- Kernel - mailweb.openeuler.org

[PATCH OLK-6.6 0/2] cgroup: Support iocost for cgroup v1
by Li Nan 10 Jan '24

10 Jan '24

Li Nan (2): cgroup: Support iocost for cgroup v1 openeuler_defconfig: enable iocost in openeuler_defconfig for x86 and arm64 block/blk-iocost.c | 27 ++++++++++++++++++++++++++ arch/arm64/configs/openeuler_defconfig | 4 +++- arch/x86/configs/openeuler_defconfig | 4 +++- block/Kconfig | 10 ++++++++++ 4 files changed, 43 insertions(+), 2 deletions(-) -- 2.39.2

2 3

[PATCH OLK-6.6 1/3] KVM: arm64: Adjust entry/exit and trap related tracepoints
by lishusen 10 Jan '24

10 Jan '24

virt inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I4IZPY CVE: NA ------------------------------------------------- Currently, we use trace_kvm_exit() to report exception type (e.g., "IRQ", "TRAP") and exception class (ESR_ELx's bit[31:26]) together. But hardware only saves the exit class to ESR_ELx on synchronous exceptions, not on asynchronous exceptions. When the guest exits due to external interrupts, we will get tracing output like: "kvm_exit: IRQ: HSR_EC: 0x0000 (UNKNOWN), PC: 0xffff87259e30" Obviously, "HSR_EC" here is meaningless. This patch splits "exit" and "trap" events by adding two tracepoints explicitly in handle_trap_exceptions(). Let trace_kvm_exit() report VM exit events, and trace_kvm_trap_exit() report VM trap events. These tracepoints are adjusted also in preparation for supporting 'perf kvm stat' on arm64. Signed-off-by: Zenghui Yu <yuzenghui(a)huawei.com> Reviewed-by: Hailiang Zhang <zhang.zhanghailiang(a)huawei.com> Signed-off-by: Zenghui Yu <yuzenghui(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> Link: https://lore.kernel.org/r/1560330526-15468-3-git-send-email-yuzenghui@huawe… Link: https://gitee.com/openeuler/kernel/commit/14b85d8d7d2d Reviewed-by: Yanan Wang <wangyanan55(a)huawei.com> Signed-off-by: Zheng Zengkai <zhengzengkai(a)huawei.com> Signed-off-by: lishusen <lishusen2(a)huawei.com> --- arch/arm64/kvm/arm.c | 4 ++-- arch/arm64/kvm/handle_exit.c | 3 +++ arch/arm64/kvm/trace_arm.h | 21 +++++++++--------- arch/arm64/kvm/trace_handle_exit.h | 35 ++++++++++++++++++++++++++++++ 4 files changed, 51 insertions(+), 12 deletions(-) diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index 70fc3d00f82c..db1ea653a471 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -1029,7 +1029,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu) /************************************************************** * Enter the guest */ - trace_kvm_entry(*vcpu_pc(vcpu)); + trace_kvm_entry(vcpu->vcpu_id, *vcpu_pc(vcpu)); guest_timing_enter_irqoff(); ret = kvm_arm_vcpu_enter_exit(vcpu); @@ -1086,7 +1086,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu) local_irq_enable(); - trace_kvm_exit(ret, kvm_vcpu_trap_get_class(vcpu), *vcpu_pc(vcpu)); + trace_kvm_exit(vcpu->vcpu_id, ret, *vcpu_pc(vcpu)); /* Exit types that need handling before we can be preempted */ handle_exit_early(vcpu, ret); diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c index 617ae6dea5d5..a8e370435e54 100644 --- a/arch/arm64/kvm/handle_exit.c +++ b/arch/arm64/kvm/handle_exit.c @@ -309,7 +309,10 @@ static int handle_trap_exceptions(struct kvm_vcpu *vcpu) exit_handle_fn exit_handler; exit_handler = kvm_get_exit_handler(vcpu); + trace_kvm_trap_enter(vcpu->vcpu_id, + kvm_vcpu_trap_get_class(vcpu)); handled = exit_handler(vcpu); + trace_kvm_trap_exit(vcpu->vcpu_id); } return handled; diff --git a/arch/arm64/kvm/trace_arm.h b/arch/arm64/kvm/trace_arm.h index 8ad53104934d..4a87477e2261 100644 --- a/arch/arm64/kvm/trace_arm.h +++ b/arch/arm64/kvm/trace_arm.h @@ -13,40 +13,41 @@ * Tracepoints for entry/exit to guest */ TRACE_EVENT(kvm_entry, - TP_PROTO(unsigned long vcpu_pc), - TP_ARGS(vcpu_pc), + TP_PROTO(unsigned int vcpu_id, unsigned long vcpu_pc), + TP_ARGS(vcpu_id, vcpu_pc), TP_STRUCT__entry( + __field( unsigned int, vcpu_id ) __field( unsigned long, vcpu_pc ) ), TP_fast_assign( + __entry->vcpu_id = vcpu_id; __entry->vcpu_pc = vcpu_pc; ), - TP_printk("PC: 0x%016lx", __entry->vcpu_pc) + TP_printk("VCPU %u: PC=0x%016lx", __entry->vcpu_id, __entry->vcpu_pc) ); TRACE_EVENT(kvm_exit, - TP_PROTO(int ret, unsigned int esr_ec, unsigned long vcpu_pc), - TP_ARGS(ret, esr_ec, vcpu_pc), + TP_PROTO(unsigned int vcpu_id, int ret, unsigned long vcpu_pc), + TP_ARGS(vcpu_id, ret, vcpu_pc), TP_STRUCT__entry( + __field( unsigned int, vcpu_id ) __field( int, ret ) - __field( unsigned int, esr_ec ) __field( unsigned long, vcpu_pc ) ), TP_fast_assign( + __entry->vcpu_id = vcpu_id; __entry->ret = ARM_EXCEPTION_CODE(ret); - __entry->esr_ec = ARM_EXCEPTION_IS_TRAP(ret) ? esr_ec : 0; __entry->vcpu_pc = vcpu_pc; ), - TP_printk("%s: HSR_EC: 0x%04x (%s), PC: 0x%016lx", + TP_printk("VCPU %u: exit_type=%s, PC=0x%016lx", + __entry->vcpu_id, __print_symbolic(__entry->ret, kvm_arm_exception_type), - __entry->esr_ec, - __print_symbolic(__entry->esr_ec, kvm_arm_exception_class), __entry->vcpu_pc) ); diff --git a/arch/arm64/kvm/trace_handle_exit.h b/arch/arm64/kvm/trace_handle_exit.h index 064a58c19f48..cda43f243f7e 100644 --- a/arch/arm64/kvm/trace_handle_exit.h +++ b/arch/arm64/kvm/trace_handle_exit.h @@ -8,6 +8,41 @@ #undef TRACE_SYSTEM #define TRACE_SYSTEM kvm +TRACE_EVENT(kvm_trap_enter, + TP_PROTO(unsigned int vcpu_id, unsigned int esr_ec), + TP_ARGS(vcpu_id, esr_ec), + + TP_STRUCT__entry( + __field(unsigned int, vcpu_id) + __field(unsigned int, esr_ec) + ), + + TP_fast_assign( + __entry->vcpu_id = vcpu_id; + __entry->esr_ec = esr_ec; + ), + + TP_printk("VCPU %u: HSR_EC=0x%04x (%s)", + __entry->vcpu_id, + __entry->esr_ec, + __print_symbolic(__entry->esr_ec, kvm_arm_exception_class)) +); + +TRACE_EVENT(kvm_trap_exit, + TP_PROTO(unsigned int vcpu_id), + TP_ARGS(vcpu_id), + + TP_STRUCT__entry( + __field(unsigned int, vcpu_id) + ), + + TP_fast_assign( + __entry->vcpu_id = vcpu_id; + ), + + TP_printk("VCPU %u", __entry->vcpu_id) +); + TRACE_EVENT(kvm_wfx_arm64, TP_PROTO(unsigned long vcpu_pc, bool is_wfe), TP_ARGS(vcpu_pc, is_wfe), -- 2.33.0

2 3

[PATCH OLK-6.6] sched/core: Change depends of SCHED_CORE
by Xia Fukun 09 Jan '24

09 Jan '24

From: Lin Shengwang <linshengwang1(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8UVUS CVE: NA -------------------------------------------------------------------------- Due to the conflict between Core Scheduling and SMT expeller, so SCHED_CORE and QOS_SCHED_SMT_EXPELLER are mutually exclusive. Signed-off-by: Lin Shengwang <linshengwang1(a)huawei.com> Signed-off-by: Xia Fukun <xiafukun(a)huawei.com> --- kernel/Kconfig.preempt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/Kconfig.preempt b/kernel/Kconfig.preempt index c2f1fd95a821..dc2a630f200a 100644 --- a/kernel/Kconfig.preempt +++ b/kernel/Kconfig.preempt @@ -117,7 +117,7 @@ config PREEMPT_DYNAMIC config SCHED_CORE bool "Core Scheduling for SMT" - depends on SCHED_SMT + depends on SCHED_SMT && !QOS_SCHED_SMT_EXPELLER help This option permits Core Scheduling, a means of coordinated task selection across SMT siblings. When enabled -- see -- 2.34.1

2 1

[PATCH OLK-6.6 v4 0/8] introduce smart_grid zone
by Yipeng Zou 09 Jan '24

09 Jan '24

The patch sets include two parts: 1. patch 1~3: Rebase smart_grid from OLK-5.10 to OLK-6.6 2. patch 4~7: introduce smart_grid zone qos and cpufreq Since V3: 1. fix some review comments Since V2: 1. move new sysctl form sysctl.c to fair.c 2. add smart_grid cmdline Since V1: 1. Don't init smart_grid cpufreq with dynamic_affinity disable. 2. Forbid smart_grid cgroup interface with dynamic_affinity disable. Hui Tang (1): sched: Introduce smart grid scheduling strategy for cfs Wang ShaoBo (2): sched: smart grid: init sched_grid_qos structure on QOS purpose config: enable CONFIG_QOS_SCHED_SMART_GRID by default Yipeng Zou (5): sched: introduce smart grid qos zone smart_grid: introduce /proc/pid/smart_grid_level smart_grid: introduce smart_grid_strategy_ctrl sysctl smart_grid: cpufreq: introduce smart_grid cpufreq control smart_grid: introduce smart_grid cmdline arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + drivers/cpufreq/cpufreq.c | 244 +++++++++++ fs/proc/array.c | 16 + fs/proc/base.c | 84 ++++ include/linux/cpufreq.h | 11 + include/linux/sched.h | 31 ++ include/linux/sched/grid_qos.h | 135 +++++++ init/Kconfig | 13 + kernel/fork.c | 15 + kernel/sched/Makefile | 1 + kernel/sched/core.c | 181 +++++++++ kernel/sched/fair.c | 537 ++++++++++++++++++++++++- kernel/sched/grid/Makefile | 2 + kernel/sched/grid/internal.h | 6 + kernel/sched/grid/power.c | 27 ++ kernel/sched/grid/qos.c | 279 +++++++++++++ kernel/sched/grid/stat.c | 47 +++ kernel/sched/sched.h | 48 ++- mm/mempolicy.c | 11 +- 20 files changed, 1679 insertions(+), 11 deletions(-) create mode 100644 include/linux/sched/grid_qos.h create mode 100644 kernel/sched/grid/Makefile create mode 100644 kernel/sched/grid/internal.h create mode 100644 kernel/sched/grid/power.c create mode 100644 kernel/sched/grid/qos.c create mode 100644 kernel/sched/grid/stat.c -- 2.34.1

2 9

[PATCH OLK-6.6] sched/smt: fix unbalance sched_smt_present dec/inc
by Xia Fukun 09 Jan '24

09 Jan '24

From: Yang Yingliang <yangyingliang(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8UR4U CVE: NA -------------------------------- If cpuset_cpu_inactive() fails in sched_cpu_deactivate(), the cpu offline failed, but sched_smt_present is decreased before calling sched_cpu_deactivate, it leads unbalance dec/inc, so fix it by increasing sched_smt_present in the error path. Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> Signed-off-by: Yu Liao <liaoyu15(a)huawei.com> Signed-off-by: Xia Fukun <xiafukun(a)huawei.com> --- kernel/sched/core.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 7a0997e7e136..adecea9c607d 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -9792,6 +9792,10 @@ int sched_cpu_deactivate(unsigned int cpu) ret = cpuset_cpu_inactive(cpu); if (ret) { balance_push_set(cpu, false); +#ifdef CONFIG_SCHED_SMT + if (cpumask_weight(cpu_smt_mask(cpu)) == 2) + static_branch_inc_cpuslocked(&sched_smt_present); +#endif set_cpu_active(cpu, true); sched_update_numa(cpu, true); return ret; -- 2.34.1

2 1

[PATCH OLK-6.6] kabi: Reserve space for perf subsystem related structures
by Yang Jihong 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8UOCZ -------------------------------- Reserve kabi space for `struct pmu`, `struct perf_event`, `struct perf_event_context`, `struct perf_cgroup_info` and `struct perf_sample_data`. Signed-off-by: Yang Jihong <yangjihong1(a)huawei.com> --- include/linux/perf_event.h | 31 +++++++++++++++++++++++++++++++ 1 file changed, 31 insertions(+) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index e846f87e2d09..a24aa95de1fc 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -62,6 +62,7 @@ struct perf_guest_info_callbacks { #include <linux/security.h> #include <linux/static_call.h> #include <linux/lockdep.h> +#include <linux/kabi.h> #include <asm/local.h> struct perf_callchain_entry { @@ -324,6 +325,11 @@ struct pmu { /* number of address filters this PMU can do */ unsigned int nr_addr_filters; + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + /* * Fully disable/enable this PMU, can be used to protect from the PMI * as well as for lazy/batch writing of the MSRs. @@ -540,6 +546,11 @@ struct pmu { * Check period value for PERF_EVENT_IOC_PERIOD ioctl. */ int (*check_period) (struct perf_event *event, u64 value); /* optional */ + + KABI_RESERVE(5) + KABI_RESERVE(6) + KABI_RESERVE(7) + KABI_RESERVE(8) }; enum perf_addr_filter_action_t { @@ -839,6 +850,13 @@ struct perf_event { * user. */ __u32 orig_type; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + KABI_RESERVE(5) + KABI_RESERVE(6) #endif /* CONFIG_PERF_EVENTS */ }; @@ -965,6 +983,11 @@ struct perf_event_context { * that until the signal is delivered. */ local_t nr_pending; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; /* @@ -1040,6 +1063,9 @@ struct perf_cgroup_info { u64 timestamp; u64 timeoffset; int active; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct perf_cgroup { @@ -1192,6 +1218,11 @@ struct perf_sample_data { u64 data_page_size; u64 code_page_size; u64 aux_size; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) } ____cacheline_aligned; /* default value for data source */ -- 2.34.1

2 1

[PATCH OLK-6.6] xfs: update the last_sync_lsn with ctx start lsn
by Long Li 09 Jan '24

09 Jan '24

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8UTJ5 CVE: NA -------------------------------- While performing the io fault injection test, I caught the following data corruption report: XFS (dm-6): Internal error ltbno + ltlen > bno at line 1976 of file fs/xfs/libxfs/xfs_alloc.c. Caller xfs_free_ag_extent+0x40b/0x930 [xfs] CPU: 7 PID: 184267 Comm: kworker/7:1 Kdump: loaded Tainted: G O 5.10.0-136.12.0.86.h1179.eulerosv2r12.x86_64 #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58-20211121_093514-szxrtosci10000 04/01/2014 Workqueue: xfs-inodegc/dm-6 xfs_inodegc_worker [xfs] Call Trace: dump_stack+0x57/0x6e xfs_corruption_error+0x81/0x90 [xfs] xfs_free_ag_extent+0x43c/0x930 [xfs] xfs_free_agfl_block+0x3b/0xd0 [xfs] xfs_agfl_free_finish_item+0x14c/0x160 [xfs] xfs_defer_finish_one+0xd5/0x220 [xfs] xfs_defer_finish_noroll+0xb5/0x210 [xfs] xfs_defer_finish+0x11/0x70 [xfs] xfs_itruncate_extents_flags+0xc1/0x240 [xfs] xfs_inactive_truncate+0xab/0xe0 [xfs] xfs_inactive+0x154/0x170 [xfs] xfs_inodegc_inactivate+0x16/0x50 [xfs] xfs_inodegc_worker+0xa0/0x110 [xfs] process_one_work+0x1b5/0x350 worker_thread+0x49/0x310 kthread+0xfe/0x140 ret_from_fork+0x22/0x30 XFS (dm-6): Corruption detected. Unmount and run xfs_repair After analyzing the disk image, I found that the cause of the problem was that the transactions were not replayed. The problem arises in that the iclog buffer IO completion updates the l_last_sync_lsn with it's own LSN. Transactions can be large enough to span many iclogs, only commit iclog goes to update l_last_sync_lsn. Since the last_sync_lsn update and the insertion of the item into the ail list releases the l_icloglock, if the ail is is empty in the meantime, the new iclog gets the last_sync_lsn as tail lsn. If the new iclog is written to disk and a shutdown occurs, the current iclog will not be able to replay in the next mount. xlog_state_done_syncing xlog_state_do_callback spin_lock(&log->l_icloglock); xlog_state_do_iclog_callbacks xlog_state_iodone_process_iclog xlog_state_set_callback last_sync_lsn = iclog->ic_header.h_lsn spin_unlock(&log->l_icloglock); ====>AIL is empty and get tail lsn for new iclog xlog_cil_process_committed xlog_cil_committed xfs_trans_committed_bulk(ctx->start_lsn) xfs_log_item_batch_insert(commit_lsn) xlog_state_clean_iclog(log, iclog) spin_lock(&log->l_icloglock); spin_unlock(&log->l_icloglock); Fix is simple, updates the l_last_sync_lsn with it's first ctx start lsn when commit iclog buffer IO completion. Even if the above happens, the iclog will be replayed as well. Signed-off-by: Long Li <leo.lilong(a)huawei.com> --- fs/xfs/xfs_log.c | 19 +++++++++++-------- 1 file changed, 11 insertions(+), 8 deletions(-) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index bd167a94a6b3..96416b6ca013 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -2704,19 +2704,22 @@ xlog_get_lowest_lsn( static void xlog_state_set_callback( struct xlog *log, - struct xlog_in_core *iclog, - xfs_lsn_t header_lsn) + struct xlog_in_core *iclog) { + struct xfs_cil_ctx *ctx; + trace_xlog_iclog_callback(iclog, _RET_IP_); iclog->ic_state = XLOG_STATE_CALLBACK; - ASSERT(XFS_LSN_CMP(atomic64_read(&log->l_last_sync_lsn), - header_lsn) <= 0); - - if (list_empty_careful(&iclog->ic_callbacks)) + ctx = list_first_entry_or_null(&iclog->ic_callbacks, + struct xfs_cil_ctx, iclog_entry); + if (!ctx) return; - atomic64_set(&log->l_last_sync_lsn, header_lsn); + ASSERT(XFS_LSN_CMP(atomic64_read(&log->l_last_sync_lsn), + ctx->start_lsn) <= 0); + + atomic64_set(&log->l_last_sync_lsn, ctx->start_lsn); xlog_grant_push_ail(log, 0); } @@ -2751,7 +2754,7 @@ xlog_state_iodone_process_iclog( lowest_lsn = xlog_get_lowest_lsn(log); if (lowest_lsn && XFS_LSN_CMP(lowest_lsn, header_lsn) < 0) return false; - xlog_state_set_callback(log, iclog, header_lsn); + xlog_state_set_callback(log, iclog); return false; default: /* -- 2.31.1

2 1

[PATCH OLK-6.6 00/10] Backport etmem feature to OLK 6.6
by Yuchen Tang 09 Jan '24

09 Jan '24

v2 -> v3 - fix code style and commit messages v1 -> v2 - make independent patches to export symbols we need - make independent patches for etmem scan and swap config - fix indentation and code style problems Yuchen Tang (10): etmem: add ETMEM feature CONFIG to mm/Kconfig etmem: add ETMEM scan feature CONFIG to mm/Kconfig mm: Export symbol __pud_trans_huge_lock mm: Export symbol walk_page_range etmem: add etmem scan feature mm: Export symbol reclaim_pages etmem: add etmem swap feature etmem: add original kernel swap enabled options etmem: add swapcache reclaim to etmem etmem: enable etmem configurations arch/arm64/configs/openeuler_defconfig | 3 + arch/x86/configs/openeuler_defconfig | 3 + fs/proc/Makefile | 3 + fs/proc/base.c | 8 + fs/proc/etmem_proc.c | 216 ++++ fs/proc/etmem_scan.c | 1383 ++++++++++++++++++++++++ fs/proc/etmem_scan.h | 149 +++ fs/proc/etmem_swap.c | 283 +++++ fs/proc/internal.h | 4 + include/linux/etmem.h | 75 ++ include/linux/mm.h | 4 + include/linux/mm_types.h | 5 + include/linux/swap.h | 1 + include/uapi/asm-generic/mman-common.h | 4 + mm/Kconfig | 33 + mm/Makefile | 1 + mm/etmem.c | 384 +++++++ mm/huge_memory.c | 1 + mm/internal.h | 1 - mm/madvise.c | 16 + mm/pagewalk.c | 1 + mm/swap_state.c | 4 + mm/vmscan.c | 16 + virt/kvm/kvm_main.c | 8 + 24 files changed, 2605 insertions(+), 1 deletion(-) create mode 100644 fs/proc/etmem_proc.c create mode 100644 fs/proc/etmem_scan.c create mode 100644 fs/proc/etmem_scan.h create mode 100644 fs/proc/etmem_swap.c create mode 100644 include/linux/etmem.h create mode 100644 mm/etmem.c -- 2.33.0

3 12

[PATCH OLK-6.6] hfs: fix null-ptr-deref in hfs_find_init()
by Long Li 09 Jan '24

09 Jan '24

From: Yang Yingliang <yangyingliang(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8URA4 CVE: CVE-2018-12928 -------------------------------- It will cause a null-ptr-deref in hfs_find_init() in fuzz test. [ 107.092729] hfs: continuing without an alternate MDB [ 107.097632] general protection fault, probably for non-canonical address 0xdffffc0000000008: 0000 [#1] SMP KASAN PTI [ 107.104679] KASAN: null-ptr-deref in range [0x0000000000000040-0x0000000000000047] [ 107.109100] CPU: 0 PID: 379 Comm: hfs_inject Not tainted 5.7.0-rc7-00001-g24627f5f2973 #897 [ 107.114142] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014 [ 107.121095] RIP: 0010:hfs_find_init+0x72/0x170 [ 107.123609] Code: c1 ea 03 80 3c 02 00 0f 85 e6 00 00 00 4c 8d 65 40 48 c7 43 18 00 00 00 00 48 b8 00 00 00 00 00 fc ff df 4c 89 e2 48 c1 ea 03 <0f> b6 04 02 84 c0 74 08 3c 03 0f 8e a5 00 00 00 8b 45 40 be c0 0c [ 107.134660] RSP: 0018:ffff88810291f3f8 EFLAGS: 00010202 [ 107.137897] RAX: dffffc0000000000 RBX: ffff88810291f468 RCX: 1ffff110175cdf05 [ 107.141874] RDX: 0000000000000008 RSI: ffff88810291f468 RDI: ffff88810291f480 [ 107.145844] RBP: 0000000000000000 R08: 0000000000000000 R09: ffffed1020381013 [ 107.149431] R10: ffff88810291f500 R11: ffffed1020381012 R12: 0000000000000040 [ 107.152315] R13: 0000000000000000 R14: ffff888101c0814a R15: ffff88810291f468 [ 107.155464] FS: 00000000009ea880(0000) GS:ffff88810c600000(0000) knlGS:0000000000000000 [ 107.159795] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 107.162987] CR2: 00005605a19dd284 CR3: 0000000103a0c006 CR4: 0000000000020ef0 [ 107.166665] Call Trace: [ 107.167969] ? find_held_lock+0x33/0x1c0 [ 107.169972] hfs_ext_read_extent+0x16b/0xb00 [ 107.172092] ? create_page_buffers+0x14e/0x1b0 [ 107.174303] ? hfs_free_extents+0x280/0x280 [ 107.176437] ? lock_downgrade+0x730/0x730 [ 107.178272] hfs_get_block+0x496/0x8a0 [ 107.179972] block_read_full_page+0x241/0x8d0 [ 107.181971] ? hfs_extend_file+0xae0/0xae0 [ 107.183814] ? end_buffer_async_read_io+0x10/0x10 [ 107.185954] ? add_to_page_cache_lru+0x13f/0x1f0 [ 107.188006] ? add_to_page_cache_locked+0x10/0x10 [ 107.190175] do_read_cache_page+0xc6a/0x1180 [ 107.192096] ? generic_file_read_iter+0x4c0/0x4c0 [ 107.194234] ? hfs_btree_open+0x408/0x1000 [ 107.196068] ? lock_downgrade+0x730/0x730 [ 107.197926] ? wake_bit_function+0x180/0x180 [ 107.199845] ? lockdep_init_map_waits+0x267/0x7c0 [ 107.201895] hfs_btree_open+0x455/0x1000 [ 107.203479] hfs_mdb_get+0x122c/0x1ae8 [ 107.205065] ? hfs_mdb_put+0x350/0x350 [ 107.206590] ? queue_work_node+0x260/0x260 [ 107.208309] ? rcu_read_lock_sched_held+0xa1/0xd0 [ 107.210227] ? lockdep_init_map_waits+0x267/0x7c0 [ 107.212144] ? lockdep_init_map_waits+0x267/0x7c0 [ 107.213979] hfs_fill_super+0x9ba/0x1280 [ 107.215444] ? bdev_name.isra.9+0xf1/0x2b0 [ 107.217028] ? hfs_remount+0x190/0x190 [ 107.218428] ? pointer+0x5da/0x710 [ 107.219745] ? file_dentry_name+0xf0/0xf0 [ 107.221262] ? mount_bdev+0xd1/0x330 [ 107.222592] ? vsnprintf+0x7bd/0x1250 [ 107.224007] ? pointer+0x710/0x710 [ 107.225332] ? down_write+0xe5/0x160 [ 107.226698] ? hfs_remount+0x190/0x190 [ 107.228120] ? snprintf+0x91/0xc0 [ 107.229388] ? vsprintf+0x10/0x10 [ 107.230628] ? sget+0x3af/0x4a0 [ 107.231848] ? hfs_remount+0x190/0x190 [ 107.233300] mount_bdev+0x26e/0x330 [ 107.234611] ? hfs_statfs+0x540/0x540 [ 107.236015] legacy_get_tree+0x101/0x1f0 [ 107.237431] ? security_capable+0x58/0x90 [ 107.238832] vfs_get_tree+0x89/0x2d0 [ 107.240082] ? ns_capable_common+0x5c/0xd0 [ 107.241521] do_mount+0xd8a/0x1720 [ 107.242727] ? lock_downgrade+0x730/0x730 [ 107.244116] ? copy_mount_string+0x20/0x20 [ 107.245557] ? _copy_from_user+0xbe/0x100 [ 107.246967] ? memdup_user+0x47/0x70 [ 107.248212] __x64_sys_mount+0x162/0x1b0 [ 107.249537] do_syscall_64+0xa5/0x4f0 [ 107.250742] entry_SYSCALL_64_after_hwframe+0x49/0xb3 [ 107.252369] RIP: 0033:0x44e8ea [ 107.253360] Code: 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 49 89 ca b8 a5 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48 [ 107.259240] RSP: 002b:00007ffd910e4c28 EFLAGS: 00000207 ORIG_RAX: 00000000000000a5 [ 107.261668] RAX: ffffffffffffffda RBX: 0000000000400400 RCX: 000000000044e8ea [ 107.263920] RDX: 000000000049321e RSI: 0000000000493222 RDI: 00007ffd910e4d00 [ 107.266177] RBP: 00007ffd910e5d10 R08: 0000000000000000 R09: 000000000000000a [ 107.268451] R10: 0000000000000001 R11: 0000000000000207 R12: 0000000000401c40 [ 107.270721] R13: 0000000000000000 R14: 00000000006ba018 R15: 0000000000000000 [ 107.273025] Modules linked in: [ 107.274029] Dumping ftrace buffer: [ 107.275121] (ftrace buffer empty) [ 107.276370] ---[ end trace c5e0b9d684f3570e ]--- We need check tree in hfs_find_init(). https://lore.kernel.org/linux-fsdevel/20180419024358.GA5215@bombadil.infrad… https://marc.info/?l=linux-fsdevel&m=152406881024567&w=2 References: CVE-2018-12928 Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> Signed-off-by: Long Li <leo.lilong(a)huawei.com> --- fs/hfs/bfind.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/hfs/bfind.c b/fs/hfs/bfind.c index ef9498a6e88a..ae44c8fa86ec 100644 --- a/fs/hfs/bfind.c +++ b/fs/hfs/bfind.c @@ -16,6 +16,8 @@ int hfs_find_init(struct hfs_btree *tree, struct hfs_find_data *fd) { void *ptr; + if (!tree) + return -EINVAL; fd->tree = tree; fd->bnode = NULL; ptr = kmalloc(tree->max_key_len * 2 + 4, GFP_KERNEL); -- 2.31.1

2 1

[PATCH OLK-6.6 v6 0/5] block: Add config to show info about opening a mounted device for write
by Li Lingfeng 09 Jan '24

09 Jan '24

v1->v2: 1) Replace BLK_DEV_DUMPINFO with BLK_DEV_WRITE_MOUNTED_QUIET, which is logically opposite; 2) Make BLK_DEV_WRITE_MOUNTED_QUIET dependent on BLK_DEV_WRITE_MOUNTED, and decide whether to set bd_writers by BLK_DEV_WRITE_MOUNTED_QUIET; 3) Move bdev_dump_info() into bdev_may_open(); 4) Add the cmdline interface to control whether to show info; 5) Pass the state of "writes blocked" to part0 when mounting a partition. v2->v3: Add restriction for modifying bd_writers. v3->v4: 1) Count bd_writers regardless of whether bdev_allow_write_mounted is set; 2) Expand the meaning of bdev_allow_write_mounted; 3) Add config option to detect writing to part0 while partitions mounted; 4) Add bd_mounters to record the number of times that part0 or partition is mounted; 5) Show info about opening a lower device for write while upper-layers mounted. v4->v5: Move introduction of bd_mounters to patch1. v5->v6: Delete redundant conflict judgment. Li Lingfeng (5): block: Count writers and mounters regardless of whether bdev_allow_write_mounted is set block: Expand the meaning of bdev_allow_write_mounted block: Add config option to detect writing to part0 while partitions mounted block: Add config option to show info about opening a mounted device for write block: Show info about opening a lower device for write while upper-layers mounted block/Kconfig | 19 +++++++ block/bdev.c | 104 ++++++++++++++++++++++++++++++++------ include/linux/blk_types.h | 1 + 3 files changed, 109 insertions(+), 15 deletions(-) -- 2.31.1

