From: Jiri Olsa <jolsa@kernel.org>
To: Oleg Nesterov <oleg@redhat.com>,
Peter Zijlstra <peterz@infradead.org>,
Alexei Starovoitov <ast@kernel.org>,
Daniel Borkmann <daniel@iogearbox.net>,
Andrii Nakryiko <andrii@kernel.org>
Cc: bpf@vger.kernel.org, Martin KaFai Lau <kafai@fb.com>,
Song Liu <songliubraving@fb.com>, Yonghong Song <yhs@fb.com>,
John Fastabend <john.fastabend@gmail.com>,
KP Singh <kpsingh@chromium.org>,
Stanislav Fomichev <sdf@fomichev.me>, Hao Luo <haoluo@google.com>,
Steven Rostedt <rostedt@goodmis.org>,
Masami Hiramatsu <mhiramat@kernel.org>,
linux-kernel@vger.kernel.org, linux-trace-kernel@vger.kernel.org
Subject: [PATCHv6 perf/core 02/16] uprobe: Add support for session consumer
Date: Thu, 10 Oct 2024 22:09:43 +0200 [thread overview]
Message-ID: <20241010200957.2750179-3-jolsa@kernel.org> (raw)
In-Reply-To: <20241010200957.2750179-1-jolsa@kernel.org>
This change allows the uprobe consumer to behave as session which
means that 'handler' and 'ret_handler' callbacks are connected in
a way that allows to:
- control execution of 'ret_handler' from 'handler' callback
- share data between 'handler' and 'ret_handler' callbacks
The session concept fits to our common use case where we do filtering
on entry uprobe and based on the result we decide to run the return
uprobe (or not).
It's also convenient to share the data between session callbacks.
To achive this we are adding new return value the uprobe consumer
can return from 'handler' callback:
UPROBE_HANDLER_IGNORE
- Ignore 'ret_handler' callback for this consumer.
And store cookie and pass it to 'ret_handler' when consumer has both
'handler' and 'ret_handler' callbacks defined.
We store shared data in the return_consumer object array as part of
the return_instance object. This way the handle_uretprobe_chain can
find related return_consumer and its shared data.
We also store entry handler return value, for cases when there are
multiple consumers on single uprobe and some of them are ignored and
some of them not, in which case the return probe gets installed and
we need to have a way to find out which consumer needs to be ignored.
The tricky part is when consumer is registered 'after' the uprobe
entry handler is hit. In such case this consumer's 'ret_handler' gets
executed as well, but it won't have the proper data pointer set,
so we can filter it out.
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Suggested-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
---
include/linux/uprobes.h | 21 +++++-
kernel/events/uprobes.c | 148 ++++++++++++++++++++++++++++++++--------
2 files changed, 139 insertions(+), 30 deletions(-)
diff --git a/include/linux/uprobes.h b/include/linux/uprobes.h
index bb265a632b91..dbaf04189548 100644
--- a/include/linux/uprobes.h
+++ b/include/linux/uprobes.h
@@ -23,8 +23,17 @@ struct inode;
struct notifier_block;
struct page;
+/*
+ * Allowed return values from uprobe consumer's handler callback
+ * with following meaning:
+ *
+ * UPROBE_HANDLER_REMOVE
+ * - Remove the uprobe breakpoint from current->mm.
+ * UPROBE_HANDLER_IGNORE
+ * - Ignore ret_handler callback for this consumer.
