It seems that the pointer-to-kretprobe "rp" within the kretprobe_holder is RCU-managed, based on the (non-rethook) implementation of get_kretprobe(). The thought behind this patch is to make use of the RCU API where possible when accessing this pointer so that the needed barriers are always in place and to self-document the code. The __rcu annotation to "rp" allows for sparse RCU checking. Plain writes done to the "rp" pointer are changed to make use of the RCU macro for assignment. For the single read, the implementation of get_kretprobe() is simplified by making use of an RCU macro which accomplishes the same, but note that the log warning text will be more generic. I did find that there is a difference in assembly generated between the usage of the RCU macros vs without. For example, on arm64, when using rcu_assign_pointer(), the corresponding store instruction is a store-release (STLR) which has an implicit barrier. When normal assignment is done, a regular store (STR) is found. In the macro case, this seems to be a result of rcu_assign_pointer() using smp_store_release() when the value to write is not NULL. Link: https://lore.kernel.org/all/20231122132058.3359-1-inwardvessel@gmail.com/
Unified · Split
| @@ -139,7 +139,7 @@ static inline bool kprobe_ftrace(struct kprobe *p) · 138 unmodified lines | |||
| 139 | * | 139 | * |
| 140 | */ | 140 | */ |
| 141 | struct kretprobe_holder { | 141 | struct kretprobe_holder { |
| 142 | struct kretprobe *rp; | 142 | struct kretprobe __rcu *rp; |
| 143 | struct objpool_head pool; | 143 | struct objpool_head pool; |
| 144 | }; | 144 | }; |
| 145 | 145 | ||
| @@ -245,10 +245,7 @@ unsigned long kretprobe_trampoline_handler(struct pt_regs *regs, · 99 unmodified lines | |||
| 245 | 245 | ||
| 246 | static nokprobe_inline struct kretprobe *get_kretprobe(struct kretprobe_instance *ri) | 246 | static nokprobe_inline struct kretprobe *get_kretprobe(struct kretprobe_instance *ri) |
| 247 | { | 247 | { |
| 248 | RCU_LOCKDEP_WARN(!rcu_read_lock_any_held(), | 248 | return rcu_dereference_check(ri->rph->rp, rcu_read_lock_any_held()); |
| 249 | "Kretprobe is accessed from instance under preemptive context"); | ||
| 250 | |||
| 251 | return READ_ONCE(ri->rph->rp); | ||
| 252 | } | 249 | } |
| 253 | 250 | ||
| 254 | static nokprobe_inline unsigned long get_kretprobe_retaddr(struct kretprobe_instance *ri) | 251 | static nokprobe_inline unsigned long get_kretprobe_retaddr(struct kretprobe_instance *ri) |
| @@ -2252,7 +2252,7 @@ int register_kretprobe(struct kretprobe *rp) · 2251 unmodified lines | |||
| 2252 | rp->rph = NULL; | 2252 | rp->rph = NULL; |
| 2253 | return -ENOMEM; | 2253 | return -ENOMEM; |
| 2254 | } | 2254 | } |
| 2255 | rp->rph->rp = rp; | 2255 | rcu_assign_pointer(rp->rph->rp, rp); |
| 2256 | rp->nmissed = 0; | 2256 | rp->nmissed = 0; |
| 2257 | /* Establish function entry probe point */ | 2257 | /* Establish function entry probe point */ |
| 2258 | ret = register_kprobe(&rp->kp); | 2258 | ret = register_kprobe(&rp->kp); |
| @@ -2300,7 +2300,7 @@ void unregister_kretprobes(struct kretprobe **rps, int num) · 41 unmodified lines | |||
| 2300 | #ifdef CONFIG_KRETPROBE_ON_RETHOOK | 2300 | #ifdef CONFIG_KRETPROBE_ON_RETHOOK |
| 2301 | rethook_free(rps[i]->rh); | 2301 | rethook_free(rps[i]->rh); |
| 2302 | #else | 2302 | #else |
| 2303 | rps[i]->rph->rp = NULL; | 2303 | rcu_assign_pointer(rps[i]->rph->rp, NULL); |
| 2304 | #endif | 2304 | #endif |
| 2305 | } | 2305 | } |
| 2306 | mutex_unlock(&kprobe_mutex); | 2306 | mutex_unlock(&kprobe_mutex); |