2 6

[PATCH V4 OLK-6.6] fs/dirty_pages: dump the number of dirty pages for each inode
by Zizhi Wo 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OHM4?from=project-issue CVE: NA ---------------------------------------- In order to analyse the IO performance when using buffer IO, it's useful to obtain the number of dirty pages for each inode in the filesystem. This feature depends on 'CONFIG_DIRTY_PAGES'. It creates 3 interfaces by using procfs. /proc/dirty/buffer_size for buffer allocation and release, /proc/dirty/page_threshold to filter result and /proc/dirty/dirty_list to get dirty pages. Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/Kconfig | 13 + fs/Makefile | 1 + fs/dirty_pages.c | 355 +++++++++++++++++++++++++ 5 files changed, 371 insertions(+) create mode 100644 fs/dirty_pages.c diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index a417275129a4..13ddaf47cff2 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -6925,6 +6925,7 @@ CONFIG_PROC_PAGE_MONITOR=y CONFIG_PROC_CHILDREN=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index c09c7fe35a54..84727e7ad415 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -8121,6 +8121,7 @@ CONFIG_PROC_PID_ARCH_STATUS=y CONFIG_PROC_CPU_RESCTRL=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/fs/Kconfig b/fs/Kconfig index 7c9d45ba9b0f..6dc83223c61b 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -170,6 +170,19 @@ source "fs/proc/Kconfig" source "fs/kernfs/Kconfig" source "fs/sysfs/Kconfig" +config DIRTY_PAGES + bool "Dumps the number of dirty pages of each file" + depends on PROC_FS + default y + help + This config supports the rendering of dirty page data to the user, + which may be useful to analyze the IO performance when using buffer + IO. + + It create 3 interfaces by using procfs. /proc/dirty/buffer_size for + buffer allocation and release; /proc/dirty/page_threshold to filter + result; /proc/dirty/dirty_list to get dirty pages. + config TMPFS bool "Tmpfs virtual memory file system support (former shm fs)" depends on SHMEM diff --git a/fs/Makefile b/fs/Makefile index bb6164200a39..cc4735c23c32 100644 --- a/fs/Makefile +++ b/fs/Makefile @@ -44,6 +44,7 @@ obj-$(CONFIG_FS_POSIX_ACL) += posix_acl.o obj-$(CONFIG_NFS_COMMON) += nfs_common/ obj-$(CONFIG_COREDUMP) += coredump.o obj-$(CONFIG_SYSCTL) += drop_caches.o sysctls.o +obj-$(CONFIG_DIRTY_PAGES) += dirty_pages.o obj-$(CONFIG_FHANDLE) += fhandle.o obj-$(CONFIG_CGROUP_FILES) += filescontrol.o diff --git a/fs/dirty_pages.c b/fs/dirty_pages.c new file mode 100644 index 000000000000..712992cba9e6 --- /dev/null +++ b/fs/dirty_pages.c @@ -0,0 +1,355 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/proc_fs.h> +#include <linux/seq_file.h> +#include <linux/uaccess.h> +#include <linux/pagemap.h> +#include <linux/pagevec.h> +#include <linux/fs.h> +#include <linux/mm.h> +#include <linux/slab.h> +#include <linux/sched.h> +#include <linux/proc_fs.h> +#include <linux/kdev_t.h> +#include <linux/vmalloc.h> +#include <linux/module.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include "internal.h" + +#define DIRTY_ROOT "dirty" /* proc root directory name */ +#define DIRTY_BUFFER_SIZE "buffer_size" /* proc file to control buffer */ +#define DIRTY_PAGES "dirty_list" /* proc file to obtain diry pages */ +#define DIRTY_LIMIT "page_threshold" /* proc file to filter result */ +#define MAX_BUF_NUM 102400 /* max buffer num in number of pages */ +#define TERM_STR "terminated\n" /* tail string for overflowed buffer */ +#define UNKNOW_STR "unknown" /* for unknown file name or fs type */ + +static char *buf_dirty; /* buffer to store number of dirty pages */ +static unsigned long buf_size; /* size of buffer in bytes */ +static unsigned long buf_num; /* size of buffer in number of pages */ +static unsigned long buf_limit; /* filter threshold of dirty pages */ +static unsigned long lock_word; /* for exclusive access to buffer */ + +static bool dirty_pages_lock(void) +{ + if (xchg(&lock_word, 1) == 1) + return false; + return true; +} + +static void dirty_pages_unlock(void) +{ + lock_word = 0; +} + +static unsigned long dump_dirtypages_inode(struct inode *inode) +{ + XA_STATE(xas, &inode->i_mapping->i_pages, 0); + unsigned long pages = 0; + void *page; + + if (!mapping_tagged(inode->i_mapping, PAGECACHE_TAG_DIRTY)) + return 0; + + xas_lock_irq(&xas); + xas_for_each_marked(&xas, page, (pgoff_t)-1, PAGECACHE_TAG_DIRTY) { + if (++pages % XA_CHECK_SCHED) + continue; + + xas_pause(&xas); + xas_unlock_irq(&xas); + cond_resched(); + xas_lock_irq(&xas); + } + xas_unlock_irq(&xas); + + return pages; +} + +static char *inode_filename(struct inode *inode) +{ + char *path; + static char buf[PATH_MAX]; + struct dentry *dentry; + + dentry = d_find_alias(inode); + if (!dentry) + return UNKNOW_STR; + + path = dentry_path_raw(dentry, buf, PATH_MAX); + dput(dentry); + + return IS_ERR_OR_NULL(path) ? UNKNOW_STR : path; +} + +static const char *sb_fstype(struct super_block *sb) +{ + if (sb->s_type && sb->s_type->name) + return sb->s_type->name; + else + return UNKNOW_STR; +} + +static inline bool is_sb_writable(struct super_block *sb) +{ + if (sb_rdonly(sb)) + return false; + + if (sb->s_writers.frozen == SB_FREEZE_COMPLETE) + return false; + + return true; +} + +/* + * dump_dirtypages_sb - dump the dirty pages of each inode in the sb + * @sb the super block + * @m the seq_file witch is initialized in proc_dpages_open + * + * For each inode in the sb, call dump_dirtypages_pages to get the number + * of dirty pages. And use seq_printf to store the result in the buffer + * if it's not less than the threshold. The inode in unusual state will + * be skipped. + */ +static void dump_dirtypages_sb(struct super_block *sb, struct seq_file *m) +{ + struct inode *inode, *toput_inode = NULL; + unsigned long limit; + + if (!is_sb_writable(sb) || !sb->s_bdev) + return; + + limit = READ_ONCE(buf_limit); + spin_lock(&sb->s_inode_list_lock); + list_for_each_entry(inode, &sb->s_inodes, i_sb_list) { + unsigned long dirty_pages; + + spin_lock(&inode->i_lock); + if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) { + spin_unlock(&inode->i_lock); + continue; + } + + __iget(inode); + spin_unlock(&inode->i_lock); + spin_unlock(&sb->s_inode_list_lock); + + dirty_pages = dump_dirtypages_inode(inode); + if (!dirty_pages || dirty_pages < limit) + goto skip; + + seq_printf(m, "FSType: %s, Dev ID: %u(%u:%u) ino %lu, dirty pages %lu, path %s\n", + sb_fstype(sb), sb->s_dev, MAJOR(sb->s_dev), + MINOR(sb->s_dev), inode->i_ino, dirty_pages, + inode_filename(inode)); + + if (seq_has_overflowed(m)) { + /* keep size > count to avoid overflow in seq_read_iter() */ + m->size += strlen(TERM_STR) + 1; + seq_puts(m, TERM_STR); + iput(inode); + iput(toput_inode); + return; + } +skip: + iput(toput_inode); + toput_inode = inode; + cond_resched(); + spin_lock(&sb->s_inode_list_lock); + } + spin_unlock(&sb->s_inode_list_lock); + iput(toput_inode); +} + +static int proc_dpages_show(struct seq_file *m, void *v) +{ + iterate_supers((void *)dump_dirtypages_sb, (void *)m); + return 0; +} + +static int proc_dpages_open(struct inode *inode, struct file *filp) +{ + int ret; + struct seq_file *m; + + if (!dirty_pages_lock()) + return -EBUSY; + + if (buf_dirty == NULL || buf_size == 0) { + pr_warn("dirty_pages: please allocate buffer before getting dirty pages\n"); + dirty_pages_unlock(); + return -ENOMEM; + } + + ret = single_open(filp, proc_dpages_show, NULL); + if (ret) { + dirty_pages_unlock(); + return ret; + } + + m = (struct seq_file *)filp->private_data; + memset(buf_dirty, 0, buf_size); + m->size = buf_size - (strlen(TERM_STR) + 1); + m->buf = buf_dirty; + + return 0; +} + +static int proc_dpages_release(struct inode *inode, struct file *file) +{ + struct seq_file *m = file->private_data; + + /* we don't want to free the buf */ + m->buf = NULL; + dirty_pages_unlock(); + return single_release(inode, file); +} + +static void free_buf_dirty(void) +{ + if (!buf_dirty) + return; + + vfree(buf_dirty); + buf_dirty = NULL; + buf_size = 0; + WRITE_ONCE(buf_num, 0); +} + +static int proc_buffer_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", READ_ONCE(buf_num)); + return 0; +} + +static int proc_buffer_open(struct inode *inode, struct file *filp) +{ + int ret; + + if (((filp->f_flags & O_ACCMODE) != O_RDONLY) && !dirty_pages_lock()) + return -EBUSY; + + ret = single_open(filp, proc_buffer_show, NULL); + if (ret) + dirty_pages_unlock(); + + return ret; +} + +static ssize_t proc_buffer_write(struct file *filp, const char __user *buf, + size_t count, loff_t *offp) +{ + int ret; + unsigned long new_buf_num, new_buf_size; + char *new_buf_dirty; + + ret = kstrtoul_from_user(buf, count, 10, &new_buf_num); + if (ret != 0 || new_buf_num > MAX_BUF_NUM) + return -EINVAL; + + if (new_buf_num == buf_num) + return count; + + if (new_buf_num == 0) { + free_buf_dirty(); + return count; + } + + new_buf_size = PAGE_SIZE * new_buf_num; + new_buf_dirty = vzalloc(new_buf_size); + if (!new_buf_dirty) + return -ENOMEM; + + free_buf_dirty(); + buf_dirty = new_buf_dirty; + buf_size = new_buf_size; + WRITE_ONCE(buf_num, new_buf_num); + + return count; +} + +static int proc_buffer_release(struct inode *inode, struct file *filp) +{ + if ((filp->f_flags & O_ACCMODE) != O_RDONLY) + dirty_pages_unlock(); + return single_release(inode, filp); +} + +static int proc_limit_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", READ_ONCE(buf_limit)); + return 0; +} + +static int proc_limit_open(struct inode *inode, struct file *filp) +{ + return single_open(filp, proc_limit_show, NULL); +} + +static ssize_t proc_limit_write(struct file *filp, const char __user *buf, + size_t count, loff_t *offp) +{ + int ret; + unsigned long limit; + + ret = kstrtoul_from_user(buf, count, 10, &limit); + if (ret != 0) + return ret; + + WRITE_ONCE(buf_limit, limit); + + return count; +} + +static const struct proc_ops proc_dpages_operations = { + .proc_open = proc_dpages_open, + .proc_read = seq_read, + .proc_release = proc_dpages_release, +}; + +static const struct proc_ops proc_buffer_operations = { + .proc_open = proc_buffer_open, + .proc_read = seq_read, + .proc_write = proc_buffer_write, + .proc_lseek = seq_lseek, + .proc_release = proc_buffer_release, +}; + +static const struct proc_ops proc_limit_operations = { + .proc_open = proc_limit_open, + .proc_read = seq_read, + .proc_write = proc_limit_write, + .proc_lseek = seq_lseek, + .proc_release = single_release, +}; + +static int __init dpages_proc_init(void) +{ + struct proc_dir_entry *dirty_dir; + + dirty_dir = proc_mkdir(DIRTY_ROOT, NULL); + if (!dirty_dir) + return -ENOMEM; + + if (!proc_create(DIRTY_PAGES, 0440, dirty_dir, &proc_dpages_operations)) + goto fail_pages; + + if (!proc_create(DIRTY_BUFFER_SIZE, 0640, dirty_dir, + &proc_buffer_operations)) + goto fail_switch; + + if (!proc_create(DIRTY_LIMIT, 0640, dirty_dir, &proc_limit_operations)) + goto fail_limit; + + return 0; + +fail_limit: + remove_proc_entry(DIRTY_BUFFER_SIZE, dirty_dir); +fail_switch: + remove_proc_entry(DIRTY_PAGES, dirty_dir); +fail_pages: + remove_proc_entry(DIRTY_ROOT, NULL); + return -ENOMEM; +} + +subsys_initcall(dpages_proc_init); -- 2.39.2

2 1

[PATCH OLK-6.6 0/2] Some simple extensions of the kfence feature
by Ze Zuo 09 Jan '24

09 Jan '24

This patchset supports the following two features: 1. The number of kfence objects is a key factor to balance kfence bug detection capabilities and performance loss, and often needs to be optimized according to actual business scenarios. The number of kfence objects can be dynamically configured kfence.num_objects cmdline, so as to reduce the release of kernel versions in actual use, thereby reducing costs. 2. The kfence feature under arm64 supports dynamic opening after system startup, which will turn all the kernel memblock memory into page-mapped mapping and increase the memory overhead, in order to save the page table memory usage, only the kfence pool supports page-mapped mapping, which can be enabled at kernel startup by setting kfence.sample _interval=-1 at kernel startup to enable the dynamic on capability, at this time it will allocate memory for kfence by default. The kfence will be allocated memory by default. Subsequent dynamic enablement and disabling will no longer consume additional memory. This feature has been tested under ARM64, x86_64 and ARM architectures, and the Kfence Kunit test results are consistent with the native ones. ChangeLog: - Add config kfence_must_early_init to perform code isolation for non-functional purposes. - Some minor changes about document, no functional changes. Ze Zuo (2): kfence: Add a module parameter to adjust kfence objects arm64: kfence: scale sample_interval to support early init for kfence. Documentation/dev-tools/kfence.rst | 8 +- arch/arm64/include/asm/kfence.h | 3 + arch/arm64/mm/mmu.c | 5 + include/linux/kfence.h | 13 +- lib/Kconfig.kfence | 28 +++++ mm/kfence/core.c | 189 +++++++++++++++++++++++++---- mm/kfence/kfence.h | 4 +- mm/kfence/kfence_test.c | 2 +- 8 files changed, 222 insertions(+), 30 deletions(-) -- 2.25.1

2 3

[PATCH OLK-6.6] pid_ns: Make pid_max per namespace
by Yi Yang 09 Jan '24

09 Jan '24

From: Li Zefan <lizefan(a)huawei.com> euler inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8TCPY ------------------------------------------------- The adjusted pid_max on the host also takes effect in the container. Move pid_max to pid_namespace to fix this problem. Signed-off-by: Li Zefan <lizefan(a)huawei.com> Signed-off-by: Yi Yang <yiyang13(a)huawei.com> --- include/linux/pid.h | 2 ++ include/linux/pid_namespace.h | 3 ++ init/Kconfig | 6 ++++ kernel/pid.c | 57 +++++++++++++++++++++++++++++++++++ kernel/pid_namespace.c | 13 ++++++++ kernel/sysctl.c | 2 ++ kernel/trace/pid_list.c | 4 +++ kernel/trace/trace.c | 4 +++ kernel/trace/trace.h | 2 ++ 9 files changed, 93 insertions(+) diff --git a/include/linux/pid.h b/include/linux/pid.h index 653a527574c4..b90bc447d2a2 100644 --- a/include/linux/pid.h +++ b/include/linux/pid.h @@ -113,8 +113,10 @@ extern void transfer_pid(struct task_struct *old, struct task_struct *new, struct pid_namespace; extern struct pid_namespace init_pid_ns; +#ifndef CONFIG_PID_MAX_PER_NAMESPACE extern int pid_max; extern int pid_max_min, pid_max_max; +#endif /* * look up a PID in the hash table. Must be called with the tasklist_lock diff --git a/include/linux/pid_namespace.h b/include/linux/pid_namespace.h index f9f9931e02d6..0a7c6dab0b7b 100644 --- a/include/linux/pid_namespace.h +++ b/include/linux/pid_namespace.h @@ -36,6 +36,9 @@ struct pid_namespace { #endif struct user_namespace *user_ns; struct ucounts *ucounts; +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + int pid_max; +#endif int reboot; /* group exit code if this pidns was rebooted */ struct ns_common ns; #if defined(CONFIG_SYSCTL) && defined(CONFIG_MEMFD_CREATE) diff --git a/init/Kconfig b/init/Kconfig index 869eea4108d0..8c07d158bb54 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -2104,3 +2104,9 @@ config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE # <asm/syscall_wrapper.h>. config ARCH_HAS_SYSCALL_WRAPPER def_bool n + +config PID_MAX_PER_NAMESPACE + bool "Make pid_max per namespace" + default y + help + Say Y here to enable make pid_max per namespace. diff --git a/kernel/pid.c b/kernel/pid.c index 383abde0c208..4e9267988dd7 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -45,6 +45,10 @@ #include <net/sock.h> #include <uapi/linux/pidfd.h> +#ifdef CONFIG_PID_MAX_PER_NAMESPACE +#include <linux/kmemleak.h> +#endif + struct pid init_struct_pid = { .count = REFCOUNT_INIT(1), .tasks = { @@ -59,12 +63,19 @@ struct pid init_struct_pid = { }, } }; +#ifndef CONFIG_PID_MAX_PER_NAMESPACE int pid_max = PID_MAX_DEFAULT; +#endif #define RESERVED_PIDS 300 +#ifndef CONFIG_PID_MAX_PER_NAMESPACE int pid_max_min = RESERVED_PIDS + 1; int pid_max_max = PID_MAX_LIMIT; +#else +static int pid_max_min = RESERVED_PIDS + 1; +static int pid_max_max = PID_MAX_LIMIT; +#endif /* * PID-map pages start out as NULL, they get allocated upon @@ -80,6 +91,9 @@ struct pid_namespace init_pid_ns = { .child_reaper = &init_task, .user_ns = &init_user_ns, .ns.inum = PROC_PID_INIT_INO, +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + .pid_max = PID_MAX_DEFAULT, +#endif #ifdef CONFIG_PID_NS .ns.ops = &pidns_operations, #endif @@ -194,7 +208,11 @@ struct pid *alloc_pid(struct pid_namespace *ns, pid_t *set_tid, tid = set_tid[ns->level - i]; retval = -EINVAL; +#ifndef CONFIG_PID_MAX_PER_NAMESPACE if (tid < 1 || tid >= pid_max) +#else + if (tid < 1 || tid >= task_active_pid_ns(current)->pid_max) +#endif goto out_free; /* * Also fail if a PID != 1 is requested and @@ -234,7 +252,11 @@ struct pid *alloc_pid(struct pid_namespace *ns, pid_t *set_tid, * a partially initialized PID (see below). */ nr = idr_alloc_cyclic(&tmp->idr, NULL, pid_min, +#ifndef CONFIG_PID_MAX_PER_NAMESPACE pid_max, GFP_ATOMIC); +#else + tmp->pid_max, GFP_ATOMIC); +#endif } spin_unlock_irq(&pidmap_lock); idr_preload_end(); @@ -646,8 +668,37 @@ SYSCALL_DEFINE2(pidfd_open, pid_t, pid, unsigned int, flags) return fd; } +#ifdef CONFIG_PID_MAX_PER_NAMESPACE +static int proc_dointvec_pidmax(struct ctl_table *table, int write, + void __user *buffer, size_t *lenp, loff_t *ppos) +{ + struct ctl_table tmp; + + tmp = *table; + tmp.data = &task_active_pid_ns(current)->pid_max; + + return proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos); +} + +static struct ctl_table pid_ctl_table[] = { + { + .procname = "pid_max", + .data = &init_pid_ns.pid_max, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = proc_dointvec_pidmax, + .extra1 = &pid_max_min, + .extra2 = &pid_max_max, + }, + {} +}; +#endif + void __init pid_idr_init(void) { +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + int pid_max = init_pid_ns.pid_max; +#endif /* Verify no one has done anything silly: */ BUILD_BUG_ON(PID_MAX_LIMIT >= PIDNS_ADDING); @@ -658,6 +709,9 @@ void __init pid_idr_init(void) PIDS_PER_CPU_MIN * num_possible_cpus()); pr_info("pid_max: default: %u minimum: %u\n", pid_max, pid_max_min); +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + init_pid_ns.pid_max = pid_max; +#endif idr_init(&init_pid_ns.idr); init_pid_ns.pid_cachep = kmem_cache_create("pid", @@ -665,6 +719,9 @@ void __init pid_idr_init(void) __alignof__(struct pid), SLAB_HWCACHE_ALIGN | SLAB_PANIC | SLAB_ACCOUNT, NULL); +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + register_sysctl_init("kernel", pid_ctl_table); +#endif } static struct file *__pidfd_fget(struct task_struct *task, int fd) diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index 619972c78774..c837b1096dcb 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -110,6 +110,9 @@ static struct pid_namespace *create_pid_namespace(struct user_namespace *user_ns ns->user_ns = get_user_ns(user_ns); ns->ucounts = ucounts; ns->pid_allocated = PIDNS_ADDING; +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + ns->pid_max = parent_pid_ns->pid_max; +#endif #if defined(CONFIG_SYSCTL) && defined(CONFIG_MEMFD_CREATE) ns->memfd_noexec_scope = pidns_memfd_noexec_scope(parent_pid_ns); #endif @@ -295,6 +298,10 @@ static int pid_ns_ctl_handler(struct ctl_table *table, int write, next = idr_get_cursor(&pid_ns->idr) - 1; tmp.data = &next; +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + tmp.extra2 = &pid_ns->pid_max; +#endif + ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos); if (!ret && write) idr_set_cursor(&pid_ns->idr, next + 1); @@ -302,7 +309,9 @@ static int pid_ns_ctl_handler(struct ctl_table *table, int write, return ret; } +#ifndef CONFIG_PID_MAX_PER_NAMESPACE extern int pid_max; +#endif static struct ctl_table pid_ns_ctl_table[] = { { .procname = "ns_last_pid", @@ -310,7 +319,11 @@ static struct ctl_table pid_ns_ctl_table[] = { .mode = 0666, /* permissions are checked in the handler */ .proc_handler = pid_ns_ctl_handler, .extra1 = SYSCTL_ZERO, +#ifndef CONFIG_PID_MAX_PER_NAMESPACE .extra2 = &pid_max, +#else + .extra2 = &init_pid_ns.pid_max, +#endif }, { } }; diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 354a2d294f52..e84df0818517 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1809,6 +1809,7 @@ static struct ctl_table kern_table[] = { .proc_handler = proc_dointvec, }, #endif +#ifndef CONFIG_PID_MAX_PER_NAMESPACE { .procname = "pid_max", .data = &pid_max, @@ -1818,6 +1819,7 @@ static struct ctl_table kern_table[] = { .extra1 = &pid_max_min, .extra2 = &pid_max_max, }, +#endif { .procname = "panic_on_oops", .data = &panic_on_oops, diff --git a/kernel/trace/pid_list.c b/kernel/trace/pid_list.c index 95106d02b32d..8056d7441d05 100644 --- a/kernel/trace/pid_list.c +++ b/kernel/trace/pid_list.c @@ -414,7 +414,11 @@ struct trace_pid_list *trace_pid_list_alloc(void) int i; /* According to linux/thread.h, pids can be no bigger that 30 bits */ +#ifndef CONFIG_PID_MAX_PER_NAMESPACE WARN_ON_ONCE(pid_max > (1 << 30)); +#else + WARN_ON_ONCE(task_active_pid_ns(current)->pid_max > (1 << 30)); +#endif pid_list = kzalloc(sizeof(*pid_list), GFP_KERNEL); if (!pid_list) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index b656cab67f67..5afa58302b06 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -5423,7 +5423,11 @@ int set_tracer_flag(struct trace_array *tr, unsigned int mask, int enabled) if (mask == TRACE_ITER_RECORD_TGID) { if (!tgid_map) { +#ifndef CONFIG_PID_MAX_PER_NAMESPACE tgid_map_max = pid_max; +#else + tgid_map_max = init_pid_ns.pid_max; +#endif map = kvcalloc(tgid_map_max + 1, sizeof(*tgid_map), GFP_KERNEL); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index d608f6128704..52123df262b2 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -695,7 +695,9 @@ extern unsigned long tracing_thresh; /* PID filtering */ +#ifndef CONFIG_PID_MAX_PER_NAMESPACE extern int pid_max; +#endif bool trace_find_filtered_pid(struct trace_pid_list *filtered_pids, pid_t search_pid); -- 2.25.1

2 1

[PATCH V4 OLK-6.6 0/8] Spark SQL scenario bpf readahead optimization synchronization to OLK-6.6
by ZhaoLong Wang 09 Jan '24

09 Jan '24

CONFIG: CONFIG_DEBUG_INFO_BTF=y CONFIG_BPF_SYSCALL=y CONFIG_NF_CONNTRACK=y CONFIG_NF_CONNTRACK_MARK=y V2: - Using helpers function, such as i_size_read(), round_up - Fix commit log Signature - Fix the user-mode test program V3 - Do not use the round_up helper function. Use the correct calculation - err = -errno before printf V4 - Drop the "Conflict" in the commit log. Hou Tao (5): vfs: add bare tracepoints for vfs read and release fs: add helper fs_file_read_do_trace() xfs: add trace for read and release of regular file ext4: add trace for the read and release of regular file selftests/bpf: add demo for file read pattern detection Yufen Yu (1): readahead: introduce FMODE_CTL_WILLNEED to read first 2MB of file ZhaoLong Wang (2): VFS: Rolling Back the fmode macro definition and structure members selftests/bpf: Update the demo file_read_pattern to run on libbpf 1.0+ fs/ext4/file.c | 4 + fs/read_write.c | 38 +++++ fs/xfs/xfs_file.c | 3 + include/linux/fs.h | 37 +++++ include/trace/events/fs.h | 33 +++++ mm/readahead.c | 40 ++++- tools/testing/selftests/bpf/Makefile | 1 + .../testing/selftests/bpf/file_read_pattern.c | 76 ++++++++++ .../bpf/progs/file_read_pattern_prog.c | 137 ++++++++++++++++++ 9 files changed, 368 insertions(+), 1 deletion(-) create mode 100644 include/trace/events/fs.h create mode 100644 tools/testing/selftests/bpf/file_read_pattern.c create mode 100644 tools/testing/selftests/bpf/progs/file_read_pattern_prog.c -- 2.39.2

2 9

[PATCH OLK-6.6 0/9] Add swap control for memcg
by Jinjiang Tu 09 Jan '24

09 Jan '24

This series patches add swap control for memory cgroup. Patch[1] add page type in memory.reclaim interface to support reclaim anon pages. Patch[3] add memory.force_swapin interface to support swap back pages proactively. Patch[4-6] add memory.swapfile interface to limit available swap device for memory cgroup. Patch[7] add memory.swap.max interface to limit usage of swap for memory cgroup. Jinjiang Tu (1): memcg: make sysctl registration more extensible Liu Shixin (7): memcg: add page type to memory.reclaim interface memcg: introduce memcg swap qos feature memcg: introduce per-memcg swapin interface memcg: add restrict to swap to cgroup1 mm/swapfile: introduce per-memcg swapfile control mm: swap_slots: add per-type slot cache config: enable memcg swap qos for x86_64 and arm64 by default zhaoxiaoqiang11 (1): memcg/swap: add ability to disable memcg swap .../admin-guide/cgroup-v1/memory.rst | 4 + Documentation/admin-guide/cgroup-v2.rst | 10 +- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + include/linux/memcontrol.h | 35 ++ include/linux/mm.h | 4 + include/linux/swap.h | 12 +- include/linux/swap_slots.h | 2 +- init/Kconfig | 9 + mm/madvise.c | 31 +- mm/memcontrol.c | 493 +++++++++++++++++- mm/swap_slots.c | 151 +++++- mm/swapfile.c | 97 +++- mm/vmscan.c | 9 + 14 files changed, 810 insertions(+), 49 deletions(-) -- 2.25.1

2 10

[PATCH OLK-6.6] arm64: config: disable network config
by Zhengchao Shao 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8URWS CVE: NA -------------------------------- Disable arm64 network config. Signed-off-by: Zhengchao Shao <shaozhengchao(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 32 +++++++++++++------------- 1 file changed, 16 insertions(+), 16 deletions(-) diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 86370e8686f5..a7d78c16002b 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -1301,7 +1301,7 @@ CONFIG_NETFILTER_NETLINK_QUEUE=m CONFIG_NETFILTER_NETLINK_LOG=m CONFIG_NETFILTER_NETLINK_OSF=m CONFIG_NF_CONNTRACK=m -CONFIG_NF_LOG_SYSLOG=m +# CONFIG_NF_LOG_SYSLOG is not set CONFIG_NETFILTER_CONNCOUNT=m CONFIG_NF_CONNTRACK_MARK=y CONFIG_NF_CONNTRACK_SECMARK=y @@ -1863,7 +1863,7 @@ CONFIG_RFKILL_LEDS=y CONFIG_RFKILL_INPUT=y CONFIG_RFKILL_GPIO=m CONFIG_NET_9P=m -CONFIG_NET_9P_FD=m +# CONFIG_NET_9P_FD is not set CONFIG_NET_9P_VIRTIO=m # CONFIG_NET_9P_RDMA is not set # CONFIG_NET_9P_DEBUG is not set @@ -2760,7 +2760,7 @@ CONFIG_NET_XGENE_V2=m CONFIG_NET_VENDOR_AQUANTIA=y # CONFIG_AQTION is not set CONFIG_NET_VENDOR_ARC=y -CONFIG_NET_VENDOR_ASIX=y +# CONFIG_NET_VENDOR_ASIX is not set # CONFIG_SPI_AX88796C is not set CONFIG_NET_VENDOR_ATHEROS=y # CONFIG_ATL2 is not set @@ -2805,16 +2805,16 @@ CONFIG_CHELSIO_IPSEC_INLINE=m CONFIG_CHELSIO_TLS_DEVICE=m # CONFIG_NET_VENDOR_CISCO is not set # CONFIG_NET_VENDOR_CORTINA is not set -CONFIG_NET_VENDOR_DAVICOM=y +# CONFIG_NET_VENDOR_DAVICOM is not set # CONFIG_DM9051 is not set CONFIG_DNET=m # CONFIG_NET_VENDOR_DEC is not set # CONFIG_NET_VENDOR_DLINK is not set # CONFIG_NET_VENDOR_EMULEX is not set -CONFIG_NET_VENDOR_ENGLEDER=y +# CONFIG_NET_VENDOR_ENGLEDER is not set # CONFIG_TSNEP is not set # CONFIG_NET_VENDOR_EZCHIP is not set -CONFIG_NET_VENDOR_FUNGIBLE=y +# CONFIG_NET_VENDOR_FUNGIBLE is not set # CONFIG_FUN_ETH is not set CONFIG_NET_VENDOR_GOOGLE=y # CONFIG_GVE is not set @@ -2852,13 +2852,13 @@ CONFIG_I40E=m CONFIG_IAVF=m CONFIG_I40EVF=m CONFIG_ICE=m -CONFIG_ICE_SWITCHDEV=y +# CONFIG_ICE_SWITCHDEV is not set CONFIG_FM10K=m # CONFIG_IGC is not set # CONFIG_JME is not set -CONFIG_NET_VENDOR_ADI=y +# CONFIG_NET_VENDOR_ADI is not set # CONFIG_ADIN1110 is not set -CONFIG_NET_VENDOR_LITEX=y +# CONFIG_NET_VENDOR_LITEX is not set # CONFIG_LITEX_LITEETH is not set # CONFIG_NET_VENDOR_MARVELL is not set CONFIG_NET_VENDOR_MELLANOX=y @@ -2895,7 +2895,7 @@ CONFIG_MLXFW=m # CONFIG_NET_VENDOR_MICROCHIP is not set CONFIG_NET_VENDOR_MICROSEMI=y # CONFIG_MSCC_OCELOT_SWITCH is not set -CONFIG_NET_VENDOR_MICROSOFT=y +# CONFIG_NET_VENDOR_MICROSOFT is not set CONFIG_NET_VENDOR_MYRI=y # CONFIG_MYRI10GE is not set # CONFIG_FEALNX is not set @@ -2906,7 +2906,7 @@ CONFIG_NET_VENDOR_NETRONOME=y CONFIG_NFP=m CONFIG_NFP_APP_FLOWER=y CONFIG_NFP_APP_ABM_NIC=y -CONFIG_NFP_NET_IPSEC=y +# CONFIG_NFP_NET_IPSEC is not set # CONFIG_NFP_DEBUG is not set # CONFIG_NET_VENDOR_NVIDIA is not set CONFIG_NET_VENDOR_OKI=y @@ -2966,10 +2966,10 @@ CONFIG_SMSC9420=m # CONFIG_NET_VENDOR_SYNOPSYS is not set # CONFIG_NET_VENDOR_TEHUTI is not set # CONFIG_NET_VENDOR_TI is not set -CONFIG_NET_VENDOR_VERTEXCOM=y +# CONFIG_NET_VENDOR_VERTEXCOM is not set # CONFIG_MSE102X is not set # CONFIG_NET_VENDOR_VIA is not set -CONFIG_NET_VENDOR_WANGXUN=y +# CONFIG_NET_VENDOR_WANGXUN is not set CONFIG_LIBWX=m CONFIG_NGBE=m CONFIG_TXGBE=m @@ -3235,7 +3235,7 @@ CONFIG_ATH10K_PCI=m CONFIG_WLAN_VENDOR_MICROCHIP=y # CONFIG_WILC1000_SDIO is not set # CONFIG_WILC1000_SPI is not set -CONFIG_WLAN_VENDOR_PURELIFI=y +# CONFIG_WLAN_VENDOR_PURELIFI is not set # CONFIG_PLFXLC is not set CONFIG_WLAN_VENDOR_RALINK=y CONFIG_RT2X00=m @@ -3262,7 +3262,7 @@ CONFIG_RT2X00_LIB_LEDS=y # CONFIG_RT2X00_DEBUG is not set # CONFIG_WLAN_VENDOR_REALTEK is not set # CONFIG_WLAN_VENDOR_RSI is not set -CONFIG_WLAN_VENDOR_SILABS=y +# CONFIG_WLAN_VENDOR_SILABS is not set # CONFIG_WFX is not set # CONFIG_WLAN_VENDOR_ST is not set # CONFIG_WLAN_VENDOR_TI is not set @@ -7028,7 +7028,7 @@ CONFIG_SUNRPC=m CONFIG_SUNRPC_GSS=m CONFIG_SUNRPC_BACKCHANNEL=y CONFIG_RPCSEC_GSS_KRB5=m -CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_AES_SHA1=y +# CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_AES_SHA1 is not set # CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_CAMELLIA is not set # CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_AES_SHA2 is not set CONFIG_SUNRPC_DEBUG=y -- 2.34.1

