| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <2802160.mvXUDI8C0e@7950hx>
Date: Mon, 13 Oct 2025 09:20:42 +0800
From: Menglong Dong <menglong.dong@...ux.dev>
To: Menglong Dong <menglong8.dong@...il.com>,
Masami Hiramatsu <mhiramat@...nel.org>
Cc: rostedt@...dmis.org, mathieu.desnoyers@...icios.com, jiang.biao@...ux.dev,
linux-kernel@...r.kernel.org, linux-trace-kernel@...r.kernel.org,
bpf@...r.kernel.org
Subject: Re: [PATCH v2 1/2] tracing: fprobe: optimization for entry only case
On 2025/10/12 12:07, Masami Hiramatsu wrote:
> Hi Menglong,
>
> On Fri, 10 Oct 2025 11:38:46 +0800
> Menglong Dong <menglong8.dong@...il.com> wrote:
>
> > For now, fgraph is used for the fprobe, even if we need trace the entry
> > only. However, the performance of ftrace is better than fgraph, and we
> > can use ftrace_ops for this case.
> >
> > Then performance of kprobe-multi increases from 54M to 69M. Before this
> > commit:
> >
> > $ ./benchs/run_bench_trigger.sh kprobe-multi
> > kprobe-multi : 54.663 ± 0.493M/s
> >
> > After this commit:
> >
> > $ ./benchs/run_bench_trigger.sh kprobe-multi
> > kprobe-multi : 69.447 ± 0.143M/s
> >
> > Mitigation is disable during the bench testing above.
> >
>
> Thanks for updating!
>
> This looks good to me. Just a nit comment below;
>
> [...]
> > @@ -379,11 +380,82 @@ static void fprobe_return(struct ftrace_graph_ret *trace,
> > NOKPROBE_SYMBOL(fprobe_return);
> >
> > static struct fgraph_ops fprobe_graph_ops = {
> > - .entryfunc = fprobe_entry,
> > + .entryfunc = fprobe_fgraph_entry,
> > .retfunc = fprobe_return,
> > };
> > static int fprobe_graph_active;
> >
> > +/* ftrace_ops callback, this processes fprobes which have only entry_handler. */
> > +static void fprobe_ftrace_entry(unsigned long ip, unsigned long parent_ip,
> > + struct ftrace_ops *ops, struct ftrace_regs *fregs)
> > +{
> > + struct fprobe_hlist_node *node;
> > + struct rhlist_head *head, *pos;
> > + struct fprobe *fp;
> > + int bit;
> > +
> > + bit = ftrace_test_recursion_trylock(ip, parent_ip);
> > + if (bit < 0)
> > + return;
> > +
>
> nit: We'd better to explain why we need this here too;
>
> /*
> * ftrace_test_recursion_trylock() disables preemption, but
> * rhltable_lookup() checks whether rcu_read_lcok is held.
> * So we take rcu_read_lock() here.
> */
It's very nice! I'll send a V3 now.
Thanks!
Menglong Dong
>
> > + rcu_read_lock();
> > + head = rhltable_lookup(&fprobe_ip_table, &ip, fprobe_rht_params);
> > +
> > + rhl_for_each_entry_rcu(node, pos, head, hlist) {
> > + if (node->addr != ip)
> > + break;
> > + fp = READ_ONCE(node->fp);
> > + if (unlikely(!fp || fprobe_disabled(fp) || fp->exit_handler))
> > + continue;
> > +
> > + if (fprobe_shared_with_kprobes(fp))
> > + __fprobe_kprobe_handler(ip, parent_ip, fp, fregs, NULL);
> > + else
> > + __fprobe_handler(ip, parent_ip, fp, fregs, NULL);
> > + }
> > + rcu_read_unlock();
> > + ftrace_test_recursion_unlock(bit);
> > +}
> > +NOKPROBE_SYMBOL(fprobe_ftrace_entry);
>
> Thank you,
>
> > +
> > +static struct ftrace_ops fprobe_ftrace_ops = {
> > + .func = fprobe_ftrace_entry,
> > + .flags = FTRACE_OPS_FL_SAVE_REGS,
> > +};
> > +static int fprobe_ftrace_active;
> > +
> > +static int fprobe_ftrace_add_ips(unsigned long *addrs, int num)
> > +{
> > + int ret;
> > +
> > + lockdep_assert_held(&fprobe_mutex);
> > +
> > + ret = ftrace_set_filter_ips(&fprobe_ftrace_ops, addrs, num, 0, 0);
> > + if (ret)
> > + return ret;
> > +
> > + if (!fprobe_ftrace_active) {
> > + ret = register_ftrace_function(&fprobe_ftrace_ops);
> > + if (ret) {
> > + ftrace_free_filter(&fprobe_ftrace_ops);
> > + return ret;
> > + }
> > + }
> > + fprobe_ftrace_active++;
> > + return 0;
> > +}
> > +
> > +static void fprobe_ftrace_remove_ips(unsigned long *addrs, int num)
> > +{
> > + lockdep_assert_held(&fprobe_mutex);
> > +
> > + fprobe_ftrace_active--;
> > + if (!fprobe_ftrace_active)
> > + unregister_ftrace_function(&fprobe_ftrace_ops);
> > + if (num)
> > + ftrace_set_filter_ips(&fprobe_ftrace_ops, addrs, num, 1, 0);
> > +}
> > +
> > /* Add @addrs to the ftrace filter and register fgraph if needed. */
> > static int fprobe_graph_add_ips(unsigned long *addrs, int num)
> > {
> > @@ -498,9 +570,12 @@ static int fprobe_module_callback(struct notifier_block *nb,
> > } while (node == ERR_PTR(-EAGAIN));
> > rhashtable_walk_exit(&iter);
> >
> > - if (alist.index > 0)
> > + if (alist.index > 0) {
> > ftrace_set_filter_ips(&fprobe_graph_ops.ops,
> > alist.addrs, alist.index, 1, 0);
> > + ftrace_set_filter_ips(&fprobe_ftrace_ops,
> > + alist.addrs, alist.index, 1, 0);
> > + }
> > mutex_unlock(&fprobe_mutex);
> >
> > kfree(alist.addrs);
> > @@ -733,7 +808,11 @@ int register_fprobe_ips(struct fprobe *fp, unsigned long *addrs, int num)
> > mutex_lock(&fprobe_mutex);
> >
> > hlist_array = fp->hlist_array;
> > - ret = fprobe_graph_add_ips(addrs, num);
> > + if (fp->exit_handler)
> > + ret = fprobe_graph_add_ips(addrs, num);
> > + else
> > + ret = fprobe_ftrace_add_ips(addrs, num);
> > +
> > if (!ret) {
> > add_fprobe_hash(fp);
> > for (i = 0; i < hlist_array->size; i++) {
> > @@ -829,7 +908,10 @@ int unregister_fprobe(struct fprobe *fp)
> > }
> > del_fprobe_hash(fp);
> >
> > - fprobe_graph_remove_ips(addrs, count);
> > + if (fp->exit_handler)
> > + fprobe_graph_remove_ips(addrs, count);
> > + else
> > + fprobe_ftrace_remove_ips(addrs, count);
> >
> > kfree_rcu(hlist_array, rcu);
> > fp->hlist_array = NULL;
>
>
>
Powered by blists - more mailing lists