+ */
#define UPROBE_HANDLER_REMOVE 1
-#define UPROBE_HANDLER_MASK 1
+#define UPROBE_HANDLER_IGNORE 2
#define MAX_URETPROBE_DEPTH 64
@@ -44,6 +53,8 @@ struct uprobe_consumer {
bool (*filter)(struct uprobe_consumer *self, struct mm_struct *mm);
struct list_head cons_node;
+
+ __u64 id; /* set when uprobe_consumer is registered */
};
#ifdef CONFIG_UPROBES
@@ -83,14 +94,22 @@ struct uprobe_task {
unsigned int depth;
};
+struct return_consumer {
+ __u64 cookie;
+ __u64 id;
+};
+
struct return_instance {
struct uprobe *uprobe;
unsigned long func;
unsigned long stack; /* stack pointer */
unsigned long orig_ret_vaddr; /* original return address */
bool chained; /* true, if instance is nested */
+ int consumers_cnt;
struct return_instance *next; /* keep as stack */
+
+ struct return_consumer consumers[] __counted_by(consumers_cnt);
};
enum rp_check {
diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
index 6b44c386a5df..4ef4b51776eb 100644
--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c
@@ -64,7 +64,7 @@ struct uprobe {
struct rcu_head rcu;
loff_t offset;
loff_t ref_ctr_offset;
- unsigned long flags;
+ unsigned long flags; /* "unsigned long" so bitops work */
/*
* The generic code assumes that it has two members of unknown type
@@ -823,8 +823,11 @@ static struct uprobe *alloc_uprobe(struct inode *inode, loff_t offset,
static void consumer_add(struct uprobe *uprobe, struct uprobe_consumer *uc)
{
+ static atomic64_t id;
+
down_write(&uprobe->consumer_rwsem);
list_add_rcu(&uc->cons_node, &uprobe->consumers);
+ uc->id = (__u64) atomic64_inc_return(&id);
up_write(&uprobe->consumer_rwsem);
}
@@ -1761,6 +1764,34 @@ static struct uprobe_task *get_utask(void)
return current->utask;
}
+static size_t ri_size(int consumers_cnt)
+{
+ struct return_instance *ri;
+
+ return sizeof(*ri) + sizeof(ri->consumers[0]) * consumers_cnt;
+}
+
+#define DEF_CNT 4
+
+static struct return_instance *alloc_return_instance(void)
+{
+ struct return_instance *ri;
+
+ ri = kzalloc(ri_size(DEF_CNT), GFP_KERNEL);
+ if (!ri)
+ return ZERO_SIZE_PTR;
+
+ ri->consumers_cnt = DEF_CNT;
+ return ri;
+}
+
+static struct return_instance *dup_return_instance(struct return_instance *old)
+{
+ size_t size = ri_size(old->consumers_cnt);
+
+ return kmemdup(old, size, GFP_KERNEL);
+}
+
static int dup_utask(struct task_struct *t, struct uprobe_task *o_utask)
{
struct uprobe_task *n_utask;
@@ -1773,11 +1804,10 @@ static int dup_utask(struct task_struct *t, struct uprobe_task *o_utask)
p = &n_utask->return_instances;
for (o = o_utask->return_instances; o; o = o->next) {
- n = kmalloc(sizeof(struct return_instance), GFP_KERNEL);
+ n = dup_return_instance(o);
if (!n)
return -ENOMEM;
- *n = *o;
/*
* uprobe's refcnt has to be positive at this point, kept by
* utask->return_instances items; return_instances can't be
@@ -1870,35 +1900,31 @@ static void cleanup_return_instances(struct uprobe_task *utask, bool chained,
utask->return_instances = ri;
}
-static void prepare_uretprobe(struct uprobe *uprobe, struct pt_regs *regs)
+static void prepare_uretprobe(struct uprobe *uprobe, struct pt_regs *regs,
+ struct return_instance *ri)
{
struct uprobe_task *utask = current->utask;
unsigned long orig_ret_vaddr, trampoline_vaddr;
- struct return_instance *ri;
bool chained;
if (!get_xol_area())
- return;
+ goto free;
if (utask->depth >= MAX_URETPROBE_DEPTH) {
printk_ratelimited(KERN_INFO "uprobe: omit uretprobe due to"
" nestedness limit pid/tgid=%d/%d\n",
current->pid, current->tgid);
- return;
+ goto free;
}
/* we need to bump refcount to store uprobe in utask */
if (!try_get_uprobe(uprobe))
- return;
-
- ri = kmalloc(sizeof(struct return_instance), GFP_KERNEL);
- if (!ri)
- goto fail;
+ goto free;
trampoline_vaddr = uprobe_get_trampoline_vaddr();
orig_ret_vaddr = arch_uretprobe_hijack_return_addr(trampoline_vaddr, regs);
if (orig_ret_vaddr == -1)
- goto fail;
+ goto put;
/* drop the entries invalidated by longjmp() */
chained = (orig_ret_vaddr == trampoline_vaddr);
@@ -1916,7 +1942,7 @@ static void prepare_uretprobe(struct uprobe *uprobe, struct pt_regs *regs)
* attack from user-space.