2 1

[PATCH OLK-6.6] x86: config: disable network config
by Zhengchao Shao 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8UM53 CVE: NA -------------------------------- Disable X86 network config. Signed-off-by: Zhengchao Shao <shaozhengchao(a)huawei.com> --- arch/x86/configs/openeuler_defconfig | 34 ++++++++++++++-------------- 1 file changed, 17 insertions(+), 17 deletions(-) diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index 1ef398404878..dd7a8a3bca0f 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -1860,7 +1860,7 @@ CONFIG_BT_CMTP=m CONFIG_BT_HIDP=m CONFIG_BT_HS=y CONFIG_BT_LE=y -CONFIG_BT_LE_L2CAP_ECRED=y +# CONFIG_BT_LE_L2CAP_ECRED is not set # CONFIG_BT_6LOWPAN is not set # CONFIG_BT_LEDS is not set # CONFIG_BT_MSFTEXT is not set @@ -1876,7 +1876,7 @@ CONFIG_BT_BCM=m CONFIG_BT_RTL=m CONFIG_BT_HCIBTUSB=m CONFIG_BT_HCIBTUSB_AUTOSUSPEND=y -CONFIG_BT_HCIBTUSB_POLL_SYNC=y +# CONFIG_BT_HCIBTUSB_POLL_SYNC is not set CONFIG_BT_HCIBTUSB_BCM=y # CONFIG_BT_HCIBTUSB_MTK is not set CONFIG_BT_HCIBTUSB_RTL=y @@ -2788,7 +2788,7 @@ CONFIG_AMD_XGBE_HAVE_ECC=y CONFIG_NET_VENDOR_AQUANTIA=y CONFIG_AQTION=m # CONFIG_NET_VENDOR_ARC is not set -CONFIG_NET_VENDOR_ASIX=y +# CONFIG_NET_VENDOR_ASIX is not set # CONFIG_SPI_AX88796C is not set CONFIG_NET_VENDOR_ATHEROS=y CONFIG_ATL2=m @@ -2836,7 +2836,7 @@ CONFIG_CHELSIO_IPSEC_INLINE=m CONFIG_NET_VENDOR_CISCO=y CONFIG_ENIC=m # CONFIG_NET_VENDOR_CORTINA is not set -CONFIG_NET_VENDOR_DAVICOM=y +# CONFIG_NET_VENDOR_DAVICOM is not set # CONFIG_DM9051 is not set CONFIG_DNET=m CONFIG_NET_VENDOR_DEC=y @@ -2851,10 +2851,10 @@ CONFIG_BE2NET_HWMON=y # CONFIG_BE2NET_BE3 is not set CONFIG_BE2NET_LANCER=y CONFIG_BE2NET_SKYHAWK=y -CONFIG_NET_VENDOR_ENGLEDER=y +# CONFIG_NET_VENDOR_ENGLEDER is not set # CONFIG_TSNEP is not set # CONFIG_NET_VENDOR_EZCHIP is not set -CONFIG_NET_VENDOR_FUNGIBLE=y +# CONFIG_NET_VENDOR_FUNGIBLE is not set # CONFIG_FUN_ETH is not set CONFIG_NET_VENDOR_GOOGLE=y # CONFIG_GVE is not set @@ -2882,14 +2882,14 @@ CONFIG_I40E_DCB=y CONFIG_IAVF=m CONFIG_I40EVF=m CONFIG_ICE=m -CONFIG_ICE_SWITCHDEV=y -CONFIG_ICE_HWTS=y +# CONFIG_ICE_SWITCHDEV is not set +#CONFIG_ICE_HWTS is not set CONFIG_FM10K=m # CONFIG_IGC is not set # CONFIG_JME is not set -CONFIG_NET_VENDOR_ADI=y +# CONFIG_NET_VENDOR_ADI is not set # CONFIG_ADIN1110 is not set -CONFIG_NET_VENDOR_LITEX=y +# CONFIG_NET_VENDOR_LITEX is not set # CONFIG_NET_VENDOR_MARVELL is not set CONFIG_NET_VENDOR_MELLANOX=y CONFIG_MLX4_EN=m @@ -2924,7 +2924,7 @@ CONFIG_MLXFW=m # CONFIG_NET_VENDOR_MICREL is not set # CONFIG_NET_VENDOR_MICROCHIP is not set # CONFIG_NET_VENDOR_MICROSEMI is not set -CONFIG_NET_VENDOR_MICROSOFT=y +# CONFIG_NET_VENDOR_MICROSOFT is not set # CONFIG_MICROSOFT_MANA is not set CONFIG_NET_VENDOR_MYRI=y CONFIG_MYRI10GE=m @@ -2938,7 +2938,7 @@ CONFIG_NET_VENDOR_NETRONOME=y CONFIG_NFP=m CONFIG_NFP_APP_FLOWER=y CONFIG_NFP_APP_ABM_NIC=y -CONFIG_NFP_NET_IPSEC=y +# CONFIG_NFP_NET_IPSEC is not set # CONFIG_NFP_DEBUG is not set # CONFIG_NET_VENDOR_NVIDIA is not set CONFIG_NET_VENDOR_OKI=y @@ -2995,10 +2995,10 @@ CONFIG_SFC_MCDI_LOGGING=y # CONFIG_NET_VENDOR_SYNOPSYS is not set # CONFIG_NET_VENDOR_TEHUTI is not set # CONFIG_NET_VENDOR_TI is not set -CONFIG_NET_VENDOR_VERTEXCOM=y +# CONFIG_NET_VENDOR_VERTEXCOM is not set # CONFIG_MSE102X is not set # CONFIG_NET_VENDOR_VIA is not set -CONFIG_NET_VENDOR_WANGXUN=y +# CONFIG_NET_VENDOR_WANGXUN is not set CONFIG_LIBWX=m CONFIG_NGBE=m CONFIG_TXGBE=m @@ -3325,7 +3325,7 @@ CONFIG_MT76x2U=m CONFIG_WLAN_VENDOR_MICROCHIP=y # CONFIG_WILC1000_SDIO is not set # CONFIG_WILC1000_SPI is not set -CONFIG_WLAN_VENDOR_PURELIFI=y +# CONFIG_WLAN_VENDOR_PURELIFI is not set # CONFIG_PLFXLC is not set CONFIG_WLAN_VENDOR_RALINK=y CONFIG_RT2X00=m @@ -3382,7 +3382,7 @@ CONFIG_RTL8XXXU=m # CONFIG_RTW88 is not set # CONFIG_RTW89 is not set # CONFIG_WLAN_VENDOR_RSI is not set -CONFIG_WLAN_VENDOR_SILABS=y +# CONFIG_WLAN_VENDOR_SILABS is not set # CONFIG_WFX is not set # CONFIG_WLAN_VENDOR_ST is not set # CONFIG_WLAN_VENDOR_TI is not set @@ -8225,7 +8225,7 @@ CONFIG_SUNRPC=m CONFIG_SUNRPC_GSS=m CONFIG_SUNRPC_BACKCHANNEL=y CONFIG_RPCSEC_GSS_KRB5=m -CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_AES_SHA1=y +# CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_AES_SHA1 is not set # CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_CAMELLIA is not set # CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_AES_SHA2 is not set CONFIG_SUNRPC_DEBUG=y -- 2.34.1

2 1

[PATCH OLK-6.6 v2] block: support to account io_ticks precisely
by Yu Kuai 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8RIME CVE: NA ----------------------------------------------- Currently, io_ticks is accounted based on sampling, specifically update_io_ticks() will always account io_ticks by 1 jiffies from bdev_start_io_acct()/blk_account_io_start(), and the result can be inaccurate, for example(HZ is 250): Test script: fio -filename=/dev/sda -bs=4k -rw=write -direct=1 -name=test -thinktime=4ms Test result: util is about 90%, while the disk is really idle. In order to account io_ticks precisely, update_io_ticks() must know if there are IO inflight already, and this requires overhead slightly, hence precise io accounting is disabled by default, and user can enable it through sysfs entry. Noted that for rq-based devcie, part_stat_local_inc/dec() and part_in_flight() is used to track inflight instead of iterating tags, which is not supposed to be used in fast path because 'tags->lock' is grabbed in blk_mq_find_and_get_req(). Signed-off-by: Yu Kuai <yukuai3(a)huawei.com> --- Changes in v2: - some minor code changes, no functional change involed; - use kstrtobool() to replace strtobool() Documentation/ABI/stable/sysfs-block | 8 ++++-- block/blk-core.c | 27 ++++++++++++++++-- block/blk-merge.c | 3 ++ block/blk-mq-debugfs.c | 2 ++ block/blk-mq.c | 11 +++++++- block/blk-sysfs.c | 42 ++++++++++++++++++++++++++-- block/blk.h | 1 + block/genhd.c | 2 +- include/linux/blk-mq.h | 1 + include/linux/blkdev.h | 3 ++ 10 files changed, 91 insertions(+), 9 deletions(-) diff --git a/Documentation/ABI/stable/sysfs-block b/Documentation/ABI/stable/sysfs-block index 1fe9a553c37b..79027bf2661a 100644 --- a/Documentation/ABI/stable/sysfs-block +++ b/Documentation/ABI/stable/sysfs-block @@ -358,8 +358,12 @@ What: /sys/block/<disk>/queue/iostats Date: January 2009 Contact: linux-block(a)vger.kernel.org Description: - [RW] This file is used to control (on/off) the iostats - accounting of the disk. + [RW] This file is used to control the iostats accounting of the + disk. If this value is 0, iostats accounting is disabled; If + this value is 1, iostats accounting is enabled, but io_ticks is + accounted by sampling and the result is not accurate; If this + value is 2, iostats accounting is enabled and io_ticks is + accounted precisely, but there will be slightly more overhead. What: /sys/block/<disk>/queue/logical_block_size diff --git a/block/blk-core.c b/block/blk-core.c index fdf25b8d6e78..13a27d1d8540 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -71,6 +71,21 @@ static struct kmem_cache *blk_requestq_cachep; */ static struct workqueue_struct *kblockd_workqueue; +static bool precise_iostat; + +static int __init precise_iostat_setup(char *str) +{ + bool precise; + + if (!kstrtobool(str, &precise)) { + precise_iostat = precise; + pr_info("precise iostat %d\n", precise_iostat); + } + + return 1; +} +__setup("precise_iostat=", precise_iostat_setup); + /** * blk_queue_flag_set - atomically set a queue flag * @flag: flag to be set @@ -441,6 +456,8 @@ struct request_queue *blk_alloc_queue(int node_id) blk_set_default_limits(&q->limits); q->nr_requests = BLKDEV_DEFAULT_RQ; + if (precise_iostat) + blk_queue_flag_set(QUEUE_FLAG_PRECISE_IO_STAT, q); return q; @@ -938,11 +955,15 @@ EXPORT_SYMBOL_GPL(iocb_bio_iopoll); void update_io_ticks(struct block_device *part, unsigned long now, bool end) { unsigned long stamp; + bool precise = blk_queue_precise_io_stat(part->bd_queue); again: stamp = READ_ONCE(part->bd_stamp); - if (unlikely(time_after(now, stamp))) { - if (likely(try_cmpxchg(&part->bd_stamp, &stamp, now))) - __part_stat_add(part, io_ticks, end ? now - stamp : 1); + if (unlikely(time_after(now, stamp)) && + likely(try_cmpxchg(&part->bd_stamp, &stamp, now))) { + if (end || (precise && part_in_flight(part))) + __part_stat_add(part, io_ticks, now - stamp); + else if (!precise) + __part_stat_add(part, io_ticks, 1); } if (part->bd_partno) { part = bdev_whole(part); diff --git a/block/blk-merge.c b/block/blk-merge.c index 65e75efa9bd3..5db8228c46fc 100644 --- a/block/blk-merge.c +++ b/block/blk-merge.c @@ -783,6 +783,9 @@ static void blk_account_io_merge_request(struct request *req) if (blk_do_io_stat(req)) { part_stat_lock(); part_stat_inc(req->part, merges[op_stat_group(req_op(req))]); + if (req->rq_flags & RQF_PRECISE_IO_STAT) + part_stat_local_dec(req->part, + in_flight[op_is_write(req_op(req))]); part_stat_unlock(); } } diff --git a/block/blk-mq-debugfs.c b/block/blk-mq-debugfs.c index aa7a1357c3e1..9ad108307344 100644 --- a/block/blk-mq-debugfs.c +++ b/block/blk-mq-debugfs.c @@ -86,6 +86,7 @@ static const char *const blk_queue_flag_name[] = { QUEUE_FLAG_NAME(FAIL_IO), QUEUE_FLAG_NAME(NONROT), QUEUE_FLAG_NAME(IO_STAT), + QUEUE_FLAG_NAME(PRECISE_IO_STAT), QUEUE_FLAG_NAME(NOXMERGES), QUEUE_FLAG_NAME(ADD_RANDOM), QUEUE_FLAG_NAME(SYNCHRONOUS), @@ -254,6 +255,7 @@ static const char *const rqf_name[] = { RQF_NAME(FAILED), RQF_NAME(QUIET), RQF_NAME(IO_STAT), + RQF_NAME(PRECISE_IO_STAT), RQF_NAME(PM), RQF_NAME(HASHED), RQF_NAME(STATS), diff --git a/block/blk-mq.c b/block/blk-mq.c index dd7c9e3eca1b..498a6306cbb5 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -360,8 +360,11 @@ static struct request *blk_mq_rq_ctx_init(struct blk_mq_alloc_data *data, if (data->flags & BLK_MQ_REQ_PM) data->rq_flags |= RQF_PM; - if (blk_queue_io_stat(q)) + if (blk_queue_io_stat(q)) { data->rq_flags |= RQF_IO_STAT; + if (blk_queue_precise_io_stat(q)) + data->rq_flags |= RQF_PRECISE_IO_STAT; + } rq->rq_flags = data->rq_flags; if (data->rq_flags & RQF_SCHED_TAGS) { @@ -994,6 +997,9 @@ static inline void blk_account_io_done(struct request *req, u64 now) update_io_ticks(req->part, jiffies, true); part_stat_inc(req->part, ios[sgrp]); part_stat_add(req->part, nsecs[sgrp], now - req->start_time_ns); + if (req->rq_flags & RQF_PRECISE_IO_STAT) + part_stat_local_dec(req->part, + in_flight[op_is_write(req_op(req))]); part_stat_unlock(); } } @@ -1016,6 +1022,9 @@ static inline void blk_account_io_start(struct request *req) part_stat_lock(); update_io_ticks(req->part, jiffies, false); + if (req->rq_flags & RQF_PRECISE_IO_STAT) + part_stat_local_inc(req->part, + in_flight[op_is_write(req_op(req))]); part_stat_unlock(); } } diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c index 63e481262336..e58b1574e023 100644 --- a/block/blk-sysfs.c +++ b/block/blk-sysfs.c @@ -303,7 +303,6 @@ queue_##name##_store(struct request_queue *q, const char *page, size_t count) \ QUEUE_SYSFS_BIT_FNS(nonrot, NONROT, 1); QUEUE_SYSFS_BIT_FNS(random, ADD_RANDOM, 0); -QUEUE_SYSFS_BIT_FNS(iostats, IO_STAT, 0); QUEUE_SYSFS_BIT_FNS(stable_writes, STABLE_WRITES, 0); #undef QUEUE_SYSFS_BIT_FNS @@ -473,6 +472,45 @@ static ssize_t queue_dax_show(struct request_queue *q, char *page) return queue_var_show(blk_queue_dax(q), page); } +static ssize_t queue_iostats_show(struct request_queue *q, char *page) +{ + int val = 0; + + if (blk_queue_io_stat(q)) + val = blk_queue_precise_io_stat(q) ? 2 : 1; + + return queue_var_show(val, page); +} + +static ssize_t +queue_iostats_store(struct request_queue *q, const char *page, size_t count) +{ + unsigned long nr; + int ret = queue_var_store(&nr, page, count); + + if (ret < 0) + return ret; + + switch (nr) { + case 0: + blk_queue_flag_clear(QUEUE_FLAG_IO_STAT, q); + blk_queue_flag_clear(QUEUE_FLAG_PRECISE_IO_STAT, q); + break; + case 1: + blk_queue_flag_set(QUEUE_FLAG_IO_STAT, q); + blk_queue_flag_clear(QUEUE_FLAG_PRECISE_IO_STAT, q); + break; + case 2: + blk_queue_flag_set(QUEUE_FLAG_IO_STAT, q); + blk_queue_flag_set(QUEUE_FLAG_PRECISE_IO_STAT, q); + break; + default: + return -EINVAL; + } + + return count; +} + #define QUEUE_RO_ENTRY(_prefix, _name) \ static struct queue_sysfs_entry _prefix##_entry = { \ .attr = { .name = _name, .mode = 0444 }, \ @@ -494,6 +532,7 @@ QUEUE_RO_ENTRY(queue_max_segments, "max_segments"); QUEUE_RO_ENTRY(queue_max_integrity_segments, "max_integrity_segments"); QUEUE_RO_ENTRY(queue_max_segment_size, "max_segment_size"); QUEUE_RW_ENTRY(elv_iosched, "scheduler"); +QUEUE_RW_ENTRY(queue_iostats, "iostats"); QUEUE_RO_ENTRY(queue_logical_block_size, "logical_block_size"); QUEUE_RO_ENTRY(queue_physical_block_size, "physical_block_size"); @@ -539,7 +578,6 @@ static struct queue_sysfs_entry queue_hw_sector_size_entry = { }; QUEUE_RW_ENTRY(queue_nonrot, "rotational"); -QUEUE_RW_ENTRY(queue_iostats, "iostats"); QUEUE_RW_ENTRY(queue_random, "add_random"); QUEUE_RW_ENTRY(queue_stable_writes, "stable_writes"); diff --git a/block/blk.h b/block/blk.h index 08a358bc0919..67915b04b3c1 100644 --- a/block/blk.h +++ b/block/blk.h @@ -344,6 +344,7 @@ static inline bool blk_do_io_stat(struct request *rq) } void update_io_ticks(struct block_device *part, unsigned long now, bool end); +unsigned int part_in_flight(struct block_device *part); static inline void req_set_nomerge(struct request_queue *q, struct request *req) { diff --git a/block/genhd.c b/block/genhd.c index 4a16a424f57d..b3ff653f3e50 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -118,7 +118,7 @@ static void part_stat_read_all(struct block_device *part, } } -static unsigned int part_in_flight(struct block_device *part) +unsigned int part_in_flight(struct block_device *part) { unsigned int inflight = 0; int cpu; diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h index 3d74f3e5b995..9dfee0644f9c 100644 --- a/include/linux/blk-mq.h +++ b/include/linux/blk-mq.h @@ -46,6 +46,7 @@ typedef __u32 __bitwise req_flags_t; #define RQF_QUIET ((__force req_flags_t)(1 << 11)) /* account into disk and partition IO statistics */ #define RQF_IO_STAT ((__force req_flags_t)(1 << 13)) +#define RQF_PRECISE_IO_STAT ((__force req_flags_t)(1 << 14)) /* runtime pm request */ #define RQF_PM ((__force req_flags_t)(1 << 15)) /* on IO scheduler merge hash */ diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 9f3bcbcb156d..bea0b5fdac74 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -536,6 +536,7 @@ struct request_queue { #define QUEUE_FLAG_NONROT 6 /* non-rotational device (SSD) */ #define QUEUE_FLAG_VIRT QUEUE_FLAG_NONROT /* paravirt device */ #define QUEUE_FLAG_IO_STAT 7 /* do disk/partitions IO accounting */ +#define QUEUE_FLAG_PRECISE_IO_STAT 8 /* do disk/partitions IO accounting precisely */ #define QUEUE_FLAG_NOXMERGES 9 /* No extended merges */ #define QUEUE_FLAG_ADD_RANDOM 10 /* Contributes to random pool */ #define QUEUE_FLAG_SYNCHRONOUS 11 /* always completes in submit context */ @@ -576,6 +577,8 @@ bool blk_queue_flag_test_and_set(unsigned int flag, struct request_queue *q); #define blk_queue_stable_writes(q) \ test_bit(QUEUE_FLAG_STABLE_WRITES, &(q)->queue_flags) #define blk_queue_io_stat(q) test_bit(QUEUE_FLAG_IO_STAT, &(q)->queue_flags) +#define blk_queue_precise_io_stat(q) \ + test_bit(QUEUE_FLAG_PRECISE_IO_STAT, &(q)->queue_flags) #define blk_queue_add_random(q) test_bit(QUEUE_FLAG_ADD_RANDOM, &(q)->queue_flags) #define blk_queue_zone_resetall(q) \ test_bit(QUEUE_FLAG_ZONE_RESETALL, &(q)->queue_flags) -- 2.39.2

2 1

[PATCH OLK-6.6 V1 00/12] infrastructure for scheduler bpf
by Guan Jing 09 Jan '24

09 Jan '24

infrastructure for scheduler bpf Guan Jing (12): sched: programmable: Introduce bpf sched sched: programmable: Add a tag for the task sched: programmable: Add a tag for the task group sched: programmable: Add user interface of task group tag sched: programmable: Add user interface of task tag sched: basic infrastructure for scheduler bpf sched: introduce bpf_sched_enable() libbpf: add support for scheduler bpf programs bpftool: recognize scheduler programs sched: Add helper functions to get cpu statistics sched: programmable: Add hook in select_task_rq_fair() sched: programmable: Add hook in can_migrate_task() arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/proc/base.c | 64 ++++++++++++++ include/linux/bpf_sched.h | 50 +++++++++++ include/linux/bpf_types.h | 4 + include/linux/sched.h | 38 ++++++++ include/linux/sched_hook_defs.h | 4 + include/uapi/linux/bpf.h | 9 ++ init/init_task.c | 3 + kernel/bpf/Kconfig | 12 +++ kernel/bpf/btf.c | 1 + kernel/bpf/syscall.c | 27 ++++++ kernel/bpf/trampoline.c | 1 + kernel/bpf/verifier.c | 11 ++- kernel/sched/bpf_sched.c | 102 +++++++++++++++++++++ kernel/sched/build_utility.c | 4 + kernel/sched/core.c | 118 +++++++++++++++++++++++++ kernel/sched/fair.c | 45 ++++++++++ kernel/sched/sched.h | 10 +++ scripts/bpf_doc.py | 4 + tools/include/uapi/linux/bpf.h | 9 ++ tools/lib/bpf/bpf.c | 1 + tools/lib/bpf/libbpf.c | 23 ++++- tools/lib/bpf/libbpf.h | 2 + tools/lib/bpf/libbpf.map | 1 + 25 files changed, 543 insertions(+), 2 deletions(-) create mode 100644 include/linux/bpf_sched.h create mode 100644 include/linux/sched_hook_defs.h create mode 100644 kernel/sched/bpf_sched.c -- 2.34.1

2 13

[PATCH V3 OLK-6.6] fs/dirty_pages: dump the number of dirty pages for each inode
by Zizhi Wo 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OHM4?from=project-issue CVE: NA ---------------------------------------- In order to analyse the IO performance when using buffer IO, it's useful to obtain the number of dirty pages for each inode in the filesystem. This feature depends on 'CONFIG_DIRTY_PAGES'. It creates 3 interfaces by using procfs. /proc/dirty/buffer_size for buffer allocation and release, /proc/dirty/page_threshold to filter result and /proc/dirty/dirty_list to get dirty pages. Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/Kconfig | 13 + fs/Makefile | 1 + fs/dirty_pages.c | 355 +++++++++++++++++++++++++ 5 files changed, 371 insertions(+) create mode 100644 fs/dirty_pages.c diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 66737e8c8673..677b4cc4668a 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -6924,6 +6924,7 @@ CONFIG_PROC_PAGE_MONITOR=y CONFIG_PROC_CHILDREN=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index fcbef6c06587..c16cdb7b2596 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -8120,6 +8120,7 @@ CONFIG_PROC_PID_ARCH_STATUS=y CONFIG_PROC_CPU_RESCTRL=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/fs/Kconfig b/fs/Kconfig index 7c9d45ba9b0f..6dc83223c61b 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -170,6 +170,19 @@ source "fs/proc/Kconfig" source "fs/kernfs/Kconfig" source "fs/sysfs/Kconfig" +config DIRTY_PAGES + bool "Dumps the number of dirty pages of each file" + depends on PROC_FS + default y + help + This config supports the rendering of dirty page data to the user, + which may be useful to analyze the IO performance when using buffer + IO. + + It create 3 interfaces by using procfs. /proc/dirty/buffer_size for + buffer allocation and release; /proc/dirty/page_threshold to filter + result; /proc/dirty/dirty_list to get dirty pages. + config TMPFS bool "Tmpfs virtual memory file system support (former shm fs)" depends on SHMEM diff --git a/fs/Makefile b/fs/Makefile index bb6164200a39..cc4735c23c32 100644 --- a/fs/Makefile +++ b/fs/Makefile @@ -44,6 +44,7 @@ obj-$(CONFIG_FS_POSIX_ACL) += posix_acl.o obj-$(CONFIG_NFS_COMMON) += nfs_common/ obj-$(CONFIG_COREDUMP) += coredump.o obj-$(CONFIG_SYSCTL) += drop_caches.o sysctls.o +obj-$(CONFIG_DIRTY_PAGES) += dirty_pages.o obj-$(CONFIG_FHANDLE) += fhandle.o obj-$(CONFIG_CGROUP_FILES) += filescontrol.o diff --git a/fs/dirty_pages.c b/fs/dirty_pages.c new file mode 100644 index 000000000000..d0aea022b68b --- /dev/null +++ b/fs/dirty_pages.c @@ -0,0 +1,355 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/proc_fs.h> +#include <linux/seq_file.h> +#include <linux/uaccess.h> +#include <linux/pagemap.h> +#include <linux/pagevec.h> +#include <linux/fs.h> +#include <linux/mm.h> +#include <linux/slab.h> +#include <linux/sched.h> +#include <linux/proc_fs.h> +#include <linux/kdev_t.h> +#include <linux/vmalloc.h> +#include <linux/module.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include "internal.h" + +#define DIRTY_ROOT "dirty" /* proc root directory name */ +#define DIRTY_BUFFER_SIZE "buffer_size" /* proc file to control buffer */ +#define DIRTY_PAGES "dirty_list" /* proc file to obtain diry pages */ +#define DIRTY_LIMIT "page_threshold" /* proc file to filter result */ +#define MAX_BUF_NUM 102400 /* max buffer num in number of pages */ +#define TERM_STR "terminated\n" /* tail string for overflowed buffer */ +#define UNKNOW_STR "unknown" /* for unknown file name or fs type */ + +static char *buf_dirty; /* buffer to store number of dirty pages */ +static unsigned long buf_size; /* size of buffer in bytes */ +static unsigned long buf_num; /* size of buffer in number of pages */ +static unsigned long buf_limit; /* filter threshold of dirty pages */ +unsigned long lock_word; /* for exclusive access to buffer */ + +static bool dirty_pages_lock(void) +{ + if (xchg(&lock_word, 1) == 1) + return false; + return true; +} + +static void dirty_pages_unlock(void) +{ + lock_word = 0; +} + +static unsigned long dump_dirtypages_inode(struct inode *inode) +{ + XA_STATE(xas, &inode->i_mapping->i_pages, 0); + unsigned long pages = 0; + void *page; + + if (!mapping_tagged(inode->i_mapping, PAGECACHE_TAG_DIRTY)) + return 0; + + xas_lock_irq(&xas); + xas_for_each_marked(&xas, page, (pgoff_t)-1, PAGECACHE_TAG_DIRTY) { + if (++pages % XA_CHECK_SCHED) + continue; + + xas_pause(&xas); + xas_unlock_irq(&xas); + cond_resched(); + xas_lock_irq(&xas); + } + xas_unlock_irq(&xas); + + return pages; +} + +static char *inode_filename(struct inode *inode) +{ + char *path; + static char buf[PATH_MAX]; + struct dentry *dentry; + + dentry = d_find_alias(inode); + if (!dentry) + return UNKNOW_STR; + + path = dentry_path_raw(dentry, buf, PATH_MAX); + dput(dentry); + + return IS_ERR_OR_NULL(path) ? UNKNOW_STR : path; +} + +static const char *sb_fstype(struct super_block *sb) +{ + if (sb->s_type && sb->s_type->name) + return sb->s_type->name; + else + return UNKNOW_STR; +} + +static inline bool is_sb_writable(struct super_block *sb) +{ + if (sb_rdonly(sb)) + return false; + + if (sb->s_writers.frozen == SB_FREEZE_COMPLETE) + return false; + + return true; +} + +/* + * dump_dirtypages_sb - dump the dirty pages of each inode in the sb + * @sb the super block + * @m the seq_file witch is initialized in proc_dpages_open + * + * For each inode in the sb, call dump_dirtypages_pages to get the number + * of dirty pages. And use seq_printf to store the result in the buffer + * if it's not less than the threshold. The inode in unusual state will + * be skipped. + */ +static void dump_dirtypages_sb(struct super_block *sb, struct seq_file *m) +{ + struct inode *inode, *toput_inode = NULL; + unsigned long limit; + + if (!is_sb_writable(sb) || !sb->s_bdev) + return; + + limit = READ_ONCE(buf_limit); + spin_lock(&sb->s_inode_list_lock); + list_for_each_entry(inode, &sb->s_inodes, i_sb_list) { + unsigned long dirty_pages; + + spin_lock(&inode->i_lock); + if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) { + spin_unlock(&inode->i_lock); + continue; + } + + __iget(inode); + spin_unlock(&inode->i_lock); + spin_unlock(&sb->s_inode_list_lock); + + dirty_pages = dump_dirtypages_inode(inode); + if (!dirty_pages || dirty_pages < limit) + goto skip; + + seq_printf(m, "FSType: %s, Dev ID: %u(%u:%u) ino %lu, dirty pages %lu, path %s\n", + sb_fstype(sb), sb->s_dev, MAJOR(sb->s_dev), + MINOR(sb->s_dev), inode->i_ino, dirty_pages, + inode_filename(inode)); + + if (seq_has_overflowed(m)) { + /* keep size > count to avoid overflow in seq_read_iter() */ + m->size += strlen(TERM_STR) + 1; + seq_puts(m, TERM_STR); + iput(inode); + iput(toput_inode); + return; + } +skip: + iput(toput_inode); + toput_inode = inode; + cond_resched(); + spin_lock(&sb->s_inode_list_lock); + } + spin_unlock(&sb->s_inode_list_lock); + iput(toput_inode); +} + +static int proc_dpages_show(struct seq_file *m, void *v) +{ + iterate_supers((void *)dump_dirtypages_sb, (void *)m); + return 0; +} + +static int proc_dpages_open(struct inode *inode, struct file *filp) +{ + int ret; + struct seq_file *m; + + if (!dirty_pages_lock()) + return -EBUSY; + + if (buf_dirty == NULL || buf_size == 0) { + pr_warn("please allocate buffer before getting dirty pages\n"); + dirty_pages_unlock(); + return -ENOMEM; + } + + ret = single_open(filp, proc_dpages_show, NULL); + if (ret) { + dirty_pages_unlock(); + return ret; + } + + m = (struct seq_file *)filp->private_data; + memset(buf_dirty, 0, buf_size); + m->size = buf_size - (strlen(TERM_STR) + 1); + m->buf = buf_dirty; + + return 0; +} + +static int proc_dpages_release(struct inode *inode, struct file *file) +{ + struct seq_file *m = file->private_data; + + /* we don't want to free the buf */ + m->buf = NULL; + dirty_pages_unlock(); + return single_release(inode, file); +} + +static void free_buf_dirty(void) +{ + if (!buf_dirty) + return; + + vfree(buf_dirty); + buf_dirty = NULL; + buf_size = 0; + WRITE_ONCE(buf_num, 0); +} + +static int proc_buffer_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", READ_ONCE(buf_num)); + return 0; +} + +static int proc_buffer_open(struct inode *inode, struct file *filp) +{ + int ret; + + if (((filp->f_flags & O_ACCMODE) != O_RDONLY) && !dirty_pages_lock()) + return -EBUSY; + + ret = single_open(filp, proc_buffer_show, NULL); + if (ret) + dirty_pages_unlock(); + + return ret; +} + +static ssize_t proc_buffer_write(struct file *filp, const char __user *buf, + size_t count, loff_t *offp) +{ + int ret; + unsigned long new_buf_num, new_buf_size; + char *new_buf_dirty; + + ret = kstrtoul_from_user(buf, count, 10, &new_buf_num); + if (ret != 0 || new_buf_num > MAX_BUF_NUM) + return -EINVAL; + + if (new_buf_num == buf_num) + return count; + + if (new_buf_num == 0) { + free_buf_dirty(); + return count; + } + + new_buf_size = PAGE_SIZE * new_buf_num; + new_buf_dirty = vzalloc(new_buf_size); + if (!new_buf_dirty) + return -ENOMEM; + + free_buf_dirty(); + buf_dirty = new_buf_dirty; + buf_size = new_buf_size; + WRITE_ONCE(buf_num, new_buf_num); + + return count; +} + +static int proc_buffer_release(struct inode *inode, struct file *filp) +{ + if ((filp->f_flags & O_ACCMODE) != O_RDONLY) + dirty_pages_unlock(); + return single_release(inode, filp); +} + +static int proc_limit_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", READ_ONCE(buf_limit)); + return 0; +} + +static int proc_limit_open(struct inode *inode, struct file *filp) +{ + return single_open(filp, proc_limit_show, NULL); +} + +static ssize_t proc_limit_write(struct file *filp, const char __user *buf, + size_t count, loff_t *offp) +{ + int ret; + unsigned long limit; + + ret = kstrtoul_from_user(buf, count, 10, &limit); + if (ret != 0) + return ret; + + WRITE_ONCE(buf_limit, limit); + + return count; +} + +static const struct proc_ops proc_dpages_operations = { + .proc_open = proc_dpages_open, + .proc_read = seq_read, + .proc_release = proc_dpages_release, +}; + +static const struct proc_ops proc_buffer_operations = { + .proc_open = proc_buffer_open, + .proc_read = seq_read, + .proc_write = proc_buffer_write, + .proc_lseek = seq_lseek, + .proc_release = proc_buffer_release, +}; + +static const struct proc_ops proc_limit_operations = { + .proc_open = proc_limit_open, + .proc_read = seq_read, + .proc_write = proc_limit_write, + .proc_lseek = seq_lseek, + .proc_release = single_release, +}; + +static int __init dpages_proc_init(void) +{ + struct proc_dir_entry *dirty_dir; + + dirty_dir = proc_mkdir(DIRTY_ROOT, NULL); + if (!dirty_dir) + return -ENOMEM; + + if (!proc_create(DIRTY_PAGES, 0440, dirty_dir, &proc_dpages_operations)) + goto fail_pages; + + if (!proc_create(DIRTY_BUFFER_SIZE, 0640, dirty_dir, + &proc_buffer_operations)) + goto fail_switch; + + if (!proc_create(DIRTY_LIMIT, 0640, dirty_dir, &proc_limit_operations)) + goto fail_limit; + + return 0; + +fail_limit: + remove_proc_entry(DIRTY_BUFFER_SIZE, dirty_dir); +fail_switch: + remove_proc_entry(DIRTY_PAGES, dirty_dir); +fail_pages: + remove_proc_entry(DIRTY_ROOT, NULL); + return -ENOMEM; +} + +subsys_initcall(dpages_proc_init); -- 2.39.2

