrol.h | 40 +++ init/Kconfig | 10 + 11 files changed, 442 insertions(+), 8 deletions(-) create mode 100644 fs/filescontrol.c create mode 100644 include/linux/filescontrol.h -- 2.34.1

2 10

2024

2023

2022

2021

2020

2019

Kernel December 2023