*/
uprobe_warn(current, "handle tail call");
- goto fail;
+ goto put;
}
orig_ret_vaddr = utask->return_instances->orig_ret_vaddr;
}
@@ -1931,9 +1957,10 @@ static void prepare_uretprobe(struct uprobe *uprobe, struct pt_regs *regs)
utask->return_instances = ri;
return;
-fail:
- kfree(ri);
+put:
put_uprobe(uprobe);
+free:
+ kfree(ri);
}
/* Prepare to single-step probed instruction out of line. */
@@ -2077,34 +2104,90 @@ static struct uprobe *find_active_uprobe_rcu(unsigned long bp_vaddr, int *is_swb
return uprobe;
}
+static struct return_instance*
+push_consumer(struct return_instance *ri, int idx, __u64 id, __u64 cookie)
+{
+ if (unlikely(ri == ZERO_SIZE_PTR))
+ return ri;
+
+ if (unlikely(idx >= ri->consumers_cnt)) {
+ struct return_instance *old_ri = ri;
+
+ ri->consumers_cnt += DEF_CNT;
+ ri = krealloc(old_ri, ri_size(old_ri->consumers_cnt), GFP_KERNEL);
+ if (!ri) {
+ kfree(old_ri);
+ return ZERO_SIZE_PTR;
+ }
+ }
+
+ ri->consumers[idx].id = id;
+ ri->consumers[idx].cookie = cookie;
+ return ri;
+}
+
+static struct return_consumer *
+return_consumer_find(struct return_instance *ri, int *iter, int id)
+{
+ struct return_consumer *ric;
+ int idx = *iter;
+
+ for (ric = &ri->consumers[idx]; idx < ri->consumers_cnt; idx++, ric++) {
+ if (ric->id == id) {
+ *iter = idx + 1;
+ return ric;
+ }
+ }
+ return NULL;
+}
+
+static bool ignore_ret_handler(int rc)
+{
+ return rc == UPROBE_HANDLER_REMOVE || rc == UPROBE_HANDLER_IGNORE;
+}
+
static void handler_chain(struct uprobe *uprobe, struct pt_regs *regs)
{
struct uprobe_consumer *uc;
- int remove = UPROBE_HANDLER_REMOVE;
- bool need_prep = false; /* prepare return uprobe, when needed */
- bool has_consumers = false;
+ bool has_consumers = false, remove = true;
+ struct return_instance *ri = NULL;
+ int push_idx = 0;
current->utask->auprobe = &uprobe->arch;
list_for_each_entry_rcu(uc, &uprobe->consumers, cons_node, rcu_read_lock_trace_held()) {
+ bool session = uc->handler && uc->ret_handler;
+ __u64 cookie = 0;
int rc = 0;
if (uc->handler) {
- rc = uc->handler(uc, regs, NULL);
- WARN(rc & ~UPROBE_HANDLER_MASK,
+ rc = uc->handler(uc, regs, &cookie);
+ WARN(rc < 0 || rc > 2,
"bad rc=0x%x from %ps()\n", rc, uc->handler);
}
- if (uc->ret_handler)
- need_prep = true;
-
- remove &= rc;
+ remove &= rc == UPROBE_HANDLER_REMOVE;
has_consumers = true;
+
+ if (!uc->ret_handler || ignore_ret_handler(rc))
+ continue;
+
+ if (!ri)
+ ri = alloc_return_instance();
+
+ if (session)
+ ri = push_consumer(ri, push_idx++, uc->id, cookie);
}
current->utask->auprobe = NULL;
- if (need_prep && !remove)
- prepare_uretprobe(uprobe, regs); /* put bp at return */
+ if (!ZERO_OR_NULL_PTR(ri)) {
+ /*
+ * The push_idx value has the final number of return consumers,
+ * and ri->consumers_cnt has number of allocated consumers.