2 1

[PATCH OLK-6.6] fs/dcache.c: avoid panic while lockref of dentry overflow
by Zizhi Wo 09 Jan '24

09 Jan '24

From: yangerkun <yangerkun(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8JDP6 --------------------------- We use lockref for dentry reference without notice that so many negative dentry under one dir can lead to overflow of lockref. This can lead to system crash if we do this under root dir. Since there is not a perfect solution, we just limit max number of dentry count up to INT_MAX / 2. Also, it will cost a lot of time from INT_MAX / 2 to INT_MAX, so we no need to do this under protection of dentry lock. Also, we limit the FILES_MAX to INT_MAX / 2, since a lot open for same file can lead to overflow too. Changelog: v1->v2: add a function to do check / add a Macro to mean INT_MAX / 2 Signed-off-by: yangerkun <yangerkun(a)huawei.com> Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- fs/dcache.c | 32 ++++++++++++++++++++++++++++---- fs/filescontrol.c | 2 +- include/linux/fs.h | 3 +++ 3 files changed, 32 insertions(+), 5 deletions(-) diff --git a/fs/dcache.c b/fs/dcache.c index 25ac74d30bff..7cdd8bbedd1c 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1834,6 +1834,18 @@ static struct dentry *__d_alloc(struct super_block *sb, const struct qstr *name) return dentry; } +static inline bool d_forbid_overflow(struct dentry *dentry) +{ + if (unlikely(d_count(dentry) >= D_COUNT_MAX)) { + shrink_dcache_parent(dentry); + + if (d_count(dentry) >= D_COUNT_MAX) + return false; + } + + return true; +} + /** * d_alloc - allocate a dcache entry * @parent: parent of entry to allocate @@ -1845,9 +1857,15 @@ static struct dentry *__d_alloc(struct super_block *sb, const struct qstr *name) */ struct dentry *d_alloc(struct dentry * parent, const struct qstr *name) { - struct dentry *dentry = __d_alloc(parent->d_sb, name); + struct dentry *dentry = NULL; + + if (unlikely(!d_forbid_overflow(parent))) + goto out; + + dentry = __d_alloc(parent->d_sb, name); if (!dentry) - return NULL; + goto out; + spin_lock(&parent->d_lock); /* * don't need child lock because it is not subject @@ -1857,7 +1875,7 @@ struct dentry *d_alloc(struct dentry * parent, const struct qstr *name) dentry->d_parent = parent; list_add(&dentry->d_child, &parent->d_subdirs); spin_unlock(&parent->d_lock); - +out: return dentry; } EXPORT_SYMBOL(d_alloc); @@ -1870,11 +1888,17 @@ EXPORT_SYMBOL(d_alloc_anon); struct dentry *d_alloc_cursor(struct dentry * parent) { - struct dentry *dentry = d_alloc_anon(parent->d_sb); + struct dentry *dentry = NULL; + + if (unlikely(!d_forbid_overflow(parent))) + goto out; + + dentry = d_alloc_anon(parent->d_sb); if (dentry) { dentry->d_flags |= DCACHE_DENTRY_CURSOR; dentry->d_parent = dget(parent); } +out: return dentry; } diff --git a/fs/filescontrol.c b/fs/filescontrol.c index 610f13a36962..0947566b9119 100644 --- a/fs/filescontrol.c +++ b/fs/filescontrol.c @@ -27,7 +27,7 @@ #include <linux/sched/signal.h> #include <linux/module.h> -#define FILES_MAX ULONG_MAX +#define FILES_MAX D_COUNT_MAX #define FILES_MAX_STR "max" static bool no_acct; diff --git a/include/linux/fs.h b/include/linux/fs.h index fb5accebdcdf..67c2ac09e700 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -47,6 +47,9 @@ #include <asm/byteorder.h> #include <uapi/linux/fs.h> +#define D_COUNT_MAX (INT_MAX / 2) + + struct backing_dev_info; struct bdi_writeback; struct bio; -- 2.39.2

2 1

[PATCH OLK-6.6 0/1] arm64: Add non nmi ipi backtrace support
by Liao Chen 09 Jan '24

09 Jan '24

Dependencies: CONFIG_DETECT_HUNG_TASK=y CONFIG_WATCHDOG=y Li Zhengyu (1): arm64: Add non nmi ipi backtrace support arch/arm64/Kconfig | 8 +++++ arch/arm64/configs/openeuler_defconfig | 1 + arch/arm64/kernel/ipi_nmi.c | 41 ++++++++++++++++++++++---- 3 files changed, 45 insertions(+), 5 deletions(-) -- 2.34.1

2 2

[PATCH OLK-6.6 0/1] arm64: Add non nmi ipi backtrace support
by Liao Chen 09 Jan '24

09 Jan '24

Dependencies: CONFIG_DETECT_HUNG_TASK=y CONFIG_WATCHDOG=y Li Zhengyu (1): arm64: Add non nmi ipi backtrace support arch/arm64/Kconfig | 8 +++++ arch/arm64/configs/openeuler_defconfig | 1 + arch/arm64/kernel/ipi_nmi.c | 41 ++++++++++++++++++++++---- 3 files changed, 45 insertions(+), 5 deletions(-) -- 2.34.1

1 1

[PATCH OLK-6.6 00/10] Backport etmem feature to OLK 6.6
by Yuchen Tang 09 Jan '24

09 Jan '24

v2 -> v3 - fix code style and commit messages v1 -> v2 - make independent patches to export symbols we need - make independent patches for etmem scan and swap config - fix indentation and code style problems Yuchen Tang (10): etmem: add ETMEM feature CONFIG to mm/Kconfig etmem: add ETMEM scan feature CONFIG to mm/Kconfig mm: Export symbol __pud_trans_huge_lock mm: Export symbol walk_page_range etmem: add etmem scan feature mm: Export symbol reclaim_pages etmem: add etmem swap feature etmem: add original kernel swap enabled options etmem: add swapcache reclaim to etmem etmem: enable etmem configurations arch/arm64/configs/openeuler_defconfig | 3 + arch/x86/configs/openeuler_defconfig | 3 + fs/proc/Makefile | 3 + fs/proc/base.c | 8 + fs/proc/etmem_proc.c | 216 ++++ fs/proc/etmem_scan.c | 1383 ++++++++++++++++++++++++ fs/proc/etmem_scan.h | 149 +++ fs/proc/etmem_swap.c | 283 +++++ fs/proc/internal.h | 4 + include/linux/etmem.h | 75 ++ include/linux/mm.h | 4 + include/linux/mm_types.h | 5 + include/linux/swap.h | 1 + include/uapi/asm-generic/mman-common.h | 4 + mm/Kconfig | 33 + mm/Makefile | 1 + mm/etmem.c | 384 +++++++ mm/huge_memory.c | 1 + mm/internal.h | 1 - mm/madvise.c | 16 + mm/pagewalk.c | 1 + mm/swap_state.c | 4 + mm/vmscan.c | 16 + virt/kvm/kvm_main.c | 8 + 24 files changed, 2605 insertions(+), 1 deletion(-) create mode 100644 fs/proc/etmem_proc.c create mode 100644 fs/proc/etmem_scan.c create mode 100644 fs/proc/etmem_scan.h create mode 100644 fs/proc/etmem_swap.c create mode 100644 include/linux/etmem.h create mode 100644 mm/etmem.c -- 2.33.0

1 10

[PATCH OLK-6.6] openeuler_defconfig: set CONFIG_NODES_SHIFT to 8 for both x86_64/ARM64
by Liu Shixin 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8UT5S CVE: NA -------------------------------- There is no extra bit in pageflags for new flag on x86_64. Since 256 nodes is enough for x86_64, reduce CONFIG_NODES_SHIFT to 8 to obtain two unused bits on x86_64. On arm64, 128 node is not enough, so raise CONFIG_NODES_SHIFT to 8. Signed-off-by: Liu Shixin <liushixin2(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 2 +- arch/x86/configs/openeuler_defconfig | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 0f0d9b3d6367..f6d76aa7fd97 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -450,7 +450,7 @@ CONFIG_NR_CPUS=4096 CONFIG_HOTPLUG_CPU=y # CONFIG_ARM64_BOOTPARAM_HOTPLUG_CPU0 is not set CONFIG_NUMA=y -CONFIG_NODES_SHIFT=7 +CONFIG_NODES_SHIFT=8 # CONFIG_HZ_100 is not set CONFIG_HZ_250=y # CONFIG_HZ_300 is not set diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index 504de984c7be..6b39cf20f8c2 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -449,7 +449,7 @@ CONFIG_NUMA_AWARE_SPINLOCKS=y CONFIG_AMD_NUMA=y CONFIG_X86_64_ACPI_NUMA=y CONFIG_NUMA_EMU=y -CONFIG_NODES_SHIFT=10 +CONFIG_NODES_SHIFT=8 CONFIG_ARCH_SPARSEMEM_ENABLE=y CONFIG_ARCH_SPARSEMEM_DEFAULT=y # CONFIG_ARCH_MEMORY_PROBE is not set -- 2.25.1

2 3

[PATCH OLK-6.6 V1 00/12] infrastructure for scheduler bpf
by Lu Jialin 09 Jan '24

09 Jan '24

From: Guan Jing <guanjing6(a)huawei.com> infrastructure for scheduler bpf Guan Jing (12): sched: programmable: Introduce bpf sched sched: programmable: Add a tag for the task sched: programmable: Add a tag for the task group sched: programmable: Add user interface of task group tag sched: programmable: Add user interface of task tag sched: basic infrastructure for scheduler bpf sched: introduce bpf_sched_enable() libbpf: add support for scheduler bpf programs bpftool: recognize scheduler programs sched: Add helper functions to get cpu statistics sched: programmable: Add hook in select_task_rq_fair() sched: programmable: Add hook in can_migrate_task() arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/proc/base.c | 64 ++++++++++++++ include/linux/bpf_sched.h | 50 +++++++++++ include/linux/bpf_types.h | 4 + include/linux/sched.h | 38 ++++++++ include/linux/sched_hook_defs.h | 4 + include/uapi/linux/bpf.h | 9 ++ init/init_task.c | 3 + kernel/bpf/Kconfig | 12 +++ kernel/bpf/btf.c | 1 + kernel/bpf/syscall.c | 27 ++++++ kernel/bpf/trampoline.c | 1 + kernel/bpf/verifier.c | 11 ++- kernel/sched/bpf_sched.c | 102 +++++++++++++++++++++ kernel/sched/build_utility.c | 4 + kernel/sched/core.c | 118 +++++++++++++++++++++++++ kernel/sched/fair.c | 45 ++++++++++ kernel/sched/sched.h | 10 +++ scripts/bpf_doc.py | 4 + tools/include/uapi/linux/bpf.h | 9 ++ tools/lib/bpf/bpf.c | 1 + tools/lib/bpf/libbpf.c | 23 ++++- tools/lib/bpf/libbpf.h | 2 + tools/lib/bpf/libbpf.map | 1 + 25 files changed, 543 insertions(+), 2 deletions(-) create mode 100644 include/linux/bpf_sched.h create mode 100644 include/linux/sched_hook_defs.h create mode 100644 kernel/sched/bpf_sched.c -- 2.34.1

1 12

[PATCH OLK-6.6 V1 00/12] infrastructure for scheduler bpf
by Guan Jing 09 Jan '24

09 Jan '24

infrastructure for scheduler bpf Guan Jing (12): sched: programmable: Introduce bpf sched sched: programmable: Add a tag for the task sched: programmable: Add a tag for the task group sched: programmable: Add user interface of task group tag sched: programmable: Add user interface of task tag sched: basic infrastructure for scheduler bpf sched: introduce bpf_sched_enable() libbpf: add support for scheduler bpf programs bpftool: recognize scheduler programs sched: Add helper functions to get cpu statistics sched: programmable: Add hook in select_task_rq_fair() sched: programmable: Add hook in can_migrate_task() arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/proc/base.c | 64 ++++++++++++++ include/linux/bpf_sched.h | 50 +++++++++++ include/linux/bpf_types.h | 4 + include/linux/sched.h | 38 ++++++++ include/linux/sched_hook_defs.h | 4 + include/uapi/linux/bpf.h | 9 ++ init/init_task.c | 3 + kernel/bpf/Kconfig | 12 +++ kernel/bpf/btf.c | 1 + kernel/bpf/syscall.c | 27 ++++++ kernel/bpf/trampoline.c | 1 + kernel/bpf/verifier.c | 11 ++- kernel/sched/bpf_sched.c | 102 +++++++++++++++++++++ kernel/sched/build_utility.c | 4 + kernel/sched/core.c | 118 +++++++++++++++++++++++++ kernel/sched/fair.c | 45 ++++++++++ kernel/sched/sched.h | 10 +++ scripts/bpf_doc.py | 4 + tools/include/uapi/linux/bpf.h | 9 ++ tools/lib/bpf/bpf.c | 1 + tools/lib/bpf/libbpf.c | 23 ++++- tools/lib/bpf/libbpf.h | 2 + tools/lib/bpf/libbpf.map | 1 + 25 files changed, 543 insertions(+), 2 deletions(-) create mode 100644 include/linux/bpf_sched.h create mode 100644 include/linux/sched_hook_defs.h create mode 100644 kernel/sched/bpf_sched.c -- 2.34.1

1 12

[PATCH OLK-6.6 0/1] iommu: set CONFIG_SMMU_BYPASS_DEV=y
by zhangnaichuan 09 Jan '24

09 Jan '24

set CONFIG_SMMU_BYPASS_DEV=y in openeuler_defconfig arch/arm64/configs/openeuler_defconfig | 1 + 1 file changed, 1 insertion(+) -- 2.33.0

2 2

[PATCH V3 OLK-6.6] fs/dirty_pages: dump the number of dirty pages for each inode
by Zizhi Wo 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OHM4?from=project-issue CVE: NA ---------------------------------------- In order to analyse the IO performance when using buffer IO, it's useful to obtain the number of dirty pages for each inode in the filesystem. This feature depends on 'CONFIG_DIRTY_PAGES'. It creates 3 interfaces by using procfs. /proc/dirty/buffer_size for buffer allocation and release, /proc/dirty/page_threshold to filter result and /proc/dirty/dirty_list to get dirty pages. Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/Kconfig | 13 + fs/Makefile | 1 + fs/dirty_pages.c | 355 +++++++++++++++++++++++++ 5 files changed, 371 insertions(+) create mode 100644 fs/dirty_pages.c diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 66737e8c8673..677b4cc4668a 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -6924,6 +6924,7 @@ CONFIG_PROC_PAGE_MONITOR=y CONFIG_PROC_CHILDREN=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index fcbef6c06587..c16cdb7b2596 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -8120,6 +8120,7 @@ CONFIG_PROC_PID_ARCH_STATUS=y CONFIG_PROC_CPU_RESCTRL=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/fs/Kconfig b/fs/Kconfig index 7c9d45ba9b0f..6dc83223c61b 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -170,6 +170,19 @@ source "fs/proc/Kconfig" source "fs/kernfs/Kconfig" source "fs/sysfs/Kconfig" +config DIRTY_PAGES + bool "Dumps the number of dirty pages of each file" + depends on PROC_FS + default y + help + This config supports the rendering of dirty page data to the user, + which may be useful to analyze the IO performance when using buffer + IO. + + It create 3 interfaces by using procfs. /proc/dirty/buffer_size for + buffer allocation and release; /proc/dirty/page_threshold to filter + result; /proc/dirty/dirty_list to get dirty pages. + config TMPFS bool "Tmpfs virtual memory file system support (former shm fs)" depends on SHMEM diff --git a/fs/Makefile b/fs/Makefile index bb6164200a39..cc4735c23c32 100644 --- a/fs/Makefile +++ b/fs/Makefile @@ -44,6 +44,7 @@ obj-$(CONFIG_FS_POSIX_ACL) += posix_acl.o obj-$(CONFIG_NFS_COMMON) += nfs_common/ obj-$(CONFIG_COREDUMP) += coredump.o obj-$(CONFIG_SYSCTL) += drop_caches.o sysctls.o +obj-$(CONFIG_DIRTY_PAGES) += dirty_pages.o obj-$(CONFIG_FHANDLE) += fhandle.o obj-$(CONFIG_CGROUP_FILES) += filescontrol.o diff --git a/fs/dirty_pages.c b/fs/dirty_pages.c new file mode 100644 index 000000000000..d0aea022b68b --- /dev/null +++ b/fs/dirty_pages.c @@ -0,0 +1,355 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/proc_fs.h> +#include <linux/seq_file.h> +#include <linux/uaccess.h> +#include <linux/pagemap.h> +#include <linux/pagevec.h> +#include <linux/fs.h> +#include <linux/mm.h> +#include <linux/slab.h> +#include <linux/sched.h> +#include <linux/proc_fs.h> +#include <linux/kdev_t.h> +#include <linux/vmalloc.h> +#include <linux/module.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include "internal.h" + +#define DIRTY_ROOT "dirty" /* proc root directory name */ +#define DIRTY_BUFFER_SIZE "buffer_size" /* proc file to control buffer */ +#define DIRTY_PAGES "dirty_list" /* proc file to obtain diry pages */ +#define DIRTY_LIMIT "page_threshold" /* proc file to filter result */ +#define MAX_BUF_NUM 102400 /* max buffer num in number of pages */ +#define TERM_STR "terminated\n" /* tail string for overflowed buffer */ +#define UNKNOW_STR "unknown" /* for unknown file name or fs type */ + +static char *buf_dirty; /* buffer to store number of dirty pages */ +static unsigned long buf_size; /* size of buffer in bytes */ +static unsigned long buf_num; /* size of buffer in number of pages */ +static unsigned long buf_limit; /* filter threshold of dirty pages */ +unsigned long lock_word; /* for exclusive access to buffer */ + +static bool dirty_pages_lock(void) +{ + if (xchg(&lock_word, 1) == 1) + return false; + return true; +} + +static void dirty_pages_unlock(void) +{ + lock_word = 0; +} + +static unsigned long dump_dirtypages_inode(struct inode *inode) +{ + XA_STATE(xas, &inode->i_mapping->i_pages, 0); + unsigned long pages = 0; + void *page; + + if (!mapping_tagged(inode->i_mapping, PAGECACHE_TAG_DIRTY)) + return 0; + + xas_lock_irq(&xas); + xas_for_each_marked(&xas, page, (pgoff_t)-1, PAGECACHE_TAG_DIRTY) { + if (++pages % XA_CHECK_SCHED) + continue; + + xas_pause(&xas); + xas_unlock_irq(&xas); + cond_resched(); + xas_lock_irq(&xas); + } + xas_unlock_irq(&xas); + + return pages; +} + +static char *inode_filename(struct inode *inode) +{ + char *path; + static char buf[PATH_MAX]; + struct dentry *dentry; + + dentry = d_find_alias(inode); + if (!dentry) + return UNKNOW_STR; + + path = dentry_path_raw(dentry, buf, PATH_MAX); + dput(dentry); + + return IS_ERR_OR_NULL(path) ? UNKNOW_STR : path; +} + +static const char *sb_fstype(struct super_block *sb) +{ + if (sb->s_type && sb->s_type->name) + return sb->s_type->name; + else + return UNKNOW_STR; +} + +static inline bool is_sb_writable(struct super_block *sb) +{ + if (sb_rdonly(sb)) + return false; + + if (sb->s_writers.frozen == SB_FREEZE_COMPLETE) + return false; + + return true; +} + +/* + * dump_dirtypages_sb - dump the dirty pages of each inode in the sb + * @sb the super block + * @m the seq_file witch is initialized in proc_dpages_open + * + * For each inode in the sb, call dump_dirtypages_pages to get the number + * of dirty pages. And use seq_printf to store the result in the buffer + * if it's not less than the threshold. The inode in unusual state will + * be skipped. + */ +static void dump_dirtypages_sb(struct super_block *sb, struct seq_file *m) +{ + struct inode *inode, *toput_inode = NULL; + unsigned long limit; + + if (!is_sb_writable(sb) || !sb->s_bdev) + return; + + limit = READ_ONCE(buf_limit); + spin_lock(&sb->s_inode_list_lock); + list_for_each_entry(inode, &sb->s_inodes, i_sb_list) { + unsigned long dirty_pages; + + spin_lock(&inode->i_lock); + if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) { + spin_unlock(&inode->i_lock); + continue; + } + + __iget(inode); + spin_unlock(&inode->i_lock); + spin_unlock(&sb->s_inode_list_lock); + + dirty_pages = dump_dirtypages_inode(inode); + if (!dirty_pages || dirty_pages < limit) + goto skip; + + seq_printf(m, "FSType: %s, Dev ID: %u(%u:%u) ino %lu, dirty pages %lu, path %s\n", + sb_fstype(sb), sb->s_dev, MAJOR(sb->s_dev), + MINOR(sb->s_dev), inode->i_ino, dirty_pages, + inode_filename(inode)); + + if (seq_has_overflowed(m)) { + /* keep size > count to avoid overflow in seq_read_iter() */ + m->size += strlen(TERM_STR) + 1; + seq_puts(m, TERM_STR); + iput(inode); + iput(toput_inode); + return; + } +skip: + iput(toput_inode); + toput_inode = inode; + cond_resched(); + spin_lock(&sb->s_inode_list_lock); + } + spin_unlock(&sb->s_inode_list_lock); + iput(toput_inode); +} + +static int proc_dpages_show(struct seq_file *m, void *v) +{ + iterate_supers((void *)dump_dirtypages_sb, (void *)m); + return 0; +} + +static int proc_dpages_open(struct inode *inode, struct file *filp) +{ + int ret; + struct seq_file *m; + + if (!dirty_pages_lock()) + return -EBUSY; + + if (buf_dirty == NULL || buf_size == 0) { + pr_warn("please allocate buffer before getting dirty pages\n"); + dirty_pages_unlock(); + return -ENOMEM; + } + + ret = single_open(filp, proc_dpages_show, NULL); + if (ret) { + dirty_pages_unlock(); + return ret; + } + + m = (struct seq_file *)filp->private_data; + memset(buf_dirty, 0, buf_size); + m->size = buf_size - (strlen(TERM_STR) + 1); + m->buf = buf_dirty; + + return 0; +} + +static int proc_dpages_release(struct inode *inode, struct file *file) +{ + struct seq_file *m = file->private_data; + + /* we don't want to free the buf */ + m->buf = NULL; + dirty_pages_unlock(); + return single_release(inode, file); +} + +static void free_buf_dirty(void) +{ + if (!buf_dirty) + return; + + vfree(buf_dirty); + buf_dirty = NULL; + buf_size = 0; + WRITE_ONCE(buf_num, 0); +} + +static int proc_buffer_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", READ_ONCE(buf_num)); + return 0; +} + +static int proc_buffer_open(struct inode *inode, struct file *filp) +{ + int ret; + + if (((filp->f_flags & O_ACCMODE) != O_RDONLY) && !dirty_pages_lock()) + return -EBUSY; + + ret = single_open(filp, proc_buffer_show, NULL); + if (ret) + dirty_pages_unlock(); + + return ret; +} + +static ssize_t proc_buffer_write(struct file *filp, const char __user *buf, + size_t count, loff_t *offp) +{ + int ret; + unsigned long new_buf_num, new_buf_size; + char *new_buf_dirty; + + ret = kstrtoul_from_user(buf, count, 10, &new_buf_num); + if (ret != 0 || new_buf_num > MAX_BUF_NUM) + return -EINVAL; + + if (new_buf_num == buf_num) + return count; + + if (new_buf_num == 0) { + free_buf_dirty(); + return count; + } + + new_buf_size = PAGE_SIZE * new_buf_num; + new_buf_dirty = vzalloc(new_buf_size); + if (!new_buf_dirty) + return -ENOMEM; + + free_buf_dirty(); + buf_dirty = new_buf_dirty; + buf_size = new_buf_size; + WRITE_ONCE(buf_num, new_buf_num); + + return count; +} + +static int proc_buffer_release(struct inode *inode, struct file *filp) +{ + if ((filp->f_flags & O_ACCMODE) != O_RDONLY) + dirty_pages_unlock(); + return single_release(inode, filp); +} + +static int proc_limit_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", READ_ONCE(buf_limit)); + return 0; +} + +static int proc_limit_open(struct inode *inode, struct file *filp) +{ + return single_open(filp, proc_limit_show, NULL); +} + +static ssize_t proc_limit_write(struct file *filp, const char __user *buf, + size_t count, loff_t *offp) +{ + int ret; + unsigned long limit; + + ret = kstrtoul_from_user(buf, count, 10, &limit); + if (ret != 0) + return ret; + + WRITE_ONCE(buf_limit, limit); + + return count; +} + +static const struct proc_ops proc_dpages_operations = { + .proc_open = proc_dpages_open, + .proc_read = seq_read, + .proc_release = proc_dpages_release, +}; + +static const struct proc_ops proc_buffer_operations = { + .proc_open = proc_buffer_open, + .proc_read = seq_read, + .proc_write = proc_buffer_write, + .proc_lseek = seq_lseek, + .proc_release = proc_buffer_release, +}; + +static const struct proc_ops proc_limit_operations = { + .proc_open = proc_limit_open, + .proc_read = seq_read, + .proc_write = proc_limit_write, + .proc_lseek = seq_lseek, + .proc_release = single_release, +}; + +static int __init dpages_proc_init(void) +{ + struct proc_dir_entry *dirty_dir; + + dirty_dir = proc_mkdir(DIRTY_ROOT, NULL); + if (!dirty_dir) + return -ENOMEM; + + if (!proc_create(DIRTY_PAGES, 0440, dirty_dir, &proc_dpages_operations)) + goto fail_pages; + + if (!proc_create(DIRTY_BUFFER_SIZE, 0640, dirty_dir, + &proc_buffer_operations)) + goto fail_switch; + + if (!proc_create(DIRTY_LIMIT, 0640, dirty_dir, &proc_limit_operations)) + goto fail_limit; + + return 0; + +fail_limit: + remove_proc_entry(DIRTY_BUFFER_SIZE, dirty_dir); +fail_switch: + remove_proc_entry(DIRTY_PAGES, dirty_dir); +fail_pages: + remove_proc_entry(DIRTY_ROOT, NULL); + return -ENOMEM; +} + +subsys_initcall(dpages_proc_init); -- 2.39.2

1 0

[PATCH OLK-6.6] fs/dcache.c: avoid panic while lockref of dentry overflow
by Zizhi Wo 09 Jan '24

09 Jan '24

From: yangerkun <yangerkun(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8JDP6 --------------------------- We use lockref for dentry reference without notice that so many negative dentry under one dir can lead to overflow of lockref. This can lead to system crash if we do this under root dir. Since there is not a perfect solution, we just limit max number of dentry count up to INT_MAX / 2. Also, it will cost a lot of time from INT_MAX / 2 to INT_MAX, so we no need to do this under protection of dentry lock. Also, we limit the FILES_MAX to INT_MAX / 2, since a lot open for same file can lead to overflow too. Changelog: v1->v2: add a function to do check / add a Macro to mean INT_MAX / 2 Signed-off-by: yangerkun <yangerkun(a)huawei.com> Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- fs/dcache.c | 32 ++++++++++++++++++++++++++++---- fs/filescontrol.c | 2 +- include/linux/fs.h | 3 +++ 3 files changed, 32 insertions(+), 5 deletions(-) diff --git a/fs/dcache.c b/fs/dcache.c index 25ac74d30bff..7cdd8bbedd1c 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1834,6 +1834,18 @@ static struct dentry *__d_alloc(struct super_block *sb, const struct qstr *name) return dentry; } +static inline bool d_forbid_overflow(struct dentry *dentry) +{ + if (unlikely(d_count(dentry) >= D_COUNT_MAX)) { + shrink_dcache_parent(dentry); + + if (d_count(dentry) >= D_COUNT_MAX) + return false; + } + + return true; +} + /** * d_alloc - allocate a dcache entry * @parent: parent of entry to allocate @@ -1845,9 +1857,15 @@ static struct dentry *__d_alloc(struct super_block *sb, const struct qstr *name) */ struct dentry *d_alloc(struct dentry * parent, const struct qstr *name) { - struct dentry *dentry = __d_alloc(parent->d_sb, name); + struct dentry *dentry = NULL; + + if (unlikely(!d_forbid_overflow(parent))) + goto out; + + dentry = __d_alloc(parent->d_sb, name); if (!dentry) - return NULL; + goto out; + spin_lock(&parent->d_lock); /* * don't need child lock because it is not subject @@ -1857,7 +1875,7 @@ struct dentry *d_alloc(struct dentry * parent, const struct qstr *name) dentry->d_parent = parent; list_add(&dentry->d_child, &parent->d_subdirs); spin_unlock(&parent->d_lock); - +out: return dentry; } EXPORT_SYMBOL(d_alloc); @@ -1870,11 +1888,17 @@ EXPORT_SYMBOL(d_alloc_anon); struct dentry *d_alloc_cursor(struct dentry * parent) { - struct dentry *dentry = d_alloc_anon(parent->d_sb); + struct dentry *dentry = NULL; + + if (unlikely(!d_forbid_overflow(parent))) + goto out; + + dentry = d_alloc_anon(parent->d_sb); if (dentry) { dentry->d_flags |= DCACHE_DENTRY_CURSOR; dentry->d_parent = dget(parent); } +out: return dentry; } diff --git a/fs/filescontrol.c b/fs/filescontrol.c index 610f13a36962..0947566b9119 100644 --- a/fs/filescontrol.c +++ b/fs/filescontrol.c @@ -27,7 +27,7 @@ #include <linux/sched/signal.h> #include <linux/module.h> -#define FILES_MAX ULONG_MAX +#define FILES_MAX D_COUNT_MAX #define FILES_MAX_STR "max" static bool no_acct; diff --git a/include/linux/fs.h b/include/linux/fs.h index fb5accebdcdf..67c2ac09e700 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -47,6 +47,9 @@ #include <asm/byteorder.h> #include <uapi/linux/fs.h> +#define D_COUNT_MAX (INT_MAX / 2) + + struct backing_dev_info; struct bdi_writeback; struct bio; -- 2.39.2

