[PATCH OLK-6.6 0/2] xSched dmem
xSched dmem Liu Kai (2): xSched/cgroup: prevent to init xse when xcu cgroup is disabled xSched: remove redundant lookup in alloc_ctx_from_vstream to enforce allocation semantics include/linux/xsched.h | 2 ++ kernel/xsched/cgroup.c | 2 +- kernel/xsched/core.c | 3 +++ kernel/xsched/vstream.c | 4 ---- 4 files changed, 6 insertions(+), 5 deletions(-) -- 2.34.1
hulk inclusion category: bugfix bugzilla: https://atomgit.com/openeuler/kernel/issues/8424 ---------------------------------------- When the XCU subsystem is not enabled, the root node of the XCU subsystem defaults to struct freezer, and the members linked list of xsched_group remains uninitialized. Attempting to add nodes to this uninitialized list triggers a kernel list corruption error: [ 3376.708041] list_add corruption. prev is NULL. [ 3376.708372] Call trace: [ 3376.708375] __list_add_valid_or_report+0x58/0xe8 [ 3376.708378] xsched_group_xse_attach+0x68/0xe8 [ 3376.708393] xsched_group_inherit+0x38/0x170 [ 3376.708397] init_xsched_entity+0x50/0x1a0 [ 3376.708400] alloc_ctx_from_vstream+0x10c/0x1c0 [ 3376.708403] vstream_bind_to_ctx+0x16c/0x1a8 [ 3376.708406] sqcq_alloc+0x12c/0x380 [ 3376.708409] vstream_alloc+0x20/0x48 [ 3376.708412] __se_sys_vstream_manage+0xa0/0x1e8 [ 3376.708414] __arm64_sys_vstream_manage+0x1c/0x30 [ 3376.708417] invoke_syscall+0x50/0x128 [ 3376.708422] el0_svc_common.constprop.0+0xc8/0xf0 [ 3376.708424] do_el0_svc+0x48/0x78 [ 3376.708426] el0_slow_syscall+0x44/0x1e8 [ 3376.708433] el0t_64_sync_handler+0xc0/0xc8 [ 3376.708435] el0t_64_sync+0x188/0x190 [ 3376.708439] ---[ end trace 0000000000000000 ]--- This error occurs because the code attempts to perform list operations on an uninitialized list head, violating kernel list integrity checks. To resolve this, add a check for xcu_cgroup_enabled() at the start of init_xsched_entity. If the XCU subsystem is disabled, return -EPERM immediately to block xsched_entity initialization, avoiding invalid list operations and ensuring system stability. Fixes: 0ca67081a163 ("xSched/cgroup: reuse SUBSYS for xcu and freezer to preserve KABI") Signed-off-by: Liu Kai <liukai284@huawei.com> --- include/linux/xsched.h | 2 ++ kernel/xsched/cgroup.c | 2 +- kernel/xsched/core.c | 3 +++ 3 files changed, 6 insertions(+), 1 deletion(-) diff --git a/include/linux/xsched.h b/include/linux/xsched.h index ff314f88169f..e9c580149d55 100644 --- a/include/linux/xsched.h +++ b/include/linux/xsched.h @@ -485,6 +485,8 @@ void xsched_quota_account(struct xsched_group *xg, s64 exec_time); void xsched_quota_check(struct xsched_group *xg, struct xsched_cu *xcu); void xsched_quota_refill(struct work_struct *work); +bool xcu_cgroup_enabled(void); + #define XCU_PERIOD_MIN_MS 1 #define XCU_QUOTA_RUNTIME_INF -1 #define XCU_SHARES_MIN 1 diff --git a/kernel/xsched/cgroup.c b/kernel/xsched/cgroup.c index d22ffc543d45..8d22aa78ea81 100644 --- a/kernel/xsched/cgroup.c +++ b/kernel/xsched/cgroup.c @@ -67,7 +67,7 @@ static int __init xcu_setup(char *str) } __setup("xcu=", xcu_setup); -static bool xcu_cgroup_enabled(void) +bool xcu_cgroup_enabled(void) { return static_branch_unlikely(&xcu_cgroup_key); } diff --git a/kernel/xsched/core.c b/kernel/xsched/core.c index 1d835fc44d0c..5e3e4e02dfac 100644 --- a/kernel/xsched/core.c +++ b/kernel/xsched/core.c @@ -463,6 +463,9 @@ int init_xsched_entity(struct xsched_context *ctx, struct vstream_info *vs) xse = &ctx->xse; #ifdef CONFIG_CGROUP_XCU + if (!xcu_cgroup_enabled()) + return -EPERM; + xsched_group_inherit(current, xse); /* inherit the scheduler class from the parent group */ class_id = xse->parent_grp->sched_class; -- 2.34.1
hulk inclusion category: cleanup bugzilla: https://atomgit.com/openeuler/kernel/issues/8424 ---------------------------------------- Refactor 'alloc_ctx_from_vstream' to strictly perform context allocation without attempting to find an existing one first. Previously, this function would check for an existing context by TGID and return early if found. However, the lookup logic is now the responsibility of the caller, who ensures this function is only invoked when a context is genuinely missing. Removing the internal lookup ensures the function behavior aligns with its name—pure allocation—and prevents redundant searches. Signed-off-by: Liu Kai <liukai284@huawei.com> --- kernel/xsched/vstream.c | 4 ---- 1 file changed, 4 deletions(-) diff --git a/kernel/xsched/vstream.c b/kernel/xsched/vstream.c index b39e97682dfb..46710d3ce1d0 100644 --- a/kernel/xsched/vstream.c +++ b/kernel/xsched/vstream.c @@ -223,10 +223,6 @@ static int alloc_ctx_from_vstream(struct vstream_info *vstream_info, struct xsched_cu *xcu = vstream_info->xcu; int ret; - *ctx = ctx_find_by_tgid_and_xcu(vstream_info->tgid, xcu); - if (*ctx) - return 0; - *ctx = kzalloc(sizeof(struct xsched_context), GFP_KERNEL); if (!*ctx) { XSCHED_ERR("Fail to alloc xsched context (tgid=%d) @ %s\n", -- 2.34.1
反馈： 您发送到kernel@openeuler.org的补丁/补丁集，已成功转换为PR！ PR链接地址： https://atomgit.com/openeuler/kernel/merge_requests/21746 邮件列表地址：https://mailweb.openeuler.org/archives/list/kernel@openeuler.org/message/SHT... FeedBack: The patch(es) which you have sent to kernel@openeuler.org mailing list has been converted to a pull request successfully! Pull request link: https://atomgit.com/openeuler/kernel/merge_requests/21746 Mailing list address: https://mailweb.openeuler.org/archives/list/kernel@openeuler.org/message/SHT...
participants (2)
-
Liu Kai -
patchwork bot