+ */
+ ri->consumers_cnt = push_idx;
+ prepare_uretprobe(uprobe, regs, ri);
+ }
if (remove && has_consumers) {
down_read(&uprobe->register_rwsem);
@@ -2123,12 +2206,19 @@ static void
handle_uretprobe_chain(struct return_instance *ri, struct pt_regs *regs)
{
struct uprobe *uprobe = ri->uprobe;
+ struct return_consumer *ric;
struct uprobe_consumer *uc;
+ int ric_idx = 0;
rcu_read_lock_trace();
list_for_each_entry_rcu(uc, &uprobe->consumers, cons_node, rcu_read_lock_trace_held()) {
- if (uc->ret_handler)
- uc->ret_handler(uc, ri->func, regs, NULL);
+ bool session = uc->handler && uc->ret_handler;
+
+ if (uc->ret_handler) {
+ ric = return_consumer_find(ri, &ric_idx, uc->id);
+ if (!session || ric)
+ uc->ret_handler(uc, ri->func, regs, ric ? &ric->cookie : NULL);
+ }
}
rcu_read_unlock_trace();
}
--
2.46.2
next prev parent reply other threads:[~2024-10-10 20:10 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-10 20:09 [PATCHv6 bpf-next,perf/core 00/16] uprobe, bpf: Add session support Jiri Olsa
2024-10-10 20:09 ` [PATCHv6 perf/core 01/16] uprobe: Add data pointer to consumer handlers Jiri Olsa
2024-10-15 22:24 ` Masami Hiramatsu
2024-10-16 6:41 ` Jiri Olsa
2024-10-10 20:09 ` Jiri Olsa [this message]
2024-10-11 11:27 ` [PATCHv6 perf/core 02/16] uprobe: Add support for session consumer Oleg Nesterov
2024-10-16 0:07 ` Masami Hiramatsu
2024-10-10 20:09 ` [PATCHv6 bpf-next 03/16] bpf: Allow return values 0 and 1 for kprobe session Jiri Olsa
2024-10-11 2:19 ` Andrii Nakryiko
2024-10-10 20:09 ` [PATCHv6 bpf-next 04/16] bpf: Force uprobe bpf program to always return 0 Jiri Olsa
2024-10-11 2:20 ` Andrii Nakryiko
2024-10-10 20:09 ` [PATCHv6 bpf-next 05/16] bpf: Add support for uprobe multi session attach Jiri Olsa
2024-10-10 20:09 ` [PATCHv6 bpf-next 06/16] bpf: Add support for uprobe multi session context Jiri Olsa
2024-10-10 20:09 ` [PATCHv6 bpf-next 07/16] libbpf: Add support for uprobe multi session attach Jiri Olsa
2024-10-11 2:21 ` Andrii Nakryiko
2024-10-10 20:09 ` [PATCHv6 bpf-next 08/16] selftests/bpf: Add uprobe session test Jiri Olsa
2024-10-11 2:22 ` Andrii Nakryiko
2024-10-10 20:09 ` [PATCHv6 bpf-next 09/16] selftests/bpf: Add uprobe session cookie test Jiri Olsa
2024-10-10 20:09 ` [PATCHv6 bpf-next 10/16] selftests/bpf: Add uprobe session recursive test Jiri Olsa
2024-10-10 20:09 ` [PATCHv6 bpf-next 11/16] selftests/bpf: Add uprobe session verifier test for return value Jiri Olsa
2024-10-11 2:23 ` Andrii Nakryiko
2024-10-10 20:09 ` [PATCHv6 bpf-next 12/16] selftests/bpf: Add kprobe " Jiri Olsa
2024-10-11 2:24 ` Andrii Nakryiko
2024-10-10 20:09 ` [PATCHv6 bpf-next 13/16] selftests/bpf: Add uprobe session single consumer test Jiri Olsa
2024-10-11 2:25 ` Andrii Nakryiko
2024-10-11 11:33 ` Jiri Olsa
2024-10-10 20:09 ` [PATCHv6 bpf-next 14/16] selftests/bpf: Scale down uprobe multi " Jiri Olsa
2024-10-11 2:27 ` Andrii Nakryiko
2024-10-11 11:36 ` Jiri Olsa
2024-10-10 20:09 ` [PATCHv6 bpf-next 15/16] selftests/bpf: Add uprobe sessions to " Jiri Olsa
2024-10-11 2:30 ` Andrii Nakryiko
2024-10-11 11:36 ` Jiri Olsa
2024-10-10 20:09 ` [PATCHv6 bpf-next 16/16] selftests/bpf: Add threads " Jiri Olsa
2024-10-11 2:35 ` Andrii Nakryiko
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20241010200957.2750179-3-jolsa@kernel.org \
--to=jolsa@kernel.org \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=haoluo@google.com \
--cc=john.fastabend@gmail.com \
--cc=kafai@fb.com \
--cc=kpsingh@chromium.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-trace-kernel@vger.kernel.org \
--cc=mhiramat@kernel.org \
--cc=oleg@redhat.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=sdf@fomichev.me \
--cc=songliubraving@fb.com \
--cc=yhs@fb.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).