1 0

[PATCH OLK-6.6] hfs: fix null-ptr-deref in hfs_find_init()
by Long Li 09 Jan '24

09 Jan '24

From: Yang Yingliang <yangyingliang(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8URA4 CVE: CVE-2018-12928 -------------------------------- It will cause a null-ptr-deref in hfs_find_init() in fuzz test. [ 107.092729] hfs: continuing without an alternate MDB [ 107.097632] general protection fault, probably for non-canonical address 0xdffffc0000000008: 0000 [#1] SMP KASAN PTI [ 107.104679] KASAN: null-ptr-deref in range [0x0000000000000040-0x0000000000000047] [ 107.109100] CPU: 0 PID: 379 Comm: hfs_inject Not tainted 5.7.0-rc7-00001-g24627f5f2973 #897 [ 107.114142] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014 [ 107.121095] RIP: 0010:hfs_find_init+0x72/0x170 [ 107.123609] Code: c1 ea 03 80 3c 02 00 0f 85 e6 00 00 00 4c 8d 65 40 48 c7 43 18 00 00 00 00 48 b8 00 00 00 00 00 fc ff df 4c 89 e2 48 c1 ea 03 <0f> b6 04 02 84 c0 74 08 3c 03 0f 8e a5 00 00 00 8b 45 40 be c0 0c [ 107.134660] RSP: 0018:ffff88810291f3f8 EFLAGS: 00010202 [ 107.137897] RAX: dffffc0000000000 RBX: ffff88810291f468 RCX: 1ffff110175cdf05 [ 107.141874] RDX: 0000000000000008 RSI: ffff88810291f468 RDI: ffff88810291f480 [ 107.145844] RBP: 0000000000000000 R08: 0000000000000000 R09: ffffed1020381013 [ 107.149431] R10: ffff88810291f500 R11: ffffed1020381012 R12: 0000000000000040 [ 107.152315] R13: 0000000000000000 R14: ffff888101c0814a R15: ffff88810291f468 [ 107.155464] FS: 00000000009ea880(0000) GS:ffff88810c600000(0000) knlGS:0000000000000000 [ 107.159795] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 107.162987] CR2: 00005605a19dd284 CR3: 0000000103a0c006 CR4: 0000000000020ef0 [ 107.166665] Call Trace: [ 107.167969] ? find_held_lock+0x33/0x1c0 [ 107.169972] hfs_ext_read_extent+0x16b/0xb00 [ 107.172092] ? create_page_buffers+0x14e/0x1b0 [ 107.174303] ? hfs_free_extents+0x280/0x280 [ 107.176437] ? lock_downgrade+0x730/0x730 [ 107.178272] hfs_get_block+0x496/0x8a0 [ 107.179972] block_read_full_page+0x241/0x8d0 [ 107.181971] ? hfs_extend_file+0xae0/0xae0 [ 107.183814] ? end_buffer_async_read_io+0x10/0x10 [ 107.185954] ? add_to_page_cache_lru+0x13f/0x1f0 [ 107.188006] ? add_to_page_cache_locked+0x10/0x10 [ 107.190175] do_read_cache_page+0xc6a/0x1180 [ 107.192096] ? generic_file_read_iter+0x4c0/0x4c0 [ 107.194234] ? hfs_btree_open+0x408/0x1000 [ 107.196068] ? lock_downgrade+0x730/0x730 [ 107.197926] ? wake_bit_function+0x180/0x180 [ 107.199845] ? lockdep_init_map_waits+0x267/0x7c0 [ 107.201895] hfs_btree_open+0x455/0x1000 [ 107.203479] hfs_mdb_get+0x122c/0x1ae8 [ 107.205065] ? hfs_mdb_put+0x350/0x350 [ 107.206590] ? queue_work_node+0x260/0x260 [ 107.208309] ? rcu_read_lock_sched_held+0xa1/0xd0 [ 107.210227] ? lockdep_init_map_waits+0x267/0x7c0 [ 107.212144] ? lockdep_init_map_waits+0x267/0x7c0 [ 107.213979] hfs_fill_super+0x9ba/0x1280 [ 107.215444] ? bdev_name.isra.9+0xf1/0x2b0 [ 107.217028] ? hfs_remount+0x190/0x190 [ 107.218428] ? pointer+0x5da/0x710 [ 107.219745] ? file_dentry_name+0xf0/0xf0 [ 107.221262] ? mount_bdev+0xd1/0x330 [ 107.222592] ? vsnprintf+0x7bd/0x1250 [ 107.224007] ? pointer+0x710/0x710 [ 107.225332] ? down_write+0xe5/0x160 [ 107.226698] ? hfs_remount+0x190/0x190 [ 107.228120] ? snprintf+0x91/0xc0 [ 107.229388] ? vsprintf+0x10/0x10 [ 107.230628] ? sget+0x3af/0x4a0 [ 107.231848] ? hfs_remount+0x190/0x190 [ 107.233300] mount_bdev+0x26e/0x330 [ 107.234611] ? hfs_statfs+0x540/0x540 [ 107.236015] legacy_get_tree+0x101/0x1f0 [ 107.237431] ? security_capable+0x58/0x90 [ 107.238832] vfs_get_tree+0x89/0x2d0 [ 107.240082] ? ns_capable_common+0x5c/0xd0 [ 107.241521] do_mount+0xd8a/0x1720 [ 107.242727] ? lock_downgrade+0x730/0x730 [ 107.244116] ? copy_mount_string+0x20/0x20 [ 107.245557] ? _copy_from_user+0xbe/0x100 [ 107.246967] ? memdup_user+0x47/0x70 [ 107.248212] __x64_sys_mount+0x162/0x1b0 [ 107.249537] do_syscall_64+0xa5/0x4f0 [ 107.250742] entry_SYSCALL_64_after_hwframe+0x49/0xb3 [ 107.252369] RIP: 0033:0x44e8ea [ 107.253360] Code: 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 49 89 ca b8 a5 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48 [ 107.259240] RSP: 002b:00007ffd910e4c28 EFLAGS: 00000207 ORIG_RAX: 00000000000000a5 [ 107.261668] RAX: ffffffffffffffda RBX: 0000000000400400 RCX: 000000000044e8ea [ 107.263920] RDX: 000000000049321e RSI: 0000000000493222 RDI: 00007ffd910e4d00 [ 107.266177] RBP: 00007ffd910e5d10 R08: 0000000000000000 R09: 000000000000000a [ 107.268451] R10: 0000000000000001 R11: 0000000000000207 R12: 0000000000401c40 [ 107.270721] R13: 0000000000000000 R14: 00000000006ba018 R15: 0000000000000000 [ 107.273025] Modules linked in: [ 107.274029] Dumping ftrace buffer: [ 107.275121] (ftrace buffer empty) [ 107.276370] ---[ end trace c5e0b9d684f3570e ]--- We need check tree in hfs_find_init(). https://lore.kernel.org/linux-fsdevel/20180419024358.GA5215@bombadil.infrad… https://marc.info/?l=linux-fsdevel&m=152406881024567&w=2 References: CVE-2018-12928 Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> Signed-off-by: Long Li <leo.lilong(a)huawei.com> --- fs/hfs/bfind.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/hfs/bfind.c b/fs/hfs/bfind.c index ef9498a6e88a..ae44c8fa86ec 100644 --- a/fs/hfs/bfind.c +++ b/fs/hfs/bfind.c @@ -16,6 +16,8 @@ int hfs_find_init(struct hfs_btree *tree, struct hfs_find_data *fd) { void *ptr; + if (!tree) + return -EINVAL; fd->tree = tree; fd->bnode = NULL; ptr = kmalloc(tree->max_key_len * 2 + 4, GFP_KERNEL); -- 2.31.1

1 0

[PATCH OLK-6.6] sched/smt: fix unbalance sched_smt_present dec/inc
by Xia Fukun 09 Jan '24

09 Jan '24

From: Yang Yingliang <yangyingliang(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8UR4U CVE: NA -------------------------------- If cpuset_cpu_inactive() fails in sched_cpu_deactivate(), the cpu offline failed, but sched_smt_present is decreased before calling sched_cpu_deactivate, it leads unbalance dec/inc, so fix it by increasing sched_smt_present in the error path. Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> Signed-off-by: Yu Liao <liaoyu15(a)huawei.com> Signed-off-by: Xia Fukun <xiafukun(a)huawei.com> --- kernel/sched/core.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 7a0997e7e136..adecea9c607d 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -9792,6 +9792,10 @@ int sched_cpu_deactivate(unsigned int cpu) ret = cpuset_cpu_inactive(cpu); if (ret) { balance_push_set(cpu, false); +#ifdef CONFIG_SCHED_SMT + if (cpumask_weight(cpu_smt_mask(cpu)) == 2) + static_branch_inc_cpuslocked(&sched_smt_present); +#endif set_cpu_active(cpu, true); sched_update_numa(cpu, true); return ret; -- 2.34.1

1 0

[PATCH OLK-6.6 0/1] iommu: set CONFIG_SMMU_BYPASS_DEV=y
by zhangnaichuan 09 Jan '24

09 Jan '24

set CONFIG_SMMU_BYPASS_DEV=y in openeuler_defconfig arch/arm64/configs/openeuler_defconfig | 1 + 1 file changed, 1 insertion(+) -- 2.33.0

1 0

[PATCH OLK-6.6 v5 0/5] block: Add config to show info about opening a mounted device for write
by Li Lingfeng 09 Jan '24

09 Jan '24

v1->v2: 1) Replace BLK_DEV_DUMPINFO with BLK_DEV_WRITE_MOUNTED_QUIET, which is logically opposite; 2) Make BLK_DEV_WRITE_MOUNTED_QUIET dependent on BLK_DEV_WRITE_MOUNTED, and decide whether to set bd_writers by BLK_DEV_WRITE_MOUNTED_QUIET; 3) Move bdev_dump_info() into bdev_may_open(); 4) Add the cmdline interface to control whether to show info; 5) Pass the state of "writes blocked" to part0 when mounting a partition. v2->v3: Add restriction for modifying bd_writers. v3->v4: 1) Count bd_writers regardless of whether bdev_allow_write_mounted is set; 2) Expand the meaning of bdev_allow_write_mounted; 3) Add config option to detect writing to part0 while partitions mounted; 4) Add bd_mounters to record the number of times that part0 or partition is mounted; 5) Show info about opening a lower device for write while upper-layers mounted. v4->v5: Move introduction of bd_mounters to patch1. Li Lingfeng (5): block: Count writers and mounters regardless of whether bdev_allow_write_mounted is set block: Expand the meaning of bdev_allow_write_mounted block: Add config option to detect writing to part0 while partitions mounted block: Add config option to show info about opening a mounted device for write block: Show info about opening a lower device for write while upper-layers mounted block/Kconfig | 19 ++++++++ block/bdev.c | 100 +++++++++++++++++++++++++++++++++----- include/linux/blk_types.h | 1 + 3 files changed, 107 insertions(+), 13 deletions(-) -- 2.31.1

2 8

[PATCH OLK-6.6 v2] block: support to account io_ticks precisely
by Yu Kuai 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8RIME CVE: NA ----------------------------------------------- Currently, io_ticks is accounted based on sampling, specifically update_io_ticks() will always account io_ticks by 1 jiffies from bdev_start_io_acct()/blk_account_io_start(), and the result can be inaccurate, for example(HZ is 250): Test script: fio -filename=/dev/sda -bs=4k -rw=write -direct=1 -name=test -thinktime=4ms Test result: util is about 90%, while the disk is really idle. In order to account io_ticks precisely, update_io_ticks() must know if there are IO inflight already, and this requires overhead slightly, hence precise io accounting is disabled by default, and user can enable it through sysfs entry. Noted that for rq-based devcie, part_stat_local_inc/dec() and part_in_flight() is used to track inflight instead of iterating tags, which is not supposed to be used in fast path because 'tags->lock' is grabbed in blk_mq_find_and_get_req(). Signed-off-by: Yu Kuai <yukuai3(a)huawei.com> --- Changes in v2: - some minor code changes, no functional change involed; Documentation/ABI/stable/sysfs-block | 8 ++++-- block/blk-core.c | 27 ++++++++++++++++-- block/blk-merge.c | 3 ++ block/blk-mq-debugfs.c | 2 ++ block/blk-mq.c | 11 +++++++- block/blk-sysfs.c | 42 ++++++++++++++++++++++++++-- block/blk.h | 1 + block/genhd.c | 2 +- include/linux/blk-mq.h | 1 + include/linux/blkdev.h | 3 ++ 10 files changed, 91 insertions(+), 9 deletions(-) diff --git a/Documentation/ABI/stable/sysfs-block b/Documentation/ABI/stable/sysfs-block index 1fe9a553c37b..79027bf2661a 100644 --- a/Documentation/ABI/stable/sysfs-block +++ b/Documentation/ABI/stable/sysfs-block @@ -358,8 +358,12 @@ What: /sys/block/<disk>/queue/iostats Date: January 2009 Contact: linux-block(a)vger.kernel.org Description: - [RW] This file is used to control (on/off) the iostats - accounting of the disk. + [RW] This file is used to control the iostats accounting of the + disk. If this value is 0, iostats accounting is disabled; If + this value is 1, iostats accounting is enabled, but io_ticks is + accounted by sampling and the result is not accurate; If this + value is 2, iostats accounting is enabled and io_ticks is + accounted precisely, but there will be slightly more overhead. What: /sys/block/<disk>/queue/logical_block_size diff --git a/block/blk-core.c b/block/blk-core.c index fdf25b8d6e78..13a27d1d8540 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -71,6 +71,21 @@ static struct kmem_cache *blk_requestq_cachep; */ static struct workqueue_struct *kblockd_workqueue; +static bool precise_iostat; + +static int __init precise_iostat_setup(char *str) +{ + bool precise; + + if (!strtobool(str, &precise)) { + precise_iostat = precise; + pr_info("precise iostat %d\n", precise_iostat); + } + + return 1; +} +__setup("precise_iostat=", precise_iostat_setup); + /** * blk_queue_flag_set - atomically set a queue flag * @flag: flag to be set @@ -441,6 +456,8 @@ struct request_queue *blk_alloc_queue(int node_id) blk_set_default_limits(&q->limits); q->nr_requests = BLKDEV_DEFAULT_RQ; + if (precise_iostat) + blk_queue_flag_set(QUEUE_FLAG_PRECISE_IO_STAT, q); return q; @@ -938,11 +955,15 @@ EXPORT_SYMBOL_GPL(iocb_bio_iopoll); void update_io_ticks(struct block_device *part, unsigned long now, bool end) { unsigned long stamp; + bool precise = blk_queue_precise_io_stat(part->bd_queue); again: stamp = READ_ONCE(part->bd_stamp); - if (unlikely(time_after(now, stamp))) { - if (likely(try_cmpxchg(&part->bd_stamp, &stamp, now))) - __part_stat_add(part, io_ticks, end ? now - stamp : 1); + if (unlikely(time_after(now, stamp)) && + likely(try_cmpxchg(&part->bd_stamp, &stamp, now))) { + if (end || (precise && part_in_flight(part))) + __part_stat_add(part, io_ticks, now - stamp); + else if (!precise) + __part_stat_add(part, io_ticks, 1); } if (part->bd_partno) { part = bdev_whole(part); diff --git a/block/blk-merge.c b/block/blk-merge.c index 65e75efa9bd3..5db8228c46fc 100644 --- a/block/blk-merge.c +++ b/block/blk-merge.c @@ -783,6 +783,9 @@ static void blk_account_io_merge_request(struct request *req) if (blk_do_io_stat(req)) { part_stat_lock(); part_stat_inc(req->part, merges[op_stat_group(req_op(req))]); + if (req->rq_flags & RQF_PRECISE_IO_STAT) + part_stat_local_dec(req->part, + in_flight[op_is_write(req_op(req))]); part_stat_unlock(); } } diff --git a/block/blk-mq-debugfs.c b/block/blk-mq-debugfs.c index aa7a1357c3e1..9ad108307344 100644 --- a/block/blk-mq-debugfs.c +++ b/block/blk-mq-debugfs.c @@ -86,6 +86,7 @@ static const char *const blk_queue_flag_name[] = { QUEUE_FLAG_NAME(FAIL_IO), QUEUE_FLAG_NAME(NONROT), QUEUE_FLAG_NAME(IO_STAT), + QUEUE_FLAG_NAME(PRECISE_IO_STAT), QUEUE_FLAG_NAME(NOXMERGES), QUEUE_FLAG_NAME(ADD_RANDOM), QUEUE_FLAG_NAME(SYNCHRONOUS), @@ -254,6 +255,7 @@ static const char *const rqf_name[] = { RQF_NAME(FAILED), RQF_NAME(QUIET), RQF_NAME(IO_STAT), + RQF_NAME(PRECISE_IO_STAT), RQF_NAME(PM), RQF_NAME(HASHED), RQF_NAME(STATS), diff --git a/block/blk-mq.c b/block/blk-mq.c index dd7c9e3eca1b..498a6306cbb5 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -360,8 +360,11 @@ static struct request *blk_mq_rq_ctx_init(struct blk_mq_alloc_data *data, if (data->flags & BLK_MQ_REQ_PM) data->rq_flags |= RQF_PM; - if (blk_queue_io_stat(q)) + if (blk_queue_io_stat(q)) { data->rq_flags |= RQF_IO_STAT; + if (blk_queue_precise_io_stat(q)) + data->rq_flags |= RQF_PRECISE_IO_STAT; + } rq->rq_flags = data->rq_flags; if (data->rq_flags & RQF_SCHED_TAGS) { @@ -994,6 +997,9 @@ static inline void blk_account_io_done(struct request *req, u64 now) update_io_ticks(req->part, jiffies, true); part_stat_inc(req->part, ios[sgrp]); part_stat_add(req->part, nsecs[sgrp], now - req->start_time_ns); + if (req->rq_flags & RQF_PRECISE_IO_STAT) + part_stat_local_dec(req->part, + in_flight[op_is_write(req_op(req))]); part_stat_unlock(); } } @@ -1016,6 +1022,9 @@ static inline void blk_account_io_start(struct request *req) part_stat_lock(); update_io_ticks(req->part, jiffies, false); + if (req->rq_flags & RQF_PRECISE_IO_STAT) + part_stat_local_inc(req->part, + in_flight[op_is_write(req_op(req))]); part_stat_unlock(); } } diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c index 63e481262336..e58b1574e023 100644 --- a/block/blk-sysfs.c +++ b/block/blk-sysfs.c @@ -303,7 +303,6 @@ queue_##name##_store(struct request_queue *q, const char *page, size_t count) \ QUEUE_SYSFS_BIT_FNS(nonrot, NONROT, 1); QUEUE_SYSFS_BIT_FNS(random, ADD_RANDOM, 0); -QUEUE_SYSFS_BIT_FNS(iostats, IO_STAT, 0); QUEUE_SYSFS_BIT_FNS(stable_writes, STABLE_WRITES, 0); #undef QUEUE_SYSFS_BIT_FNS @@ -473,6 +472,45 @@ static ssize_t queue_dax_show(struct request_queue *q, char *page) return queue_var_show(blk_queue_dax(q), page); } +static ssize_t queue_iostats_show(struct request_queue *q, char *page) +{ + int val = 0; + + if (blk_queue_io_stat(q)) + val = blk_queue_precise_io_stat(q) ? 2 : 1; + + return queue_var_show(val, page); +} + +static ssize_t +queue_iostats_store(struct request_queue *q, const char *page, size_t count) +{ + unsigned long nr; + int ret = queue_var_store(&nr, page, count); + + if (ret < 0) + return ret; + + switch (nr) { + case 0: + blk_queue_flag_clear(QUEUE_FLAG_IO_STAT, q); + blk_queue_flag_clear(QUEUE_FLAG_PRECISE_IO_STAT, q); + break; + case 1: + blk_queue_flag_set(QUEUE_FLAG_IO_STAT, q); + blk_queue_flag_clear(QUEUE_FLAG_PRECISE_IO_STAT, q); + break; + case 2: + blk_queue_flag_set(QUEUE_FLAG_IO_STAT, q); + blk_queue_flag_set(QUEUE_FLAG_PRECISE_IO_STAT, q); + break; + default: + return -EINVAL; + } + + return count; +} + #define QUEUE_RO_ENTRY(_prefix, _name) \ static struct queue_sysfs_entry _prefix##_entry = { \ .attr = { .name = _name, .mode = 0444 }, \ @@ -494,6 +532,7 @@ QUEUE_RO_ENTRY(queue_max_segments, "max_segments"); QUEUE_RO_ENTRY(queue_max_integrity_segments, "max_integrity_segments"); QUEUE_RO_ENTRY(queue_max_segment_size, "max_segment_size"); QUEUE_RW_ENTRY(elv_iosched, "scheduler"); +QUEUE_RW_ENTRY(queue_iostats, "iostats"); QUEUE_RO_ENTRY(queue_logical_block_size, "logical_block_size"); QUEUE_RO_ENTRY(queue_physical_block_size, "physical_block_size"); @@ -539,7 +578,6 @@ static struct queue_sysfs_entry queue_hw_sector_size_entry = { }; QUEUE_RW_ENTRY(queue_nonrot, "rotational"); -QUEUE_RW_ENTRY(queue_iostats, "iostats"); QUEUE_RW_ENTRY(queue_random, "add_random"); QUEUE_RW_ENTRY(queue_stable_writes, "stable_writes"); diff --git a/block/blk.h b/block/blk.h index 08a358bc0919..67915b04b3c1 100644 --- a/block/blk.h +++ b/block/blk.h @@ -344,6 +344,7 @@ static inline bool blk_do_io_stat(struct request *rq) } void update_io_ticks(struct block_device *part, unsigned long now, bool end); +unsigned int part_in_flight(struct block_device *part); static inline void req_set_nomerge(struct request_queue *q, struct request *req) { diff --git a/block/genhd.c b/block/genhd.c index 4a16a424f57d..b3ff653f3e50 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -118,7 +118,7 @@ static void part_stat_read_all(struct block_device *part, } } -static unsigned int part_in_flight(struct block_device *part) +unsigned int part_in_flight(struct block_device *part) { unsigned int inflight = 0; int cpu; diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h index 3d74f3e5b995..9dfee0644f9c 100644 --- a/include/linux/blk-mq.h +++ b/include/linux/blk-mq.h @@ -46,6 +46,7 @@ typedef __u32 __bitwise req_flags_t; #define RQF_QUIET ((__force req_flags_t)(1 << 11)) /* account into disk and partition IO statistics */ #define RQF_IO_STAT ((__force req_flags_t)(1 << 13)) +#define RQF_PRECISE_IO_STAT ((__force req_flags_t)(1 << 14)) /* runtime pm request */ #define RQF_PM ((__force req_flags_t)(1 << 15)) /* on IO scheduler merge hash */ diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 9f3bcbcb156d..bea0b5fdac74 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -536,6 +536,7 @@ struct request_queue { #define QUEUE_FLAG_NONROT 6 /* non-rotational device (SSD) */ #define QUEUE_FLAG_VIRT QUEUE_FLAG_NONROT /* paravirt device */ #define QUEUE_FLAG_IO_STAT 7 /* do disk/partitions IO accounting */ +#define QUEUE_FLAG_PRECISE_IO_STAT 8 /* do disk/partitions IO accounting precisely */ #define QUEUE_FLAG_NOXMERGES 9 /* No extended merges */ #define QUEUE_FLAG_ADD_RANDOM 10 /* Contributes to random pool */ #define QUEUE_FLAG_SYNCHRONOUS 11 /* always completes in submit context */ @@ -576,6 +577,8 @@ bool blk_queue_flag_test_and_set(unsigned int flag, struct request_queue *q); #define blk_queue_stable_writes(q) \ test_bit(QUEUE_FLAG_STABLE_WRITES, &(q)->queue_flags) #define blk_queue_io_stat(q) test_bit(QUEUE_FLAG_IO_STAT, &(q)->queue_flags) +#define blk_queue_precise_io_stat(q) \ + test_bit(QUEUE_FLAG_PRECISE_IO_STAT, &(q)->queue_flags) #define blk_queue_add_random(q) test_bit(QUEUE_FLAG_ADD_RANDOM, &(q)->queue_flags) #define blk_queue_zone_resetall(q) \ test_bit(QUEUE_FLAG_ZONE_RESETALL, &(q)->queue_flags) -- 2.39.2

1 0

[PATCH OLK-6.6] fs/dcache.c: avoid panic while lockref of dentry overflow
by Zizhi Wo 09 Jan '24

09 Jan '24

From: yangerkun <yangerkun(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8JDP6 --------------------------- We use lockref for dentry reference without notice that so many negative dentry under one dir can lead to overflow of lockref. This can lead to system crash if we do this under root dir. Since there is not a perfect solution, we just limit max number of dentry count up to INT_MAX / 2. Also, it will cost a lot of time from INT_MAX / 2 to INT_MAX, so we no need to do this under protection of dentry lock. Also, we limit the FILES_MAX to INT_MAX / 2, since a lot open for same file can lead to overflow too. Changelog: v1->v2: add a function to do check / add a Macro to mean INT_MAX / 2 Signed-off-by: yangerkun <yangerkun(a)huawei.com> Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- fs/dcache.c | 32 ++++++++++++++++++++++++++++---- fs/filescontrol.c | 2 +- include/linux/fs.h | 3 +++ 3 files changed, 32 insertions(+), 5 deletions(-) diff --git a/fs/dcache.c b/fs/dcache.c index 25ac74d30bff..7cdd8bbedd1c 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1834,6 +1834,18 @@ static struct dentry *__d_alloc(struct super_block *sb, const struct qstr *name) return dentry; } +static inline bool d_forbid_overflow(struct dentry *dentry) +{ + if (unlikely(d_count(dentry) >= D_COUNT_MAX)) { + shrink_dcache_parent(dentry); + + if (d_count(dentry) >= D_COUNT_MAX) + return false; + } + + return true; +} + /** * d_alloc - allocate a dcache entry * @parent: parent of entry to allocate @@ -1845,9 +1857,15 @@ static struct dentry *__d_alloc(struct super_block *sb, const struct qstr *name) */ struct dentry *d_alloc(struct dentry * parent, const struct qstr *name) { - struct dentry *dentry = __d_alloc(parent->d_sb, name); + struct dentry *dentry = NULL; + + if (unlikely(!d_forbid_overflow(parent))) + goto out; + + dentry = __d_alloc(parent->d_sb, name); if (!dentry) - return NULL; + goto out; + spin_lock(&parent->d_lock); /* * don't need child lock because it is not subject @@ -1857,7 +1875,7 @@ struct dentry *d_alloc(struct dentry * parent, const struct qstr *name) dentry->d_parent = parent; list_add(&dentry->d_child, &parent->d_subdirs); spin_unlock(&parent->d_lock); - +out: return dentry; } EXPORT_SYMBOL(d_alloc); @@ -1870,11 +1888,17 @@ EXPORT_SYMBOL(d_alloc_anon); struct dentry *d_alloc_cursor(struct dentry * parent) { - struct dentry *dentry = d_alloc_anon(parent->d_sb); + struct dentry *dentry = NULL; + + if (unlikely(!d_forbid_overflow(parent))) + goto out; + + dentry = d_alloc_anon(parent->d_sb); if (dentry) { dentry->d_flags |= DCACHE_DENTRY_CURSOR; dentry->d_parent = dget(parent); } +out: return dentry; } diff --git a/fs/filescontrol.c b/fs/filescontrol.c index 610f13a36962..0947566b9119 100644 --- a/fs/filescontrol.c +++ b/fs/filescontrol.c @@ -27,7 +27,7 @@ #include <linux/sched/signal.h> #include <linux/module.h> -#define FILES_MAX ULONG_MAX +#define FILES_MAX D_COUNT_MAX #define FILES_MAX_STR "max" static bool no_acct; diff --git a/include/linux/fs.h b/include/linux/fs.h index fb5accebdcdf..67c2ac09e700 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -47,6 +47,9 @@ #include <asm/byteorder.h> #include <uapi/linux/fs.h> +#define D_COUNT_MAX (INT_MAX / 2) + + struct backing_dev_info; struct bdi_writeback; struct bio; -- 2.39.2

1 0

[PATCH RESEND OLK-6.6] kabi: reserve space for bpf related structures
by Pu Lehui 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8UOP9 -------------------------------- reserve space for bpf related structures. Signed-off-by: Pu Lehui <pulehui(a)huawei.com> --- include/linux/bpf.h | 115 +++++++++++++++++++++++++++++++++++ include/linux/bpf_verifier.h | 36 +++++++++++ include/uapi/linux/bpf.h | 16 +++++ 3 files changed, 167 insertions(+) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index 392f581af2ce..f37e681cd4f5 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -29,6 +29,7 @@ #include <linux/rcupdate_trace.h> #include <linux/static_call.h> #include <linux/memcontrol.h> +#include <linux/kabi.h> struct bpf_verifier_env; struct bpf_verifier_log; @@ -163,6 +164,11 @@ struct bpf_map_ops { u64 (*map_mem_usage)(const struct bpf_map *map); + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + /* BTF id of struct allocated by map_alloc */ int *map_btf_id; @@ -288,6 +294,11 @@ struct bpf_map { bool bypass_spec_v1; bool frozen; /* write-once; write-protected by freeze_mutex */ s64 __percpu *elem_count; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; static inline const char *btf_field_type_name(enum btf_field_type type) @@ -664,6 +675,13 @@ enum bpf_type_flag { /* DYNPTR points to xdp_buff */ DYNPTR_TYPE_XDP = BIT(16 + BPF_BASE_TYPE_BITS), +#ifdef CONFIG_KABI_RESERVE + BPF_TYPE_FLAG_RESERVE_1 = BIT(17 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_2 = BIT(18 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_3 = BIT(19 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_4 = BIT(20 + BPF_BASE_TYPE_BITS), +#endif + __BPF_TYPE_FLAG_MAX, __BPF_TYPE_LAST_FLAG = __BPF_TYPE_FLAG_MAX - 1, }; @@ -730,6 +748,13 @@ enum bpf_arg_type { /* Pointer to valid memory of size known at compile time. */ ARG_PTR_TO_FIXED_SIZE_MEM = MEM_FIXED_SIZE | ARG_PTR_TO_MEM, +#ifdef CONFIG_KABI_RESERVE + BPF_ARG_TYPE_RESERVE_1, + BPF_ARG_TYPE_RESERVE_2, + BPF_ARG_TYPE_RESERVE_3, + BPF_ARG_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -760,6 +785,13 @@ enum bpf_return_type { RET_PTR_TO_BTF_ID_OR_NULL = PTR_MAYBE_NULL | RET_PTR_TO_BTF_ID, RET_PTR_TO_BTF_ID_TRUSTED = PTR_TRUSTED | RET_PTR_TO_BTF_ID, +#ifdef CONFIG_KABI_RESERVE + BPF_RET_TYPE_RESERVE_1, + BPF_RET_TYPE_RESERVE_2, + BPF_RET_TYPE_RESERVE_3, + BPF_RET_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -807,6 +839,9 @@ struct bpf_func_proto { }; int *ret_btf_id; /* return value btf_id */ bool (*allowed)(const struct bpf_prog *prog); + + KABI_RESERVE(1) + KABI_RESERVE(2) }; /* bpf_context is intentionally undefined structure. Pointer to bpf_context is @@ -875,6 +910,13 @@ enum bpf_reg_type { PTR_TO_TCP_SOCK_OR_NULL = PTR_MAYBE_NULL | PTR_TO_TCP_SOCK, PTR_TO_BTF_ID_OR_NULL = PTR_MAYBE_NULL | PTR_TO_BTF_ID, +#ifdef CONFIG_KABI_RESERVE + BPF_REG_TYPE_RESERVE_1, + BPF_REG_TYPE_RESERVE_2, + BPF_REG_TYPE_RESERVE_3, + BPF_REG_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -895,6 +937,9 @@ struct bpf_insn_access_aux { }; }; struct bpf_verifier_log *log; /* for verbose logs */ + + KABI_RESERVE(1) + KABI_RESERVE(2) }; static inline void @@ -942,6 +987,11 @@ struct bpf_verifier_ops { int (*btf_struct_access)(struct bpf_verifier_log *log, const struct bpf_reg_state *reg, int off, int size); + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_prog_offload_ops { @@ -1121,6 +1171,9 @@ struct bpf_tramp_image { struct rcu_head rcu; struct work_struct work; }; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_trampoline { @@ -1149,6 +1202,11 @@ struct bpf_trampoline { /* Executable image of trampoline */ struct bpf_tramp_image *cur_image; struct module *mod; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_attach_target_info { @@ -1180,6 +1238,11 @@ struct bpf_dispatcher { struct static_call_key *sc_key; void *sc_tramp; #endif + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; static __always_inline __nocfi unsigned int bpf_dispatcher_nop_func( @@ -1339,6 +1402,9 @@ static inline bool bpf_prog_has_trampoline(const struct bpf_prog *prog) struct bpf_func_info_aux { u16 linkage; bool unreliable; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; enum bpf_jit_poke_reason { @@ -1361,6 +1427,9 @@ struct bpf_jit_poke_descriptor { u8 adj_off; u16 reason; u32 insn_idx; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; /* reg_type info for ctx arguments */ @@ -1368,6 +1437,9 @@ struct bpf_ctx_arg_aux { u32 offset; enum bpf_reg_type reg_type; u32 btf_id; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct btf_mod_pair { @@ -1465,6 +1537,15 @@ struct bpf_prog_aux { struct work_struct work; struct rcu_head rcu; }; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + KABI_RESERVE(5) + KABI_RESERVE(6) + KABI_RESERVE(7) + KABI_RESERVE(8) }; struct bpf_prog { @@ -1499,6 +1580,11 @@ struct bpf_prog { DECLARE_FLEX_ARRAY(struct sock_filter, insns); DECLARE_FLEX_ARRAY(struct bpf_insn, insnsi); }; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_array_aux { @@ -1507,6 +1593,9 @@ struct bpf_array_aux { struct bpf_map *map; struct mutex poke_mutex; struct work_struct work; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_link { @@ -1516,6 +1605,9 @@ struct bpf_link { const struct bpf_link_ops *ops; struct bpf_prog *prog; struct work_struct work; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_link_ops { @@ -1529,12 +1621,18 @@ struct bpf_link_ops { struct bpf_link_info *info); int (*update_map)(struct bpf_link *link, struct bpf_map *new_map, struct bpf_map *old_map); + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_tramp_link { struct bpf_link link; struct hlist_node tramp_hlist; u64 cookie; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_shim_tramp_link { @@ -1554,6 +1652,9 @@ struct bpf_link_primer { struct file *file; int fd; u32 id; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_struct_ops_value; @@ -1626,6 +1727,11 @@ struct bpf_struct_ops { struct btf_func_model func_models[BPF_STRUCT_OPS_MAX_NR_MEMBERS]; u32 type_id; u32 value_id; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #if defined(CONFIG_BPF_JIT) && defined(CONFIG_BPF_SYSCALL) @@ -1727,6 +1833,9 @@ struct bpf_array { DECLARE_FLEX_ARRAY(void *, ptrs) __aligned(8); DECLARE_FLEX_ARRAY(void __percpu *, pptrs) __aligned(8); }; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; #define BPF_COMPLEXITY_LIMIT_INSNS 1000000 /* yes. 1M insns */ @@ -2222,6 +2331,9 @@ struct bpf_iter_aux_info { enum bpf_iter_task_type type; u32 pid; } task; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; typedef int (*bpf_iter_attach_target_t)(struct bpf_prog *prog, @@ -2252,6 +2364,9 @@ struct bpf_iter_reg { u32 feature; struct bpf_ctx_arg_aux ctx_arg_info[BPF_ITER_CTX_ARG_MAX]; const struct bpf_iter_seq_info *seq_info; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_iter_meta { diff --git a/include/linux/bpf_verifier.h b/include/linux/bpf_verifier.h index b6e58dab8e27..78ad34886e17 100644 --- a/include/linux/bpf_verifier.h +++ b/include/linux/bpf_verifier.h @@ -8,6 +8,7 @@ #include <linux/btf.h> /* for struct btf and btf_id() */ #include <linux/filter.h> /* for MAX_BPF_STACK */ #include <linux/tnum.h> +#include <linux/kabi.h> /* Maximum variable offset umax_value permitted when resolving memory accesses. * In practice this is far bigger than any realistic pointer offset; this limit @@ -225,6 +226,11 @@ struct bpf_reg_state { enum bpf_reg_liveness live; /* if (!precise && SCALAR_VALUE) min/max/tnum don't affect safety */ bool precise; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; enum bpf_stack_slot_type { @@ -273,6 +279,11 @@ struct bpf_reference_state { * exiting a callback function. */ int callback_ref; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; /* state of the program: @@ -306,6 +317,11 @@ struct bpf_func_state { struct bpf_reference_state *refs; int allocated_stack; struct bpf_stack_state *stack; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_idx_pair { @@ -383,6 +399,11 @@ struct bpf_verifier_state { */ struct bpf_idx_pair *jmp_history; u32 jmp_history_cnt; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #define bpf_get_spilled_reg(slot, frame) \ @@ -490,6 +511,11 @@ struct bpf_insn_aux_data { * this instruction, regardless of any heuristics */ bool force_checkpoint; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #define MAX_USED_MAPS 64 /* max number of maps accessed by one eBPF program */ @@ -541,6 +567,11 @@ struct bpf_subprog_info { bool tail_call_reachable; bool has_ld_abs; bool is_async_cb; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_verifier_env; @@ -640,6 +671,11 @@ struct bpf_verifier_env { * e.g., in reg_type_str() to generate reg_type string */ char tmp_str_buf[TMP_STR_BUF_LEN]; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; __printf(2, 0) void bpf_verifier_vlog(struct bpf_verifier_log *log, diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h index 4924f0cde1bc..d0171d5461a5 100644 --- a/include/uapi/linux/bpf.h +++ b/include/uapi/linux/bpf.h @@ -944,6 +944,10 @@ enum bpf_map_type { BPF_MAP_TYPE_BLOOM_FILTER, BPF_MAP_TYPE_USER_RINGBUF, BPF_MAP_TYPE_CGRP_STORAGE, + BPF_MAP_TYPE_RESERVE_1, + BPF_MAP_TYPE_RESERVE_2, + BPF_MAP_TYPE_RESERVE_3, + BPF_MAP_TYPE_RESERVE_4, }; /* Note that tracing related programs such as @@ -988,6 +992,10 @@ enum bpf_prog_type { BPF_PROG_TYPE_SK_LOOKUP, BPF_PROG_TYPE_SYSCALL, /* a program that can execute syscalls */ BPF_PROG_TYPE_NETFILTER, + BPF_PROG_TYPE_RESERVE_1, + BPF_PROG_TYPE_RESERVE_2, + BPF_PROG_TYPE_RESERVE_3, + BPF_PROG_TYPE_RESERVE_4, }; enum bpf_attach_type { @@ -1040,6 +1048,10 @@ enum bpf_attach_type { BPF_TCX_INGRESS, BPF_TCX_EGRESS, BPF_TRACE_UPROBE_MULTI, + BPF_ATTACH_TYPE_RESERVE_1, + BPF_ATTACH_TYPE_RESERVE_2, + BPF_ATTACH_TYPE_RESERVE_3, + BPF_ATTACH_TYPE_RESERVE_4, __MAX_BPF_ATTACH_TYPE }; @@ -1059,6 +1071,10 @@ enum bpf_link_type { BPF_LINK_TYPE_NETFILTER = 10, BPF_LINK_TYPE_TCX = 11, BPF_LINK_TYPE_UPROBE_MULTI = 12, + BPF_LINK_TYPE_RESERVE_1 = 13, + BPF_LINK_TYPE_RESERVE_2 = 14, + BPF_LINK_TYPE_RESERVE_3 = 15, + BPF_LINK_TYPE_RESERVE_4 = 16, MAX_BPF_LINK_TYPE, }; -- 2.34.1

1 0

[PATCH OLK-6.6] kabi: reserve space for bpf related structures
by Pu Lehui 09 Jan '24

09 Jan '24

From: Pu Lehui <pulehui(a)huaweicloud.com> hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8UOP9 -------------------------------- reserve space for bpf related structures. Signed-off-by: Pu Lehui <pulehui(a)huawei.com> --- include/linux/bpf.h | 115 +++++++++++++++++++++++++++++++++++ include/linux/bpf_verifier.h | 36 +++++++++++ include/uapi/linux/bpf.h | 16 +++++ 3 files changed, 167 insertions(+) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index 392f581af2ce..f37e681cd4f5 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -29,6 +29,7 @@ #include <linux/rcupdate_trace.h> #include <linux/static_call.h> #include <linux/memcontrol.h> +#include <linux/kabi.h> struct bpf_verifier_env; struct bpf_verifier_log; @@ -163,6 +164,11 @@ struct bpf_map_ops { u64 (*map_mem_usage)(const struct bpf_map *map); + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + /* BTF id of struct allocated by map_alloc */ int *map_btf_id; @@ -288,6 +294,11 @@ struct bpf_map { bool bypass_spec_v1; bool frozen; /* write-once; write-protected by freeze_mutex */ s64 __percpu *elem_count; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; static inline const char *btf_field_type_name(enum btf_field_type type) @@ -664,6 +675,13 @@ enum bpf_type_flag { /* DYNPTR points to xdp_buff */ DYNPTR_TYPE_XDP = BIT(16 + BPF_BASE_TYPE_BITS), +#ifdef CONFIG_KABI_RESERVE + BPF_TYPE_FLAG_RESERVE_1 = BIT(17 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_2 = BIT(18 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_3 = BIT(19 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_4 = BIT(20 + BPF_BASE_TYPE_BITS), +#endif + __BPF_TYPE_FLAG_MAX, __BPF_TYPE_LAST_FLAG = __BPF_TYPE_FLAG_MAX - 1, }; @@ -730,6 +748,13 @@ enum bpf_arg_type { /* Pointer to valid memory of size known at compile time. */ ARG_PTR_TO_FIXED_SIZE_MEM = MEM_FIXED_SIZE | ARG_PTR_TO_MEM, +#ifdef CONFIG_KABI_RESERVE + BPF_ARG_TYPE_RESERVE_1, + BPF_ARG_TYPE_RESERVE_2, + BPF_ARG_TYPE_RESERVE_3, + BPF_ARG_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -760,6 +785,13 @@ enum bpf_return_type { RET_PTR_TO_BTF_ID_OR_NULL = PTR_MAYBE_NULL | RET_PTR_TO_BTF_ID, RET_PTR_TO_BTF_ID_TRUSTED = PTR_TRUSTED | RET_PTR_TO_BTF_ID, +#ifdef CONFIG_KABI_RESERVE + BPF_RET_TYPE_RESERVE_1, + BPF_RET_TYPE_RESERVE_2, + BPF_RET_TYPE_RESERVE_3, + BPF_RET_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -807,6 +839,9 @@ struct bpf_func_proto { }; int *ret_btf_id; /* return value btf_id */ bool (*allowed)(const struct bpf_prog *prog); + + KABI_RESERVE(1) + KABI_RESERVE(2) }; /* bpf_context is intentionally undefined structure. Pointer to bpf_context is @@ -875,6 +910,13 @@ enum bpf_reg_type { PTR_TO_TCP_SOCK_OR_NULL = PTR_MAYBE_NULL | PTR_TO_TCP_SOCK, PTR_TO_BTF_ID_OR_NULL = PTR_MAYBE_NULL | PTR_TO_BTF_ID, +#ifdef CONFIG_KABI_RESERVE + BPF_REG_TYPE_RESERVE_1, + BPF_REG_TYPE_RESERVE_2, + BPF_REG_TYPE_RESERVE_3, + BPF_REG_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -895,6 +937,9 @@ struct bpf_insn_access_aux { }; }; struct bpf_verifier_log *log; /* for verbose logs */ + + KABI_RESERVE(1) + KABI_RESERVE(2) }; static inline void @@ -942,6 +987,11 @@ struct bpf_verifier_ops { int (*btf_struct_access)(struct bpf_verifier_log *log, const struct bpf_reg_state *reg, int off, int size); + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_prog_offload_ops { @@ -1121,6 +1171,9 @@ struct bpf_tramp_image { struct rcu_head rcu; struct work_struct work; }; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_trampoline { @@ -1149,6 +1202,11 @@ struct bpf_trampoline { /* Executable image of trampoline */ struct bpf_tramp_image *cur_image; struct module *mod; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_attach_target_info { @@ -1180,6 +1238,11 @@ struct bpf_dispatcher { struct static_call_key *sc_key; void *sc_tramp; #endif + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; static __always_inline __nocfi unsigned int bpf_dispatcher_nop_func( @@ -1339,6 +1402,9 @@ static inline bool bpf_prog_has_trampoline(const struct bpf_prog *prog) struct bpf_func_info_aux { u16 linkage; bool unreliable; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; enum bpf_jit_poke_reason { @@ -1361,6 +1427,9 @@ struct bpf_jit_poke_descriptor { u8 adj_off; u16 reason; u32 insn_idx; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; /* reg_type info for ctx arguments */ @@ -1368,6 +1437,9 @@ struct bpf_ctx_arg_aux { u32 offset; enum bpf_reg_type reg_type; u32 btf_id; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct btf_mod_pair { @@ -1465,6 +1537,15 @@ struct bpf_prog_aux { struct work_struct work; struct rcu_head rcu; }; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + KABI_RESERVE(5) + KABI_RESERVE(6) + KABI_RESERVE(7) + KABI_RESERVE(8) }; struct bpf_prog { @@ -1499,6 +1580,11 @@ struct bpf_prog { DECLARE_FLEX_ARRAY(struct sock_filter, insns); DECLARE_FLEX_ARRAY(struct bpf_insn, insnsi); }; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_array_aux { @@ -1507,6 +1593,9 @@ struct bpf_array_aux { struct bpf_map *map; struct mutex poke_mutex; struct work_struct work; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_link { @@ -1516,6 +1605,9 @@ struct bpf_link { const struct bpf_link_ops *ops; struct bpf_prog *prog; struct work_struct work; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_link_ops { @@ -1529,12 +1621,18 @@ struct bpf_link_ops { struct bpf_link_info *info); int (*update_map)(struct bpf_link *link, struct bpf_map *new_map, struct bpf_map *old_map); + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_tramp_link { struct bpf_link link; struct hlist_node tramp_hlist; u64 cookie; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_shim_tramp_link { @@ -1554,6 +1652,9 @@ struct bpf_link_primer { struct file *file; int fd; u32 id; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_struct_ops_value; @@ -1626,6 +1727,11 @@ struct bpf_struct_ops { struct btf_func_model func_models[BPF_STRUCT_OPS_MAX_NR_MEMBERS]; u32 type_id; u32 value_id; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #if defined(CONFIG_BPF_JIT) && defined(CONFIG_BPF_SYSCALL) @@ -1727,6 +1833,9 @@ struct bpf_array { DECLARE_FLEX_ARRAY(void *, ptrs) __aligned(8); DECLARE_FLEX_ARRAY(void __percpu *, pptrs) __aligned(8); }; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; #define BPF_COMPLEXITY_LIMIT_INSNS 1000000 /* yes. 1M insns */ @@ -2222,6 +2331,9 @@ struct bpf_iter_aux_info { enum bpf_iter_task_type type; u32 pid; } task; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; typedef int (*bpf_iter_attach_target_t)(struct bpf_prog *prog, @@ -2252,6 +2364,9 @@ struct bpf_iter_reg { u32 feature; struct bpf_ctx_arg_aux ctx_arg_info[BPF_ITER_CTX_ARG_MAX]; const struct bpf_iter_seq_info *seq_info; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_iter_meta { diff --git a/include/linux/bpf_verifier.h b/include/linux/bpf_verifier.h index b6e58dab8e27..78ad34886e17 100644 --- a/include/linux/bpf_verifier.h +++ b/include/linux/bpf_verifier.h @@ -8,6 +8,7 @@ #include <linux/btf.h> /* for struct btf and btf_id() */ #include <linux/filter.h> /* for MAX_BPF_STACK */ #include <linux/tnum.h> +#include <linux/kabi.h> /* Maximum variable offset umax_value permitted when resolving memory accesses. * In practice this is far bigger than any realistic pointer offset; this limit @@ -225,6 +226,11 @@ struct bpf_reg_state { enum bpf_reg_liveness live; /* if (!precise && SCALAR_VALUE) min/max/tnum don't affect safety */ bool precise; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; enum bpf_stack_slot_type { @@ -273,6 +279,11 @@ struct bpf_reference_state { * exiting a callback function. */ int callback_ref; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; /* state of the program: @@ -306,6 +317,11 @@ struct bpf_func_state { struct bpf_reference_state *refs; int allocated_stack; struct bpf_stack_state *stack; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_idx_pair { @@ -383,6 +399,11 @@ struct bpf_verifier_state { */ struct bpf_idx_pair *jmp_history; u32 jmp_history_cnt; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #define bpf_get_spilled_reg(slot, frame) \ @@ -490,6 +511,11 @@ struct bpf_insn_aux_data { * this instruction, regardless of any heuristics */ bool force_checkpoint; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #define MAX_USED_MAPS 64 /* max number of maps accessed by one eBPF program */ @@ -541,6 +567,11 @@ struct bpf_subprog_info { bool tail_call_reachable; bool has_ld_abs; bool is_async_cb; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_verifier_env; @@ -640,6 +671,11 @@ struct bpf_verifier_env { * e.g., in reg_type_str() to generate reg_type string */ char tmp_str_buf[TMP_STR_BUF_LEN]; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; __printf(2, 0) void bpf_verifier_vlog(struct bpf_verifier_log *log, diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h index 4924f0cde1bc..d0171d5461a5 100644 --- a/include/uapi/linux/bpf.h +++ b/include/uapi/linux/bpf.h @@ -944,6 +944,10 @@ enum bpf_map_type { BPF_MAP_TYPE_BLOOM_FILTER, BPF_MAP_TYPE_USER_RINGBUF, BPF_MAP_TYPE_CGRP_STORAGE, + BPF_MAP_TYPE_RESERVE_1, + BPF_MAP_TYPE_RESERVE_2, + BPF_MAP_TYPE_RESERVE_3, + BPF_MAP_TYPE_RESERVE_4, }; /* Note that tracing related programs such as @@ -988,6 +992,10 @@ enum bpf_prog_type { BPF_PROG_TYPE_SK_LOOKUP, BPF_PROG_TYPE_SYSCALL, /* a program that can execute syscalls */ BPF_PROG_TYPE_NETFILTER, + BPF_PROG_TYPE_RESERVE_1, + BPF_PROG_TYPE_RESERVE_2, + BPF_PROG_TYPE_RESERVE_3, + BPF_PROG_TYPE_RESERVE_4, }; enum bpf_attach_type { @@ -1040,6 +1048,10 @@ enum bpf_attach_type { BPF_TCX_INGRESS, BPF_TCX_EGRESS, BPF_TRACE_UPROBE_MULTI, + BPF_ATTACH_TYPE_RESERVE_1, + BPF_ATTACH_TYPE_RESERVE_2, + BPF_ATTACH_TYPE_RESERVE_3, + BPF_ATTACH_TYPE_RESERVE_4, __MAX_BPF_ATTACH_TYPE }; @@ -1059,6 +1071,10 @@ enum bpf_link_type { BPF_LINK_TYPE_NETFILTER = 10, BPF_LINK_TYPE_TCX = 11, BPF_LINK_TYPE_UPROBE_MULTI = 12, + BPF_LINK_TYPE_RESERVE_1 = 13, + BPF_LINK_TYPE_RESERVE_2 = 14, + BPF_LINK_TYPE_RESERVE_3 = 15, + BPF_LINK_TYPE_RESERVE_4 = 16, MAX_BPF_LINK_TYPE, }; -- 2.34.1

1 0

[PATCH OLK-6.6] kabi: reserve space for bpf related structures
by Pu Lehui 09 Jan '24

09 Jan '24

From: Pu Lehui <pulehui(a)huaweicloud.com> hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8UOP9 -------------------------------- reserve space for bpf related structures. Signed-off-by: Pu Lehui <pulehui(a)huawei.com> --- include/linux/bpf.h | 115 +++++++++++++++++++++++++++++++++++ include/linux/bpf_verifier.h | 36 +++++++++++ include/uapi/linux/bpf.h | 16 +++++ 3 files changed, 167 insertions(+) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index 392f581af2ce..f37e681cd4f5 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -29,6 +29,7 @@ #include <linux/rcupdate_trace.h> #include <linux/static_call.h> #include <linux/memcontrol.h> +#include <linux/kabi.h> struct bpf_verifier_env; struct bpf_verifier_log; @@ -163,6 +164,11 @@ struct bpf_map_ops { u64 (*map_mem_usage)(const struct bpf_map *map); + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + /* BTF id of struct allocated by map_alloc */ int *map_btf_id; @@ -288,6 +294,11 @@ struct bpf_map { bool bypass_spec_v1; bool frozen; /* write-once; write-protected by freeze_mutex */ s64 __percpu *elem_count; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; static inline const char *btf_field_type_name(enum btf_field_type type) @@ -664,6 +675,13 @@ enum bpf_type_flag { /* DYNPTR points to xdp_buff */ DYNPTR_TYPE_XDP = BIT(16 + BPF_BASE_TYPE_BITS), +#ifdef CONFIG_KABI_RESERVE + BPF_TYPE_FLAG_RESERVE_1 = BIT(17 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_2 = BIT(18 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_3 = BIT(19 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_4 = BIT(20 + BPF_BASE_TYPE_BITS), +#endif + __BPF_TYPE_FLAG_MAX, __BPF_TYPE_LAST_FLAG = __BPF_TYPE_FLAG_MAX - 1, }; @@ -730,6 +748,13 @@ enum bpf_arg_type { /* Pointer to valid memory of size known at compile time. */ ARG_PTR_TO_FIXED_SIZE_MEM = MEM_FIXED_SIZE | ARG_PTR_TO_MEM, +#ifdef CONFIG_KABI_RESERVE + BPF_ARG_TYPE_RESERVE_1, + BPF_ARG_TYPE_RESERVE_2, + BPF_ARG_TYPE_RESERVE_3, + BPF_ARG_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -760,6 +785,13 @@ enum bpf_return_type { RET_PTR_TO_BTF_ID_OR_NULL = PTR_MAYBE_NULL | RET_PTR_TO_BTF_ID, RET_PTR_TO_BTF_ID_TRUSTED = PTR_TRUSTED | RET_PTR_TO_BTF_ID, +#ifdef CONFIG_KABI_RESERVE + BPF_RET_TYPE_RESERVE_1, + BPF_RET_TYPE_RESERVE_2, + BPF_RET_TYPE_RESERVE_3, + BPF_RET_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -807,6 +839,9 @@ struct bpf_func_proto { }; int *ret_btf_id; /* return value btf_id */ bool (*allowed)(const struct bpf_prog *prog); + + KABI_RESERVE(1) + KABI_RESERVE(2) }; /* bpf_context is intentionally undefined structure. Pointer to bpf_context is @@ -875,6 +910,13 @@ enum bpf_reg_type { PTR_TO_TCP_SOCK_OR_NULL = PTR_MAYBE_NULL | PTR_TO_TCP_SOCK, PTR_TO_BTF_ID_OR_NULL = PTR_MAYBE_NULL | PTR_TO_BTF_ID, +#ifdef CONFIG_KABI_RESERVE + BPF_REG_TYPE_RESERVE_1, + BPF_REG_TYPE_RESERVE_2, + BPF_REG_TYPE_RESERVE_3, + BPF_REG_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -895,6 +937,9 @@ struct bpf_insn_access_aux { }; }; struct bpf_verifier_log *log; /* for verbose logs */ + + KABI_RESERVE(1) + KABI_RESERVE(2) }; static inline void @@ -942,6 +987,11 @@ struct bpf_verifier_ops { int (*btf_struct_access)(struct bpf_verifier_log *log, const struct bpf_reg_state *reg, int off, int size); + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_prog_offload_ops { @@ -1121,6 +1171,9 @@ struct bpf_tramp_image { struct rcu_head rcu; struct work_struct work; }; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_trampoline { @@ -1149,6 +1202,11 @@ struct bpf_trampoline { /* Executable image of trampoline */ struct bpf_tramp_image *cur_image; struct module *mod; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_attach_target_info { @@ -1180,6 +1238,11 @@ struct bpf_dispatcher { struct static_call_key *sc_key; void *sc_tramp; #endif + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; static __always_inline __nocfi unsigned int bpf_dispatcher_nop_func( @@ -1339,6 +1402,9 @@ static inline bool bpf_prog_has_trampoline(const struct bpf_prog *prog) struct bpf_func_info_aux { u16 linkage; bool unreliable; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; enum bpf_jit_poke_reason { @@ -1361,6 +1427,9 @@ struct bpf_jit_poke_descriptor { u8 adj_off; u16 reason; u32 insn_idx; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; /* reg_type info for ctx arguments */ @@ -1368,6 +1437,9 @@ struct bpf_ctx_arg_aux { u32 offset; enum bpf_reg_type reg_type; u32 btf_id; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct btf_mod_pair { @@ -1465,6 +1537,15 @@ struct bpf_prog_aux { struct work_struct work; struct rcu_head rcu; }; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + KABI_RESERVE(5) + KABI_RESERVE(6) + KABI_RESERVE(7) + KABI_RESERVE(8) }; struct bpf_prog { @@ -1499,6 +1580,11 @@ struct bpf_prog { DECLARE_FLEX_ARRAY(struct sock_filter, insns); DECLARE_FLEX_ARRAY(struct bpf_insn, insnsi); }; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_array_aux { @@ -1507,6 +1593,9 @@ struct bpf_array_aux { struct bpf_map *map; struct mutex poke_mutex; struct work_struct work; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_link { @@ -1516,6 +1605,9 @@ struct bpf_link { const struct bpf_link_ops *ops; struct bpf_prog *prog; struct work_struct work; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_link_ops { @@ -1529,12 +1621,18 @@ struct bpf_link_ops { struct bpf_link_info *info); int (*update_map)(struct bpf_link *link, struct bpf_map *new_map, struct bpf_map *old_map); + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_tramp_link { struct bpf_link link; struct hlist_node tramp_hlist; u64 cookie; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_shim_tramp_link { @@ -1554,6 +1652,9 @@ struct bpf_link_primer { struct file *file; int fd; u32 id; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_struct_ops_value; @@ -1626,6 +1727,11 @@ struct bpf_struct_ops { struct btf_func_model func_models[BPF_STRUCT_OPS_MAX_NR_MEMBERS]; u32 type_id; u32 value_id; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #if defined(CONFIG_BPF_JIT) && defined(CONFIG_BPF_SYSCALL) @@ -1727,6 +1833,9 @@ struct bpf_array { DECLARE_FLEX_ARRAY(void *, ptrs) __aligned(8); DECLARE_FLEX_ARRAY(void __percpu *, pptrs) __aligned(8); }; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; #define BPF_COMPLEXITY_LIMIT_INSNS 1000000 /* yes. 1M insns */ @@ -2222,6 +2331,9 @@ struct bpf_iter_aux_info { enum bpf_iter_task_type type; u32 pid; } task; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; typedef int (*bpf_iter_attach_target_t)(struct bpf_prog *prog, @@ -2252,6 +2364,9 @@ struct bpf_iter_reg { u32 feature; struct bpf_ctx_arg_aux ctx_arg_info[BPF_ITER_CTX_ARG_MAX]; const struct bpf_iter_seq_info *seq_info; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_iter_meta { diff --git a/include/linux/bpf_verifier.h b/include/linux/bpf_verifier.h index b6e58dab8e27..78ad34886e17 100644 --- a/include/linux/bpf_verifier.h +++ b/include/linux/bpf_verifier.h @@ -8,6 +8,7 @@ #include <linux/btf.h> /* for struct btf and btf_id() */ #include <linux/filter.h> /* for MAX_BPF_STACK */ #include <linux/tnum.h> +#include <linux/kabi.h> /* Maximum variable offset umax_value permitted when resolving memory accesses. * In practice this is far bigger than any realistic pointer offset; this limit @@ -225,6 +226,11 @@ struct bpf_reg_state { enum bpf_reg_liveness live; /* if (!precise && SCALAR_VALUE) min/max/tnum don't affect safety */ bool precise; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; enum bpf_stack_slot_type { @@ -273,6 +279,11 @@ struct bpf_reference_state { * exiting a callback function. */ int callback_ref; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; /* state of the program: @@ -306,6 +317,11 @@ struct bpf_func_state { struct bpf_reference_state *refs; int allocated_stack; struct bpf_stack_state *stack; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_idx_pair { @@ -383,6 +399,11 @@ struct bpf_verifier_state { */ struct bpf_idx_pair *jmp_history; u32 jmp_history_cnt; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #define bpf_get_spilled_reg(slot, frame) \ @@ -490,6 +511,11 @@ struct bpf_insn_aux_data { * this instruction, regardless of any heuristics */ bool force_checkpoint; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #define MAX_USED_MAPS 64 /* max number of maps accessed by one eBPF program */ @@ -541,6 +567,11 @@ struct bpf_subprog_info { bool tail_call_reachable; bool has_ld_abs; bool is_async_cb; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_verifier_env; @@ -640,6 +671,11 @@ struct bpf_verifier_env { * e.g., in reg_type_str() to generate reg_type string */ char tmp_str_buf[TMP_STR_BUF_LEN]; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; __printf(2, 0) void bpf_verifier_vlog(struct bpf_verifier_log *log, diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h index 4924f0cde1bc..d0171d5461a5 100644 --- a/include/uapi/linux/bpf.h +++ b/include/uapi/linux/bpf.h @@ -944,6 +944,10 @@ enum bpf_map_type { BPF_MAP_TYPE_BLOOM_FILTER, BPF_MAP_TYPE_USER_RINGBUF, BPF_MAP_TYPE_CGRP_STORAGE, + BPF_MAP_TYPE_RESERVE_1, + BPF_MAP_TYPE_RESERVE_2, + BPF_MAP_TYPE_RESERVE_3, + BPF_MAP_TYPE_RESERVE_4, }; /* Note that tracing related programs such as @@ -988,6 +992,10 @@ enum bpf_prog_type { BPF_PROG_TYPE_SK_LOOKUP, BPF_PROG_TYPE_SYSCALL, /* a program that can execute syscalls */ BPF_PROG_TYPE_NETFILTER, + BPF_PROG_TYPE_RESERVE_1, + BPF_PROG_TYPE_RESERVE_2, + BPF_PROG_TYPE_RESERVE_3, + BPF_PROG_TYPE_RESERVE_4, }; enum bpf_attach_type { @@ -1040,6 +1048,10 @@ enum bpf_attach_type { BPF_TCX_INGRESS, BPF_TCX_EGRESS, BPF_TRACE_UPROBE_MULTI, + BPF_ATTACH_TYPE_RESERVE_1, + BPF_ATTACH_TYPE_RESERVE_2, + BPF_ATTACH_TYPE_RESERVE_3, + BPF_ATTACH_TYPE_RESERVE_4, __MAX_BPF_ATTACH_TYPE }; @@ -1059,6 +1071,10 @@ enum bpf_link_type { BPF_LINK_TYPE_NETFILTER = 10, BPF_LINK_TYPE_TCX = 11, BPF_LINK_TYPE_UPROBE_MULTI = 12, + BPF_LINK_TYPE_RESERVE_1 = 13, + BPF_LINK_TYPE_RESERVE_2 = 14, + BPF_LINK_TYPE_RESERVE_3 = 15, + BPF_LINK_TYPE_RESERVE_4 = 16, MAX_BPF_LINK_TYPE, }; -- 2.34.1

1 0

[PATCH OLK-6.6] kabi: reserve space for bpf related structures
by Pu Lehui 09 Jan '24

09 Jan '24

From: Pu Lehui <pulehui(a)huaweicloud.com> hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8UOP9 -------------------------------- reserve space for bpf related structures. Signed-off-by: Pu Lehui <pulehui(a)huawei.com> --- include/linux/bpf.h | 115 +++++++++++++++++++++++++++++++++++ include/linux/bpf_verifier.h | 36 +++++++++++ include/uapi/linux/bpf.h | 16 +++++ 3 files changed, 167 insertions(+) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index 392f581af2ce..f37e681cd4f5 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -29,6 +29,7 @@ #include <linux/rcupdate_trace.h> #include <linux/static_call.h> #include <linux/memcontrol.h> +#include <linux/kabi.h> struct bpf_verifier_env; struct bpf_verifier_log; @@ -163,6 +164,11 @@ struct bpf_map_ops { u64 (*map_mem_usage)(const struct bpf_map *map); + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + /* BTF id of struct allocated by map_alloc */ int *map_btf_id; @@ -288,6 +294,11 @@ struct bpf_map { bool bypass_spec_v1; bool frozen; /* write-once; write-protected by freeze_mutex */ s64 __percpu *elem_count; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; static inline const char *btf_field_type_name(enum btf_field_type type) @@ -664,6 +675,13 @@ enum bpf_type_flag { /* DYNPTR points to xdp_buff */ DYNPTR_TYPE_XDP = BIT(16 + BPF_BASE_TYPE_BITS), +#ifdef CONFIG_KABI_RESERVE + BPF_TYPE_FLAG_RESERVE_1 = BIT(17 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_2 = BIT(18 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_3 = BIT(19 + BPF_BASE_TYPE_BITS), + BPF_TYPE_FLAG_RESERVE_4 = BIT(20 + BPF_BASE_TYPE_BITS), +#endif + __BPF_TYPE_FLAG_MAX, __BPF_TYPE_LAST_FLAG = __BPF_TYPE_FLAG_MAX - 1, }; @@ -730,6 +748,13 @@ enum bpf_arg_type { /* Pointer to valid memory of size known at compile time. */ ARG_PTR_TO_FIXED_SIZE_MEM = MEM_FIXED_SIZE | ARG_PTR_TO_MEM, +#ifdef CONFIG_KABI_RESERVE + BPF_ARG_TYPE_RESERVE_1, + BPF_ARG_TYPE_RESERVE_2, + BPF_ARG_TYPE_RESERVE_3, + BPF_ARG_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -760,6 +785,13 @@ enum bpf_return_type { RET_PTR_TO_BTF_ID_OR_NULL = PTR_MAYBE_NULL | RET_PTR_TO_BTF_ID, RET_PTR_TO_BTF_ID_TRUSTED = PTR_TRUSTED | RET_PTR_TO_BTF_ID, +#ifdef CONFIG_KABI_RESERVE + BPF_RET_TYPE_RESERVE_1, + BPF_RET_TYPE_RESERVE_2, + BPF_RET_TYPE_RESERVE_3, + BPF_RET_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -807,6 +839,9 @@ struct bpf_func_proto { }; int *ret_btf_id; /* return value btf_id */ bool (*allowed)(const struct bpf_prog *prog); + + KABI_RESERVE(1) + KABI_RESERVE(2) }; /* bpf_context is intentionally undefined structure. Pointer to bpf_context is @@ -875,6 +910,13 @@ enum bpf_reg_type { PTR_TO_TCP_SOCK_OR_NULL = PTR_MAYBE_NULL | PTR_TO_TCP_SOCK, PTR_TO_BTF_ID_OR_NULL = PTR_MAYBE_NULL | PTR_TO_BTF_ID, +#ifdef CONFIG_KABI_RESERVE + BPF_REG_TYPE_RESERVE_1, + BPF_REG_TYPE_RESERVE_2, + BPF_REG_TYPE_RESERVE_3, + BPF_REG_TYPE_RESERVE_4, +#endif + /* This must be the last entry. Its purpose is to ensure the enum is * wide enough to hold the higher bits reserved for bpf_type_flag. */ @@ -895,6 +937,9 @@ struct bpf_insn_access_aux { }; }; struct bpf_verifier_log *log; /* for verbose logs */ + + KABI_RESERVE(1) + KABI_RESERVE(2) }; static inline void @@ -942,6 +987,11 @@ struct bpf_verifier_ops { int (*btf_struct_access)(struct bpf_verifier_log *log, const struct bpf_reg_state *reg, int off, int size); + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_prog_offload_ops { @@ -1121,6 +1171,9 @@ struct bpf_tramp_image { struct rcu_head rcu; struct work_struct work; }; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_trampoline { @@ -1149,6 +1202,11 @@ struct bpf_trampoline { /* Executable image of trampoline */ struct bpf_tramp_image *cur_image; struct module *mod; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_attach_target_info { @@ -1180,6 +1238,11 @@ struct bpf_dispatcher { struct static_call_key *sc_key; void *sc_tramp; #endif + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; static __always_inline __nocfi unsigned int bpf_dispatcher_nop_func( @@ -1339,6 +1402,9 @@ static inline bool bpf_prog_has_trampoline(const struct bpf_prog *prog) struct bpf_func_info_aux { u16 linkage; bool unreliable; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; enum bpf_jit_poke_reason { @@ -1361,6 +1427,9 @@ struct bpf_jit_poke_descriptor { u8 adj_off; u16 reason; u32 insn_idx; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; /* reg_type info for ctx arguments */ @@ -1368,6 +1437,9 @@ struct bpf_ctx_arg_aux { u32 offset; enum bpf_reg_type reg_type; u32 btf_id; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct btf_mod_pair { @@ -1465,6 +1537,15 @@ struct bpf_prog_aux { struct work_struct work; struct rcu_head rcu; }; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + KABI_RESERVE(5) + KABI_RESERVE(6) + KABI_RESERVE(7) + KABI_RESERVE(8) }; struct bpf_prog { @@ -1499,6 +1580,11 @@ struct bpf_prog { DECLARE_FLEX_ARRAY(struct sock_filter, insns); DECLARE_FLEX_ARRAY(struct bpf_insn, insnsi); }; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_array_aux { @@ -1507,6 +1593,9 @@ struct bpf_array_aux { struct bpf_map *map; struct mutex poke_mutex; struct work_struct work; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_link { @@ -1516,6 +1605,9 @@ struct bpf_link { const struct bpf_link_ops *ops; struct bpf_prog *prog; struct work_struct work; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_link_ops { @@ -1529,12 +1621,18 @@ struct bpf_link_ops { struct bpf_link_info *info); int (*update_map)(struct bpf_link *link, struct bpf_map *new_map, struct bpf_map *old_map); + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_tramp_link { struct bpf_link link; struct hlist_node tramp_hlist; u64 cookie; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_shim_tramp_link { @@ -1554,6 +1652,9 @@ struct bpf_link_primer { struct file *file; int fd; u32 id; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_struct_ops_value; @@ -1626,6 +1727,11 @@ struct bpf_struct_ops { struct btf_func_model func_models[BPF_STRUCT_OPS_MAX_NR_MEMBERS]; u32 type_id; u32 value_id; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #if defined(CONFIG_BPF_JIT) && defined(CONFIG_BPF_SYSCALL) @@ -1727,6 +1833,9 @@ struct bpf_array { DECLARE_FLEX_ARRAY(void *, ptrs) __aligned(8); DECLARE_FLEX_ARRAY(void __percpu *, pptrs) __aligned(8); }; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; #define BPF_COMPLEXITY_LIMIT_INSNS 1000000 /* yes. 1M insns */ @@ -2222,6 +2331,9 @@ struct bpf_iter_aux_info { enum bpf_iter_task_type type; u32 pid; } task; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; typedef int (*bpf_iter_attach_target_t)(struct bpf_prog *prog, @@ -2252,6 +2364,9 @@ struct bpf_iter_reg { u32 feature; struct bpf_ctx_arg_aux ctx_arg_info[BPF_ITER_CTX_ARG_MAX]; const struct bpf_iter_seq_info *seq_info; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct bpf_iter_meta { diff --git a/include/linux/bpf_verifier.h b/include/linux/bpf_verifier.h index b6e58dab8e27..78ad34886e17 100644 --- a/include/linux/bpf_verifier.h +++ b/include/linux/bpf_verifier.h @@ -8,6 +8,7 @@ #include <linux/btf.h> /* for struct btf and btf_id() */ #include <linux/filter.h> /* for MAX_BPF_STACK */ #include <linux/tnum.h> +#include <linux/kabi.h> /* Maximum variable offset umax_value permitted when resolving memory accesses. * In practice this is far bigger than any realistic pointer offset; this limit @@ -225,6 +226,11 @@ struct bpf_reg_state { enum bpf_reg_liveness live; /* if (!precise && SCALAR_VALUE) min/max/tnum don't affect safety */ bool precise; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; enum bpf_stack_slot_type { @@ -273,6 +279,11 @@ struct bpf_reference_state { * exiting a callback function. */ int callback_ref; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; /* state of the program: @@ -306,6 +317,11 @@ struct bpf_func_state { struct bpf_reference_state *refs; int allocated_stack; struct bpf_stack_state *stack; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_idx_pair { @@ -383,6 +399,11 @@ struct bpf_verifier_state { */ struct bpf_idx_pair *jmp_history; u32 jmp_history_cnt; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #define bpf_get_spilled_reg(slot, frame) \ @@ -490,6 +511,11 @@ struct bpf_insn_aux_data { * this instruction, regardless of any heuristics */ bool force_checkpoint; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; #define MAX_USED_MAPS 64 /* max number of maps accessed by one eBPF program */ @@ -541,6 +567,11 @@ struct bpf_subprog_info { bool tail_call_reachable; bool has_ld_abs; bool is_async_cb; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; struct bpf_verifier_env; @@ -640,6 +671,11 @@ struct bpf_verifier_env { * e.g., in reg_type_str() to generate reg_type string */ char tmp_str_buf[TMP_STR_BUF_LEN]; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; __printf(2, 0) void bpf_verifier_vlog(struct bpf_verifier_log *log, diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h index 4924f0cde1bc..d0171d5461a5 100644 --- a/include/uapi/linux/bpf.h +++ b/include/uapi/linux/bpf.h @@ -944,6 +944,10 @@ enum bpf_map_type { BPF_MAP_TYPE_BLOOM_FILTER, BPF_MAP_TYPE_USER_RINGBUF, BPF_MAP_TYPE_CGRP_STORAGE, + BPF_MAP_TYPE_RESERVE_1, + BPF_MAP_TYPE_RESERVE_2, + BPF_MAP_TYPE_RESERVE_3, + BPF_MAP_TYPE_RESERVE_4, }; /* Note that tracing related programs such as @@ -988,6 +992,10 @@ enum bpf_prog_type { BPF_PROG_TYPE_SK_LOOKUP, BPF_PROG_TYPE_SYSCALL, /* a program that can execute syscalls */ BPF_PROG_TYPE_NETFILTER, + BPF_PROG_TYPE_RESERVE_1, + BPF_PROG_TYPE_RESERVE_2, + BPF_PROG_TYPE_RESERVE_3, + BPF_PROG_TYPE_RESERVE_4, }; enum bpf_attach_type { @@ -1040,6 +1048,10 @@ enum bpf_attach_type { BPF_TCX_INGRESS, BPF_TCX_EGRESS, BPF_TRACE_UPROBE_MULTI, + BPF_ATTACH_TYPE_RESERVE_1, + BPF_ATTACH_TYPE_RESERVE_2, + BPF_ATTACH_TYPE_RESERVE_3, + BPF_ATTACH_TYPE_RESERVE_4, __MAX_BPF_ATTACH_TYPE }; @@ -1059,6 +1071,10 @@ enum bpf_link_type { BPF_LINK_TYPE_NETFILTER = 10, BPF_LINK_TYPE_TCX = 11, BPF_LINK_TYPE_UPROBE_MULTI = 12, + BPF_LINK_TYPE_RESERVE_1 = 13, + BPF_LINK_TYPE_RESERVE_2 = 14, + BPF_LINK_TYPE_RESERVE_3 = 15, + BPF_LINK_TYPE_RESERVE_4 = 16, MAX_BPF_LINK_TYPE, }; -- 2.34.1

1 0

[PATCH V4 OLK-6.6 0/8] Spark SQL scenario bpf readahead optimization synchronization to OLK-6.6
by ZhaoLong Wang 09 Jan '24

09 Jan '24

CONFIG: CONFIG_DEBUG_INFO_BTF=y CONFIG_BPF_SYSCALL=y CONFIG_NF_CONNTRACK=y CONFIG_NF_CONNTRACK_MARK=y V2: - Using helpers function, such as i_size_read(), round_up - Fix commit log Signature - Fix the user-mode test program V3 - Do not use the round_up helper function. Use the correct calculation - err = -errno before printf V4 - Drop the "Conflict" in the commit log. Hou Tao (5): vfs: add bare tracepoints for vfs read and release fs: add helper fs_file_read_do_trace() xfs: add trace for read and release of regular file ext4: add trace for the read and release of regular file selftests/bpf: add demo for file read pattern detection Yufen Yu (1): readahead: introduce FMODE_CTL_WILLNEED to read first 2MB of file ZhaoLong Wang (2): VFS: Rolling Back the fmode macro definition and structure members selftests/bpf: Update the demo file_read_pattern to run on libbpf 1.0+ fs/ext4/file.c | 4 + fs/read_write.c | 38 +++++ fs/xfs/xfs_file.c | 3 + include/linux/fs.h | 37 +++++ include/trace/events/fs.h | 33 +++++ mm/readahead.c | 40 ++++- tools/testing/selftests/bpf/Makefile | 1 + .../testing/selftests/bpf/file_read_pattern.c | 76 ++++++++++ .../bpf/progs/file_read_pattern_prog.c | 137 ++++++++++++++++++ 9 files changed, 368 insertions(+), 1 deletion(-) create mode 100644 include/trace/events/fs.h create mode 100644 tools/testing/selftests/bpf/file_read_pattern.c create mode 100644 tools/testing/selftests/bpf/progs/file_read_pattern_prog.c -- 2.39.2

1 8

[PATCH OLK-6.6 0/1] arm64: Add non nmi ipi backtrace support
by Liao Chen 09 Jan '24

09 Jan '24

Dependencies: CONFIG_DETECT_HUNG_TASK=y CONFIG_WATCHDOG=y Li Zhengyu (1): arm64: Add non nmi ipi backtrace support arch/arm64/Kconfig | 8 +++++ arch/arm64/configs/openeuler_defconfig | 1 + arch/arm64/kernel/ipi_nmi.c | 41 ++++++++++++++++++++++---- 3 files changed, 45 insertions(+), 5 deletions(-) -- 2.34.1

1 1

[PATCH OLK-6.6] kabi: Reserve space for perf subsystem related structures
by Yang Jihong 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8UOCZ -------------------------------- Reserve kabi space for `struct pmu`, `struct perf_event`, `struct perf_event_context`, `struct perf_cgroup_info` and `struct perf_sample_data`. Signed-off-by: Yang Jihong <yangjihong1(a)huawei.com> --- include/linux/perf_event.h | 31 +++++++++++++++++++++++++++++++ 1 file changed, 31 insertions(+) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index e846f87e2d09..a24aa95de1fc 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -62,6 +62,7 @@ struct perf_guest_info_callbacks { #include <linux/security.h> #include <linux/static_call.h> #include <linux/lockdep.h> +#include <linux/kabi.h> #include <asm/local.h> struct perf_callchain_entry { @@ -324,6 +325,11 @@ struct pmu { /* number of address filters this PMU can do */ unsigned int nr_addr_filters; + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + /* * Fully disable/enable this PMU, can be used to protect from the PMI * as well as for lazy/batch writing of the MSRs. @@ -540,6 +546,11 @@ struct pmu { * Check period value for PERF_EVENT_IOC_PERIOD ioctl. */ int (*check_period) (struct perf_event *event, u64 value); /* optional */ + + KABI_RESERVE(5) + KABI_RESERVE(6) + KABI_RESERVE(7) + KABI_RESERVE(8) }; enum perf_addr_filter_action_t { @@ -839,6 +850,13 @@ struct perf_event { * user. */ __u32 orig_type; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) + KABI_RESERVE(5) + KABI_RESERVE(6) #endif /* CONFIG_PERF_EVENTS */ }; @@ -965,6 +983,11 @@ struct perf_event_context { * that until the signal is delivered. */ local_t nr_pending; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) }; /* @@ -1040,6 +1063,9 @@ struct perf_cgroup_info { u64 timestamp; u64 timeoffset; int active; + + KABI_RESERVE(1) + KABI_RESERVE(2) }; struct perf_cgroup { @@ -1192,6 +1218,11 @@ struct perf_sample_data { u64 data_page_size; u64 code_page_size; u64 aux_size; + + KABI_RESERVE(1) + KABI_RESERVE(2) + KABI_RESERVE(3) + KABI_RESERVE(4) } ____cacheline_aligned; /* default value for data source */ -- 2.34.1

1 0

[PATCH OLK-6.6 V1 00/12] infrastructure for scheduler bpf
by Guan Jing 09 Jan '24

09 Jan '24

infrastructure for scheduler bpf Guan Jing (12): sched: programmable: Introduce bpf sched sched: programmable: Add a tag for the task sched: programmable: Add a tag for the task group sched: programmable: Add user interface of task group tag sched: programmable: Add user interface of task tag sched: basic infrastructure for scheduler bpf sched: introduce bpf_sched_enable() libbpf: add support for scheduler bpf programs bpftool: recognize scheduler programs sched: Add helper functions to get cpu statistics sched: programmable: Add hook in select_task_rq_fair() sched: programmable: Add hook in can_migrate_task() arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/proc/base.c | 64 ++++++++++++++ include/linux/bpf_sched.h | 50 +++++++++++ include/linux/bpf_types.h | 4 + include/linux/sched.h | 38 ++++++++ include/linux/sched_hook_defs.h | 4 + include/uapi/linux/bpf.h | 9 ++ init/init_task.c | 3 + kernel/bpf/Kconfig | 12 +++ kernel/bpf/btf.c | 1 + kernel/bpf/syscall.c | 27 ++++++ kernel/bpf/trampoline.c | 1 + kernel/bpf/verifier.c | 11 ++- kernel/sched/bpf_sched.c | 102 +++++++++++++++++++++ kernel/sched/build_utility.c | 4 + kernel/sched/core.c | 118 +++++++++++++++++++++++++ kernel/sched/fair.c | 45 ++++++++++ kernel/sched/sched.h | 10 +++ scripts/bpf_doc.py | 4 + tools/include/uapi/linux/bpf.h | 9 ++ tools/lib/bpf/bpf.c | 1 + tools/lib/bpf/libbpf.c | 23 ++++- tools/lib/bpf/libbpf.h | 2 + tools/lib/bpf/libbpf.map | 1 + 25 files changed, 543 insertions(+), 2 deletions(-) create mode 100644 include/linux/bpf_sched.h create mode 100644 include/linux/sched_hook_defs.h create mode 100644 kernel/sched/bpf_sched.c -- 2.34.1

1 12

[PATCH V3 OLK-6.6] fs/dirty_pages: dump the number of dirty pages for each inode
by Zizhi Wo 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OHM4?from=project-issue CVE: NA ---------------------------------------- In order to analyse the IO performance when using buffer IO, it's useful to obtain the number of dirty pages for each inode in the filesystem. This feature depends on 'CONFIG_DIRTY_PAGES'. It creates 3 interfaces by using procfs. /proc/dirty/buffer_size for buffer allocation and release, /proc/dirty/page_threshold to filter result and /proc/dirty/dirty_list to get dirty pages. Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/Kconfig | 13 + fs/Makefile | 1 + fs/dirty_pages.c | 355 +++++++++++++++++++++++++ 5 files changed, 371 insertions(+) create mode 100644 fs/dirty_pages.c diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 66737e8c8673..677b4cc4668a 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -6924,6 +6924,7 @@ CONFIG_PROC_PAGE_MONITOR=y CONFIG_PROC_CHILDREN=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index fcbef6c06587..c16cdb7b2596 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -8120,6 +8120,7 @@ CONFIG_PROC_PID_ARCH_STATUS=y CONFIG_PROC_CPU_RESCTRL=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/fs/Kconfig b/fs/Kconfig index 7c9d45ba9b0f..6dc83223c61b 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -170,6 +170,19 @@ source "fs/proc/Kconfig" source "fs/kernfs/Kconfig" source "fs/sysfs/Kconfig" +config DIRTY_PAGES + bool "Dumps the number of dirty pages of each file" + depends on PROC_FS + default y + help + This config supports the rendering of dirty page data to the user, + which may be useful to analyze the IO performance when using buffer + IO. + + It create 3 interfaces by using procfs. /proc/dirty/buffer_size for + buffer allocation and release; /proc/dirty/page_threshold to filter + result; /proc/dirty/dirty_list to get dirty pages. + config TMPFS bool "Tmpfs virtual memory file system support (former shm fs)" depends on SHMEM diff --git a/fs/Makefile b/fs/Makefile index bb6164200a39..cc4735c23c32 100644 --- a/fs/Makefile +++ b/fs/Makefile @@ -44,6 +44,7 @@ obj-$(CONFIG_FS_POSIX_ACL) += posix_acl.o obj-$(CONFIG_NFS_COMMON) += nfs_common/ obj-$(CONFIG_COREDUMP) += coredump.o obj-$(CONFIG_SYSCTL) += drop_caches.o sysctls.o +obj-$(CONFIG_DIRTY_PAGES) += dirty_pages.o obj-$(CONFIG_FHANDLE) += fhandle.o obj-$(CONFIG_CGROUP_FILES) += filescontrol.o diff --git a/fs/dirty_pages.c b/fs/dirty_pages.c new file mode 100644 index 000000000000..d0aea022b68b --- /dev/null +++ b/fs/dirty_pages.c @@ -0,0 +1,355 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/proc_fs.h> +#include <linux/seq_file.h> +#include <linux/uaccess.h> +#include <linux/pagemap.h> +#include <linux/pagevec.h> +#include <linux/fs.h> +#include <linux/mm.h> +#include <linux/slab.h> +#include <linux/sched.h> +#include <linux/proc_fs.h> +#include <linux/kdev_t.h> +#include <linux/vmalloc.h> +#include <linux/module.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include "internal.h" + +#define DIRTY_ROOT "dirty" /* proc root directory name */ +#define DIRTY_BUFFER_SIZE "buffer_size" /* proc file to control buffer */ +#define DIRTY_PAGES "dirty_list" /* proc file to obtain diry pages */ +#define DIRTY_LIMIT "page_threshold" /* proc file to filter result */ +#define MAX_BUF_NUM 102400 /* max buffer num in number of pages */ +#define TERM_STR "terminated\n" /* tail string for overflowed buffer */ +#define UNKNOW_STR "unknown" /* for unknown file name or fs type */ + +static char *buf_dirty; /* buffer to store number of dirty pages */ +static unsigned long buf_size; /* size of buffer in bytes */ +static unsigned long buf_num; /* size of buffer in number of pages */ +static unsigned long buf_limit; /* filter threshold of dirty pages */ +unsigned long lock_word; /* for exclusive access to buffer */ + +static bool dirty_pages_lock(void) +{ + if (xchg(&lock_word, 1) == 1) + return false; + return true; +} + +static void dirty_pages_unlock(void) +{ + lock_word = 0; +} + +static unsigned long dump_dirtypages_inode(struct inode *inode) +{ + XA_STATE(xas, &inode->i_mapping->i_pages, 0); + unsigned long pages = 0; + void *page; + + if (!mapping_tagged(inode->i_mapping, PAGECACHE_TAG_DIRTY)) + return 0; + + xas_lock_irq(&xas); + xas_for_each_marked(&xas, page, (pgoff_t)-1, PAGECACHE_TAG_DIRTY) { + if (++pages % XA_CHECK_SCHED) + continue; + + xas_pause(&xas); + xas_unlock_irq(&xas); + cond_resched(); + xas_lock_irq(&xas); + } + xas_unlock_irq(&xas); + + return pages; +} + +static char *inode_filename(struct inode *inode) +{ + char *path; + static char buf[PATH_MAX]; + struct dentry *dentry; + + dentry = d_find_alias(inode); + if (!dentry) + return UNKNOW_STR; + + path = dentry_path_raw(dentry, buf, PATH_MAX); + dput(dentry); + + return IS_ERR_OR_NULL(path) ? UNKNOW_STR : path; +} + +static const char *sb_fstype(struct super_block *sb) +{ + if (sb->s_type && sb->s_type->name) + return sb->s_type->name; + else + return UNKNOW_STR; +} + +static inline bool is_sb_writable(struct super_block *sb) +{ + if (sb_rdonly(sb)) + return false; + + if (sb->s_writers.frozen == SB_FREEZE_COMPLETE) + return false; + + return true; +} + +/* + * dump_dirtypages_sb - dump the dirty pages of each inode in the sb + * @sb the super block + * @m the seq_file witch is initialized in proc_dpages_open + * + * For each inode in the sb, call dump_dirtypages_pages to get the number + * of dirty pages. And use seq_printf to store the result in the buffer + * if it's not less than the threshold. The inode in unusual state will + * be skipped. + */ +static void dump_dirtypages_sb(struct super_block *sb, struct seq_file *m) +{ + struct inode *inode, *toput_inode = NULL; + unsigned long limit; + + if (!is_sb_writable(sb) || !sb->s_bdev) + return; + + limit = READ_ONCE(buf_limit); + spin_lock(&sb->s_inode_list_lock); + list_for_each_entry(inode, &sb->s_inodes, i_sb_list) { + unsigned long dirty_pages; + + spin_lock(&inode->i_lock); + if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) { + spin_unlock(&inode->i_lock); + continue; + } + + __iget(inode); + spin_unlock(&inode->i_lock); + spin_unlock(&sb->s_inode_list_lock); + + dirty_pages = dump_dirtypages_inode(inode); + if (!dirty_pages || dirty_pages < limit) + goto skip; + + seq_printf(m, "FSType: %s, Dev ID: %u(%u:%u) ino %lu, dirty pages %lu, path %s\n", + sb_fstype(sb), sb->s_dev, MAJOR(sb->s_dev), + MINOR(sb->s_dev), inode->i_ino, dirty_pages, + inode_filename(inode)); + + if (seq_has_overflowed(m)) { + /* keep size > count to avoid overflow in seq_read_iter() */ + m->size += strlen(TERM_STR) + 1; + seq_puts(m, TERM_STR); + iput(inode); + iput(toput_inode); + return; + } +skip: + iput(toput_inode); + toput_inode = inode; + cond_resched(); + spin_lock(&sb->s_inode_list_lock); + } + spin_unlock(&sb->s_inode_list_lock); + iput(toput_inode); +} + +static int proc_dpages_show(struct seq_file *m, void *v) +{ + iterate_supers((void *)dump_dirtypages_sb, (void *)m); + return 0; +} + +static int proc_dpages_open(struct inode *inode, struct file *filp) +{ + int ret; + struct seq_file *m; + + if (!dirty_pages_lock()) + return -EBUSY; + + if (buf_dirty == NULL || buf_size == 0) { + pr_warn("please allocate buffer before getting dirty pages\n"); + dirty_pages_unlock(); + return -ENOMEM; + } + + ret = single_open(filp, proc_dpages_show, NULL); + if (ret) { + dirty_pages_unlock(); + return ret; + } + + m = (struct seq_file *)filp->private_data; + memset(buf_dirty, 0, buf_size); + m->size = buf_size - (strlen(TERM_STR) + 1); + m->buf = buf_dirty; + + return 0; +} + +static int proc_dpages_release(struct inode *inode, struct file *file) +{ + struct seq_file *m = file->private_data; + + /* we don't want to free the buf */ + m->buf = NULL; + dirty_pages_unlock(); + return single_release(inode, file); +} + +static void free_buf_dirty(void) +{ + if (!buf_dirty) + return; + + vfree(buf_dirty); + buf_dirty = NULL; + buf_size = 0; + WRITE_ONCE(buf_num, 0); +} + +static int proc_buffer_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", READ_ONCE(buf_num)); + return 0; +} + +static int proc_buffer_open(struct inode *inode, struct file *filp) +{ + int ret; + + if (((filp->f_flags & O_ACCMODE) != O_RDONLY) && !dirty_pages_lock()) + return -EBUSY; + + ret = single_open(filp, proc_buffer_show, NULL); + if (ret) + dirty_pages_unlock(); + + return ret; +} + +static ssize_t proc_buffer_write(struct file *filp, const char __user *buf, + size_t count, loff_t *offp) +{ + int ret; + unsigned long new_buf_num, new_buf_size; + char *new_buf_dirty; + + ret = kstrtoul_from_user(buf, count, 10, &new_buf_num); + if (ret != 0 || new_buf_num > MAX_BUF_NUM) + return -EINVAL; + + if (new_buf_num == buf_num) + return count; + + if (new_buf_num == 0) { + free_buf_dirty(); + return count; + } + + new_buf_size = PAGE_SIZE * new_buf_num; + new_buf_dirty = vzalloc(new_buf_size); + if (!new_buf_dirty) + return -ENOMEM; + + free_buf_dirty(); + buf_dirty = new_buf_dirty; + buf_size = new_buf_size; + WRITE_ONCE(buf_num, new_buf_num); + + return count; +} + +static int proc_buffer_release(struct inode *inode, struct file *filp) +{ + if ((filp->f_flags & O_ACCMODE) != O_RDONLY) + dirty_pages_unlock(); + return single_release(inode, filp); +} + +static int proc_limit_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", READ_ONCE(buf_limit)); + return 0; +} + +static int proc_limit_open(struct inode *inode, struct file *filp) +{ + return single_open(filp, proc_limit_show, NULL); +} + +static ssize_t proc_limit_write(struct file *filp, const char __user *buf, + size_t count, loff_t *offp) +{ + int ret; + unsigned long limit; + + ret = kstrtoul_from_user(buf, count, 10, &limit); + if (ret != 0) + return ret; + + WRITE_ONCE(buf_limit, limit); + + return count; +} + +static const struct proc_ops proc_dpages_operations = { + .proc_open = proc_dpages_open, + .proc_read = seq_read, + .proc_release = proc_dpages_release, +}; + +static const struct proc_ops proc_buffer_operations = { + .proc_open = proc_buffer_open, + .proc_read = seq_read, + .proc_write = proc_buffer_write, + .proc_lseek = seq_lseek, + .proc_release = proc_buffer_release, +}; + +static const struct proc_ops proc_limit_operations = { + .proc_open = proc_limit_open, + .proc_read = seq_read, + .proc_write = proc_limit_write, + .proc_lseek = seq_lseek, + .proc_release = single_release, +}; + +static int __init dpages_proc_init(void) +{ + struct proc_dir_entry *dirty_dir; + + dirty_dir = proc_mkdir(DIRTY_ROOT, NULL); + if (!dirty_dir) + return -ENOMEM; + + if (!proc_create(DIRTY_PAGES, 0440, dirty_dir, &proc_dpages_operations)) + goto fail_pages; + + if (!proc_create(DIRTY_BUFFER_SIZE, 0640, dirty_dir, + &proc_buffer_operations)) + goto fail_switch; + + if (!proc_create(DIRTY_LIMIT, 0640, dirty_dir, &proc_limit_operations)) + goto fail_limit; + + return 0; + +fail_limit: + remove_proc_entry(DIRTY_BUFFER_SIZE, dirty_dir); +fail_switch: + remove_proc_entry(DIRTY_PAGES, dirty_dir); +fail_pages: + remove_proc_entry(DIRTY_ROOT, NULL); + return -ENOMEM; +} + +subsys_initcall(dpages_proc_init); -- 2.39.2

1 0

[PATCH OLK-6.6] fs/dirty_pages: dump the number of dirty pages for each inode
by Zizhi Wo 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OHM4?from=project-issue CVE: NA ---------------------------------------- In order to analyse the IO performance when using buffer IO, it's useful to obtain the number of dirty pages for each inode in the filesystem. This feature depends on 'CONFIG_DIRTY_PAGES'. It creates 3 interfaces by using procfs. /proc/dirty/buffer_size for buffer allocation and release, /proc/dirty/page_threshold to filter result and /proc/dirty/dirty_list to get dirty pages. Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/Kconfig | 13 + fs/Makefile | 1 + fs/dirty_pages.c | 350 +++++++++++++++++++++++++ 5 files changed, 366 insertions(+) create mode 100644 fs/dirty_pages.c diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 66737e8c8673..677b4cc4668a 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -6924,6 +6924,7 @@ CONFIG_PROC_PAGE_MONITOR=y CONFIG_PROC_CHILDREN=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index fcbef6c06587..c16cdb7b2596 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -8120,6 +8120,7 @@ CONFIG_PROC_PID_ARCH_STATUS=y CONFIG_PROC_CPU_RESCTRL=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/fs/Kconfig b/fs/Kconfig index 7c9d45ba9b0f..6dc83223c61b 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -170,6 +170,19 @@ source "fs/proc/Kconfig" source "fs/kernfs/Kconfig" source "fs/sysfs/Kconfig" +config DIRTY_PAGES + bool "Dumps the number of dirty pages of each file" + depends on PROC_FS + default y + help + This config supports the rendering of dirty page data to the user, + which may be useful to analyze the IO performance when using buffer + IO. + + It create 3 interfaces by using procfs. /proc/dirty/buffer_size for + buffer allocation and release; /proc/dirty/page_threshold to filter + result; /proc/dirty/dirty_list to get dirty pages. + config TMPFS bool "Tmpfs virtual memory file system support (former shm fs)" depends on SHMEM diff --git a/fs/Makefile b/fs/Makefile index bb6164200a39..cc4735c23c32 100644 --- a/fs/Makefile +++ b/fs/Makefile @@ -44,6 +44,7 @@ obj-$(CONFIG_FS_POSIX_ACL) += posix_acl.o obj-$(CONFIG_NFS_COMMON) += nfs_common/ obj-$(CONFIG_COREDUMP) += coredump.o obj-$(CONFIG_SYSCTL) += drop_caches.o sysctls.o +obj-$(CONFIG_DIRTY_PAGES) += dirty_pages.o obj-$(CONFIG_FHANDLE) += fhandle.o obj-$(CONFIG_CGROUP_FILES) += filescontrol.o diff --git a/fs/dirty_pages.c b/fs/dirty_pages.c new file mode 100644 index 000000000000..209a76a6aa46 --- /dev/null +++ b/fs/dirty_pages.c @@ -0,0 +1,350 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/proc_fs.h> +#include <linux/seq_file.h> +#include <linux/uaccess.h> +#include <linux/pagemap.h> +#include <linux/pagevec.h> +#include <linux/fs.h> +#include <linux/mm.h> +#include <linux/slab.h> +#include <linux/sched.h> +#include <linux/proc_fs.h> +#include <linux/kdev_t.h> +#include <linux/vmalloc.h> +#include <linux/module.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include "internal.h" + +#define DIRTY_ROOT "dirty" /* proc root directory name */ +#define DIRTY_BUFFER_SIZE "buffer_size" /* proc file to control buffer */ +#define DIRTY_PAGES "dirty_list" /* proc file to obtain diry pages */ +#define DIRTY_LIMIT "page_threshold" /* proc file to filter result */ +#define MAX_BUF_NUM 102400 /* max buffer num in number of pages */ +#define TERM_STR "terminated\n" /* tail string for overflowed buffer */ +#define UNKNOW_STR "unknown" /* for unknown file name or fs type */ + +static char *buf_dirty; /* buffer to store number of dirty pages */ +static unsigned long buf_size; /* size of buffer in bytes */ +static unsigned long buf_num; /* size of buffer in number of pages */ +static unsigned long buf_limit; /* filter threshold of dirty pages */ +unsigned long lock_word; /* for exclusive access to buffer */ + +static bool dirty_pages_lock(void) +{ + if (xchg(&lock_word, 1) == 1) + return false; + return true; +} + +static void dirty_pages_unlock(void) +{ + lock_word = 0; +} + +static unsigned long dump_dirtypages_inode(struct inode *inode) +{ + XA_STATE(xas, &inode->i_mapping->i_pages, 0); + unsigned long pages = 0; + void *page; + + if (!mapping_tagged(inode->i_mapping, PAGECACHE_TAG_DIRTY)) + return 0; + + xas_lock_irq(&xas); + xas_for_each_marked(&xas, page, (pgoff_t)-1, PAGECACHE_TAG_DIRTY) { + if (++pages % XA_CHECK_SCHED) + continue; + + xas_pause(&xas); + xas_unlock_irq(&xas); + cond_resched(); + xas_lock_irq(&xas); + } + xas_unlock_irq(&xas); + + return pages; +} + +static char *inode_filename(struct inode *inode) +{ + char *path; + static char buf[PATH_MAX]; + struct dentry *dentry = d_find_alias(inode); + + if (!dentry) + return UNKNOW_STR; + + path = dentry_path_raw(dentry, buf, PATH_MAX); + dput(dentry); + + return IS_ERR_OR_NULL(path) ? UNKNOW_STR : path; +} + +static const char *sb_fstype(struct super_block *sb) +{ + if (sb->s_type && sb->s_type->name) + return sb->s_type->name; + else + return UNKNOW_STR; +} + +static inline bool is_sb_writable(struct super_block *sb) +{ + if (sb_rdonly(sb)) + return false; + + if (sb->s_writers.frozen == SB_FREEZE_COMPLETE) + return false; + + return true; +} + +/* + * dump_dirtypages_sb - dump the dirty pages of each inode in the sb + * @sb the super block + * @m the seq_file witch is initialized in proc_dpages_open + * + * For each inode in the sb, call dump_dirtypages_pages to get the number + * of dirty pages. And use seq_printf to store the result in the buffer + * if it's not less than the threshold. The inode in unusual state will + * be skipped. + */ +static void dump_dirtypages_sb(struct super_block *sb, struct seq_file *m) +{ + struct inode *inode, *toput_inode = NULL; + unsigned long limit = READ_ONCE(buf_limit); + + if (!is_sb_writable(sb) || !sb->s_bdev) + return; + + spin_lock(&sb->s_inode_list_lock); + list_for_each_entry(inode, &sb->s_inodes, i_sb_list) { + unsigned long dirty_pages; + + spin_lock(&inode->i_lock); + if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) { + spin_unlock(&inode->i_lock); + continue; + } + + __iget(inode); + spin_unlock(&inode->i_lock); + spin_unlock(&sb->s_inode_list_lock); + + dirty_pages = dump_dirtypages_inode(inode); + if (!dirty_pages || dirty_pages < limit) + goto skip; + + seq_printf(m, "FSType: %s, Dev ID: %u(%u:%u) ino %lu, dirty pages %lu, path %s\n", + sb_fstype(sb), sb->s_dev, MAJOR(sb->s_dev), + MINOR(sb->s_dev), inode->i_ino, dirty_pages, + inode_filename(inode)); + + if (seq_has_overflowed(m)) { + /* keep size > count to avoid overflow in seq_read_iter() */ + m->size += strlen(TERM_STR) + 1; + seq_puts(m, TERM_STR); + iput(inode); + iput(toput_inode); + return; + } +skip: + iput(toput_inode); + toput_inode = inode; + cond_resched(); + spin_lock(&sb->s_inode_list_lock); + } + spin_unlock(&sb->s_inode_list_lock); + iput(toput_inode); +} + +static int proc_dpages_show(struct seq_file *m, void *v) +{ + iterate_supers((void *)dump_dirtypages_sb, (void *)m); + return 0; +} + +static int proc_dpages_open(struct inode *inode, struct file *filp) +{ + int ret; + struct seq_file *m; + + if (!dirty_pages_lock()) + return -EBUSY; + + if (buf_dirty == NULL || buf_size == 0) { + pr_warn("please allocate buffer before getting dirty pages\n"); + dirty_pages_unlock(); + return -ENOMEM; + } + + ret = single_open(filp, proc_dpages_show, NULL); + if (ret) { + dirty_pages_unlock(); + return ret; + } + + m = (struct seq_file *)filp->private_data; + memset(buf_dirty, 0, buf_size); + m->size = buf_size - (strlen(TERM_STR) + 1); + m->buf = buf_dirty; + + return 0; +} + +static int proc_dpages_release(struct inode *inode, struct file *file) +{ + struct seq_file *m = file->private_data; + + /* we don't want to free the buf */ + m->buf = NULL; + dirty_pages_unlock(); + return single_release(inode, file); +} + +static void free_buf_dirty(void) +{ + if (!buf_dirty) + return; + + vfree(buf_dirty); + buf_dirty = NULL; + buf_size = 0; + WRITE_ONCE(buf_num, 0); +} + +static int proc_buffer_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", READ_ONCE(buf_num)); + return 0; +} + +static int proc_buffer_open(struct inode *inode, struct file *filp) +{ + int ret; + + if (((filp->f_flags & O_ACCMODE) != O_RDONLY) && !dirty_pages_lock()) + return -EBUSY; + + ret = single_open(filp, proc_buffer_show, NULL); + if (ret) + dirty_pages_unlock(); + + return ret; +} + +static ssize_t proc_buffer_write(struct file *filp, const char *buf, size_t count, + loff_t *offp) +{ + unsigned long new_buf_num, new_buf_size; + char *new_buf_dirty; + int ret = kstrtoul_from_user(buf, count, 10, &new_buf_num); + + if (ret != 0 || new_buf_num > MAX_BUF_NUM) + return -EINVAL; + + if (new_buf_num == buf_num) + return count; + + if (new_buf_num == 0) { + free_buf_dirty(); + return count; + } + + new_buf_size = PAGE_SIZE * new_buf_num; + new_buf_dirty = vzalloc(new_buf_size); + if (!new_buf_dirty) + return -ENOMEM; + + free_buf_dirty(); + buf_dirty = new_buf_dirty; + buf_size = new_buf_size; + WRITE_ONCE(buf_num, new_buf_num); + + return count; +} + +static int proc_buffer_release(struct inode *inode, struct file *filp) +{ + if ((filp->f_flags & O_ACCMODE) != O_RDONLY) + dirty_pages_unlock(); + return single_release(inode, filp); +} + +static int proc_limit_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", READ_ONCE(buf_limit)); + return 0; +} + +static int proc_limit_open(struct inode *inode, struct file *filp) +{ + return single_open(filp, proc_limit_show, NULL); +} + +static ssize_t proc_limit_write(struct file *filp, const char *buf, size_t count, + loff_t *offp) +{ + unsigned long limit; + int ret = kstrtoul_from_user(buf, count, 10, &limit); + + if (ret != 0) + return ret; + + WRITE_ONCE(buf_limit, limit); + + return count; +} + +static const struct proc_ops proc_dpages_operations = { + .proc_open = proc_dpages_open, + .proc_read = seq_read, + .proc_release = proc_dpages_release, +}; + +static const struct proc_ops proc_buffer_operations = { + .proc_open = proc_buffer_open, + .proc_read = seq_read, + .proc_write = proc_buffer_write, + .proc_lseek = seq_lseek, + .proc_release = proc_buffer_release, +}; + +static const struct proc_ops proc_limit_operations = { + .proc_open = proc_limit_open, + .proc_read = seq_read, + .proc_write = proc_limit_write, + .proc_lseek = seq_lseek, + .proc_release = single_release, +}; + +static int __init dpages_proc_init(void) +{ + struct proc_dir_entry *dirty_dir = proc_mkdir(DIRTY_ROOT, NULL); + + if (!dirty_dir) + return -ENOMEM; + + if (!proc_create(DIRTY_PAGES, 0440, dirty_dir, &proc_dpages_operations)) + goto fail_pages; + + if (!proc_create(DIRTY_BUFFER_SIZE, 0640, dirty_dir, + &proc_buffer_operations)) + goto fail_switch; + + if (!proc_create(DIRTY_LIMIT, 0640, dirty_dir, &proc_limit_operations)) + goto fail_limit; + + return 0; + +fail_limit: + remove_proc_entry(DIRTY_BUFFER_SIZE, dirty_dir); +fail_switch: + remove_proc_entry(DIRTY_PAGES, dirty_dir); +fail_pages: + remove_proc_entry(DIRTY_ROOT, NULL); + return -ENOMEM; +} + +subsys_initcall(dpages_proc_init); -- 2.39.2

2 1

[PATCH OLK-6.6 V1 00/12] infrastructure for scheduler bpf
by Guan Jing 09 Jan '24

09 Jan '24

infrastructure for scheduler bpf Guan Jing (12): sched: programmable: Introduce bpf sched sched: programmable: Add a tag for the task sched: programmable: Add a tag for the task group sched: programmable: Add user interface of task group tag sched: programmable: Add user interface of task tag sched: basic infrastructure for scheduler bpf sched: introduce bpf_sched_enable() libbpf: add support for scheduler bpf programs bpftool: recognize scheduler programs sched: Add helper functions to get cpu statistics sched: programmable: Add hook in select_task_rq_fair() sched: programmable: Add hook in can_migrate_task() arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/proc/base.c | 64 ++++++++++++++ include/linux/bpf_sched.h | 50 +++++++++++ include/linux/bpf_types.h | 4 + include/linux/sched.h | 38 ++++++++ include/linux/sched_hook_defs.h | 4 + include/uapi/linux/bpf.h | 9 ++ init/init_task.c | 3 + kernel/bpf/Kconfig | 12 +++ kernel/bpf/btf.c | 1 + kernel/bpf/syscall.c | 27 ++++++ kernel/bpf/trampoline.c | 1 + kernel/bpf/verifier.c | 11 ++- kernel/sched/bpf_sched.c | 102 +++++++++++++++++++++ kernel/sched/build_utility.c | 4 + kernel/sched/core.c | 118 +++++++++++++++++++++++++ kernel/sched/fair.c | 45 ++++++++++ kernel/sched/sched.h | 10 +++ scripts/bpf_doc.py | 4 + tools/include/uapi/linux/bpf.h | 9 ++ tools/lib/bpf/bpf.c | 1 + tools/lib/bpf/libbpf.c | 23 ++++- tools/lib/bpf/libbpf.h | 2 + tools/lib/bpf/libbpf.map | 1 + 25 files changed, 543 insertions(+), 2 deletions(-) create mode 100644 include/linux/bpf_sched.h create mode 100644 include/linux/sched_hook_defs.h create mode 100644 kernel/sched/bpf_sched.c -- 2.34.1

1 12

[PATCH OLK-6.6] x86: config: disable network config
by Zhengchao Shao 09 Jan '24

09 Jan '24

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8UM53 CVE: NA -------------------------------- Disable X86 network config. Signed-off-by: Zhengchao Shao <shaozhengchao(a)huawei.com> --- arch/x86/configs/openeuler_defconfig | 34 ++++++++++++++-------------- 1 file changed, 17 insertions(+), 17 deletions(-) diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index 1ef398404878..dd7a8a3bca0f 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -1860,7 +1860,7 @@ CONFIG_BT_CMTP=m CONFIG_BT_HIDP=m CONFIG_BT_HS=y CONFIG_BT_LE=y -CONFIG_BT_LE_L2CAP_ECRED=y +# CONFIG_BT_LE_L2CAP_ECRED is not set # CONFIG_BT_6LOWPAN is not set # CONFIG_BT_LEDS is not set # CONFIG_BT_MSFTEXT is not set @@ -1876,7 +1876,7 @@ CONFIG_BT_BCM=m CONFIG_BT_RTL=m CONFIG_BT_HCIBTUSB=m CONFIG_BT_HCIBTUSB_AUTOSUSPEND=y -CONFIG_BT_HCIBTUSB_POLL_SYNC=y +# CONFIG_BT_HCIBTUSB_POLL_SYNC is not set CONFIG_BT_HCIBTUSB_BCM=y # CONFIG_BT_HCIBTUSB_MTK is not set CONFIG_BT_HCIBTUSB_RTL=y @@ -2788,7 +2788,7 @@ CONFIG_AMD_XGBE_HAVE_ECC=y CONFIG_NET_VENDOR_AQUANTIA=y CONFIG_AQTION=m # CONFIG_NET_VENDOR_ARC is not set -CONFIG_NET_VENDOR_ASIX=y +# CONFIG_NET_VENDOR_ASIX is not set # CONFIG_SPI_AX88796C is not set CONFIG_NET_VENDOR_ATHEROS=y CONFIG_ATL2=m @@ -2836,7 +2836,7 @@ CONFIG_CHELSIO_IPSEC_INLINE=m CONFIG_NET_VENDOR_CISCO=y CONFIG_ENIC=m # CONFIG_NET_VENDOR_CORTINA is not set -CONFIG_NET_VENDOR_DAVICOM=y +# CONFIG_NET_VENDOR_DAVICOM is not set # CONFIG_DM9051 is not set CONFIG_DNET=m CONFIG_NET_VENDOR_DEC=y @@ -2851,10 +2851,10 @@ CONFIG_BE2NET_HWMON=y # CONFIG_BE2NET_BE3 is not set CONFIG_BE2NET_LANCER=y CONFIG_BE2NET_SKYHAWK=y -CONFIG_NET_VENDOR_ENGLEDER=y +# CONFIG_NET_VENDOR_ENGLEDER is not set # CONFIG_TSNEP is not set # CONFIG_NET_VENDOR_EZCHIP is not set -CONFIG_NET_VENDOR_FUNGIBLE=y +# CONFIG_NET_VENDOR_FUNGIBLE is not set # CONFIG_FUN_ETH is not set CONFIG_NET_VENDOR_GOOGLE=y # CONFIG_GVE is not set @@ -2882,14 +2882,14 @@ CONFIG_I40E_DCB=y CONFIG_IAVF=m CONFIG_I40EVF=m CONFIG_ICE=m -CONFIG_ICE_SWITCHDEV=y -CONFIG_ICE_HWTS=y +# CONFIG_ICE_SWITCHDEV is not set +#CONFIG_ICE_HWTS is not set CONFIG_FM10K=m # CONFIG_IGC is not set # CONFIG_JME is not set -CONFIG_NET_VENDOR_ADI=y +# CONFIG_NET_VENDOR_ADI is not set # CONFIG_ADIN1110 is not set -CONFIG_NET_VENDOR_LITEX=y +# CONFIG_NET_VENDOR_LITEX is not set # CONFIG_NET_VENDOR_MARVELL is not set CONFIG_NET_VENDOR_MELLANOX=y CONFIG_MLX4_EN=m @@ -2924,7 +2924,7 @@ CONFIG_MLXFW=m # CONFIG_NET_VENDOR_MICREL is not set # CONFIG_NET_VENDOR_MICROCHIP is not set # CONFIG_NET_VENDOR_MICROSEMI is not set -CONFIG_NET_VENDOR_MICROSOFT=y +# CONFIG_NET_VENDOR_MICROSOFT is not set # CONFIG_MICROSOFT_MANA is not set CONFIG_NET_VENDOR_MYRI=y CONFIG_MYRI10GE=m @@ -2938,7 +2938,7 @@ CONFIG_NET_VENDOR_NETRONOME=y CONFIG_NFP=m CONFIG_NFP_APP_FLOWER=y CONFIG_NFP_APP_ABM_NIC=y -CONFIG_NFP_NET_IPSEC=y +# CONFIG_NFP_NET_IPSEC is not set # CONFIG_NFP_DEBUG is not set # CONFIG_NET_VENDOR_NVIDIA is not set CONFIG_NET_VENDOR_OKI=y @@ -2995,10 +2995,10 @@ CONFIG_SFC_MCDI_LOGGING=y # CONFIG_NET_VENDOR_SYNOPSYS is not set # CONFIG_NET_VENDOR_TEHUTI is not set # CONFIG_NET_VENDOR_TI is not set -CONFIG_NET_VENDOR_VERTEXCOM=y +# CONFIG_NET_VENDOR_VERTEXCOM is not set # CONFIG_MSE102X is not set # CONFIG_NET_VENDOR_VIA is not set -CONFIG_NET_VENDOR_WANGXUN=y +# CONFIG_NET_VENDOR_WANGXUN is not set CONFIG_LIBWX=m CONFIG_NGBE=m CONFIG_TXGBE=m @@ -3325,7 +3325,7 @@ CONFIG_MT76x2U=m CONFIG_WLAN_VENDOR_MICROCHIP=y # CONFIG_WILC1000_SDIO is not set # CONFIG_WILC1000_SPI is not set -CONFIG_WLAN_VENDOR_PURELIFI=y +# CONFIG_WLAN_VENDOR_PURELIFI is not set # CONFIG_PLFXLC is not set CONFIG_WLAN_VENDOR_RALINK=y CONFIG_RT2X00=m @@ -3382,7 +3382,7 @@ CONFIG_RTL8XXXU=m # CONFIG_RTW88 is not set # CONFIG_RTW89 is not set # CONFIG_WLAN_VENDOR_RSI is not set -CONFIG_WLAN_VENDOR_SILABS=y +# CONFIG_WLAN_VENDOR_SILABS is not set # CONFIG_WFX is not set # CONFIG_WLAN_VENDOR_ST is not set # CONFIG_WLAN_VENDOR_TI is not set @@ -8225,7 +8225,7 @@ CONFIG_SUNRPC=m CONFIG_SUNRPC_GSS=m CONFIG_SUNRPC_BACKCHANNEL=y CONFIG_RPCSEC_GSS_KRB5=m -CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_AES_SHA1=y +# CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_AES_SHA1 is not set # CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_CAMELLIA is not set # CONFIG_RPCSEC_GSS_KRB5_ENCTYPES_AES_SHA2 is not set CONFIG_SUNRPC_DEBUG=y -- 2.34.1

1 0

[PATCH OLK-6.6 0/1] arm64: Add non nmi ipi backtrace support
by Liao Chen 09 Jan '24

09 Jan '24

From: linruizhe <linruizhe(a)huawei.com> Dependencies: CONFIG_DETECT_HUNG_TASK=y CONFIG_WATCHDOG=y Li Zhengyu (1): arm64: Add non nmi ipi backtrace support arch/arm64/Kconfig | 8 +++++ arch/arm64/configs/openeuler_defconfig | 1 + arch/arm64/kernel/ipi_nmi.c | 41 ++++++++++++++++++++++---- 3 files changed, 45 insertions(+), 5 deletions(-) -- 2.25.1

1 1

[PATCH OLK-6.6] pid_ns: Make pid_max per namespace
by Yi Yang 09 Jan '24

09 Jan '24

From: Li Zefan <lizefan(a)huawei.com> euler inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8TCPY ------------------------------------------------- The adjusted pid_max on the host also takes effect in the container. Move pid_max to pid_namespace to fix this problem. Signed-off-by: Li Zefan <lizefan(a)huawei.com> Signed-off-by: Yi Yang <yiyang13(a)huawei.com> --- include/linux/pid.h | 2 ++ include/linux/pid_namespace.h | 3 ++ init/Kconfig | 6 ++++ kernel/pid.c | 57 +++++++++++++++++++++++++++++++++++ kernel/pid_namespace.c | 13 ++++++++ kernel/sysctl.c | 2 ++ kernel/trace/pid_list.c | 4 +++ kernel/trace/trace.c | 4 +++ kernel/trace/trace.h | 2 ++ 9 files changed, 93 insertions(+) diff --git a/include/linux/pid.h b/include/linux/pid.h index 653a527574c4..b90bc447d2a2 100644 --- a/include/linux/pid.h +++ b/include/linux/pid.h @@ -113,8 +113,10 @@ extern void transfer_pid(struct task_struct *old, struct task_struct *new, struct pid_namespace; extern struct pid_namespace init_pid_ns; +#ifndef CONFIG_PID_MAX_PER_NAMESPACE extern int pid_max; extern int pid_max_min, pid_max_max; +#endif /* * look up a PID in the hash table. Must be called with the tasklist_lock diff --git a/include/linux/pid_namespace.h b/include/linux/pid_namespace.h index f9f9931e02d6..0a7c6dab0b7b 100644 --- a/include/linux/pid_namespace.h +++ b/include/linux/pid_namespace.h @@ -36,6 +36,9 @@ struct pid_namespace { #endif struct user_namespace *user_ns; struct ucounts *ucounts; +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + int pid_max; +#endif int reboot; /* group exit code if this pidns was rebooted */ struct ns_common ns; #if defined(CONFIG_SYSCTL) && defined(CONFIG_MEMFD_CREATE) diff --git a/init/Kconfig b/init/Kconfig index 869eea4108d0..8c07d158bb54 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -2104,3 +2104,9 @@ config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE # <asm/syscall_wrapper.h>. config ARCH_HAS_SYSCALL_WRAPPER def_bool n + +config PID_MAX_PER_NAMESPACE + bool "Make pid_max per namespace" + default y + help + Say Y here to enable make pid_max per namespace. diff --git a/kernel/pid.c b/kernel/pid.c index 383abde0c208..4e9267988dd7 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -45,6 +45,10 @@ #include <net/sock.h> #include <uapi/linux/pidfd.h> +#ifdef CONFIG_PID_MAX_PER_NAMESPACE +#include <linux/kmemleak.h> +#endif + struct pid init_struct_pid = { .count = REFCOUNT_INIT(1), .tasks = { @@ -59,12 +63,19 @@ struct pid init_struct_pid = { }, } }; +#ifndef CONFIG_PID_MAX_PER_NAMESPACE int pid_max = PID_MAX_DEFAULT; +#endif #define RESERVED_PIDS 300 +#ifndef CONFIG_PID_MAX_PER_NAMESPACE int pid_max_min = RESERVED_PIDS + 1; int pid_max_max = PID_MAX_LIMIT; +#else +static int pid_max_min = RESERVED_PIDS + 1; +static int pid_max_max = PID_MAX_LIMIT; +#endif /* * PID-map pages start out as NULL, they get allocated upon @@ -80,6 +91,9 @@ struct pid_namespace init_pid_ns = { .child_reaper = &init_task, .user_ns = &init_user_ns, .ns.inum = PROC_PID_INIT_INO, +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + .pid_max = PID_MAX_DEFAULT, +#endif #ifdef CONFIG_PID_NS .ns.ops = &pidns_operations, #endif @@ -194,7 +208,11 @@ struct pid *alloc_pid(struct pid_namespace *ns, pid_t *set_tid, tid = set_tid[ns->level - i]; retval = -EINVAL; +#ifndef CONFIG_PID_MAX_PER_NAMESPACE if (tid < 1 || tid >= pid_max) +#else + if (tid < 1 || tid >= task_active_pid_ns(current)->pid_max) +#endif goto out_free; /* * Also fail if a PID != 1 is requested and @@ -234,7 +252,11 @@ struct pid *alloc_pid(struct pid_namespace *ns, pid_t *set_tid, * a partially initialized PID (see below). */ nr = idr_alloc_cyclic(&tmp->idr, NULL, pid_min, +#ifndef CONFIG_PID_MAX_PER_NAMESPACE pid_max, GFP_ATOMIC); +#else + tmp->pid_max, GFP_ATOMIC); +#endif } spin_unlock_irq(&pidmap_lock); idr_preload_end(); @@ -646,8 +668,37 @@ SYSCALL_DEFINE2(pidfd_open, pid_t, pid, unsigned int, flags) return fd; } +#ifdef CONFIG_PID_MAX_PER_NAMESPACE +static int proc_dointvec_pidmax(struct ctl_table *table, int write, + void __user *buffer, size_t *lenp, loff_t *ppos) +{ + struct ctl_table tmp; + + tmp = *table; + tmp.data = &task_active_pid_ns(current)->pid_max; + + return proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos); +} + +static struct ctl_table pid_ctl_table[] = { + { + .procname = "pid_max", + .data = &init_pid_ns.pid_max, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = proc_dointvec_pidmax, + .extra1 = &pid_max_min, + .extra2 = &pid_max_max, + }, + {} +}; +#endif + void __init pid_idr_init(void) { +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + int pid_max = init_pid_ns.pid_max; +#endif /* Verify no one has done anything silly: */ BUILD_BUG_ON(PID_MAX_LIMIT >= PIDNS_ADDING); @@ -658,6 +709,9 @@ void __init pid_idr_init(void) PIDS_PER_CPU_MIN * num_possible_cpus()); pr_info("pid_max: default: %u minimum: %u\n", pid_max, pid_max_min); +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + init_pid_ns.pid_max = pid_max; +#endif idr_init(&init_pid_ns.idr); init_pid_ns.pid_cachep = kmem_cache_create("pid", @@ -665,6 +719,9 @@ void __init pid_idr_init(void) __alignof__(struct pid), SLAB_HWCACHE_ALIGN | SLAB_PANIC | SLAB_ACCOUNT, NULL); +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + register_sysctl_init("kernel", pid_ctl_table); +#endif } static struct file *__pidfd_fget(struct task_struct *task, int fd) diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index 619972c78774..c837b1096dcb 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -110,6 +110,9 @@ static struct pid_namespace *create_pid_namespace(struct user_namespace *user_ns ns->user_ns = get_user_ns(user_ns); ns->ucounts = ucounts; ns->pid_allocated = PIDNS_ADDING; +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + ns->pid_max = parent_pid_ns->pid_max; +#endif #if defined(CONFIG_SYSCTL) && defined(CONFIG_MEMFD_CREATE) ns->memfd_noexec_scope = pidns_memfd_noexec_scope(parent_pid_ns); #endif @@ -295,6 +298,10 @@ static int pid_ns_ctl_handler(struct ctl_table *table, int write, next = idr_get_cursor(&pid_ns->idr) - 1; tmp.data = &next; +#ifdef CONFIG_PID_MAX_PER_NAMESPACE + tmp.extra2 = &pid_ns->pid_max; +#endif + ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos); if (!ret && write) idr_set_cursor(&pid_ns->idr, next + 1); @@ -302,7 +309,9 @@ static int pid_ns_ctl_handler(struct ctl_table *table, int write, return ret; } +#ifndef CONFIG_PID_MAX_PER_NAMESPACE extern int pid_max; +#endif static struct ctl_table pid_ns_ctl_table[] = { { .procname = "ns_last_pid", @@ -310,7 +319,11 @@ static struct ctl_table pid_ns_ctl_table[] = { .mode = 0666, /* permissions are checked in the handler */ .proc_handler = pid_ns_ctl_handler, .extra1 = SYSCTL_ZERO, +#ifndef CONFIG_PID_MAX_PER_NAMESPACE .extra2 = &pid_max, +#else + .extra2 = &init_pid_ns.pid_max, +#endif }, { } }; diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 354a2d294f52..e84df0818517 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1809,6 +1809,7 @@ static struct ctl_table kern_table[] = { .proc_handler = proc_dointvec, }, #endif +#ifndef CONFIG_PID_MAX_PER_NAMESPACE { .procname = "pid_max", .data = &pid_max, @@ -1818,6 +1819,7 @@ static struct ctl_table kern_table[] = { .extra1 = &pid_max_min, .extra2 = &pid_max_max, }, +#endif { .procname = "panic_on_oops", .data = &panic_on_oops, diff --git a/kernel/trace/pid_list.c b/kernel/trace/pid_list.c index 95106d02b32d..8056d7441d05 100644 --- a/kernel/trace/pid_list.c +++ b/kernel/trace/pid_list.c @@ -414,7 +414,11 @@ struct trace_pid_list *trace_pid_list_alloc(void) int i; /* According to linux/thread.h, pids can be no bigger that 30 bits */ +#ifndef CONFIG_PID_MAX_PER_NAMESPACE WARN_ON_ONCE(pid_max > (1 << 30)); +#else + WARN_ON_ONCE(task_active_pid_ns(current)->pid_max > (1 << 30)); +#endif pid_list = kzalloc(sizeof(*pid_list), GFP_KERNEL); if (!pid_list) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index b656cab67f67..5afa58302b06 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -5423,7 +5423,11 @@ int set_tracer_flag(struct trace_array *tr, unsigned int mask, int enabled) if (mask == TRACE_ITER_RECORD_TGID) { if (!tgid_map) { +#ifndef CONFIG_PID_MAX_PER_NAMESPACE tgid_map_max = pid_max; +#else + tgid_map_max = init_pid_ns.pid_max; +#endif map = kvcalloc(tgid_map_max + 1, sizeof(*tgid_map), GFP_KERNEL); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index d608f6128704..52123df262b2 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -695,7 +695,9 @@ extern unsigned long tracing_thresh; /* PID filtering */ +#ifndef CONFIG_PID_MAX_PER_NAMESPACE extern int pid_max; +#endif bool trace_find_filtered_pid(struct trace_pid_list *filtered_pids, pid_t search_pid); -- 2.25.1

1 0

2024

2023

2022

2021

2020

2019

Kernel