From: Alexei Starovoitov <ast@kernel.org>
To: <davem@davemloft.net>
Cc: <daniel@iogearbox.net>, <peterz@infradead.org>,
<rostedt@goodmis.org>, <x86@kernel.org>, <netdev@vger.kernel.org>,
<bpf@vger.kernel.org>, <kernel-team@fb.com>
Subject: [PATCH bpf-next 1/7] bpf, ftrace: temporary workaround
Date: Sat, 2 Nov 2019 15:00:19 -0700 [thread overview]
Message-ID: <20191102220025.2475981-2-ast@kernel.org> (raw)
In-Reply-To: <20191102220025.2475981-1-ast@kernel.org>
Add temporary workaround until proper register_ftrace_direct() api lands. This
commit must be reverted during upcoming merge window. The hack functions don't
have safety checks that ftrace api performs.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
---
arch/x86/kernel/ftrace.c | 36 ++++++++++++++++++++++++++++++++++++
include/linux/bpf.h | 5 +++++
kernel/bpf/core.c | 30 ++++++++++++++++++++++++++++++
3 files changed, 71 insertions(+)
diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c
index 024c3053dbba..0fd7643da92d 100644
--- a/arch/x86/kernel/ftrace.c
+++ b/arch/x86/kernel/ftrace.c
@@ -265,6 +265,42 @@ static int update_ftrace_func(unsigned long ip, void *new)
return ret;
}
+/*
+ * There are no safety checks here. It's a temporary hack until proper
+ * register_ftrace_direct() api lands. These functions blindly rewrite kernel
+ * text. Proper register_ftrace_direct() should check that IP is one of allowed
+ * fentry points. unregister_ftrace_direct() and modify_ftrace_direct() should
+ * do similar checks.
+ */
+int arch_register_ftrace_hack(unsigned long ip, unsigned long addr)
+{
+ u8 *new;
+ int ret;
+
+ ftrace_arch_code_modify_prepare();
+ new = ftrace_call_replace(ip, addr);
+ ret = update_ftrace_func(ip, new);
+ ftrace_arch_code_modify_post_process();
+ return ret;
+}
+
+int arch_unregister_ftrace_hack(unsigned long ip, unsigned long addr)
+{
+ u8 *old;
+ int ret;
+
+ ftrace_arch_code_modify_prepare();
+ old = (void *)ftrace_nop_replace();
+ ret = update_ftrace_func(ip, old);
+ ftrace_arch_code_modify_post_process();
+ return ret;
+}
+
+int arch_modify_ftrace_hack(unsigned long ip, unsigned long addr)
+{
+ return arch_register_ftrace_hack(ip, addr);
+}
+
int ftrace_update_ftrace_func(ftrace_func_t func)
{
unsigned long ip = (unsigned long)(&ftrace_call);
diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 7c7f518811a6..a8941f113298 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -1157,4 +1157,9 @@ static inline u32 bpf_xdp_sock_convert_ctx_access(enum bpf_access_type type,
}
#endif /* CONFIG_INET */
+/* workaround */
+int register_ftrace_direct(unsigned long ip, unsigned long addr);
+int unregister_ftrace_direct(unsigned long ip, unsigned long addr);
+int modify_ftrace_direct(unsigned long ip, unsigned long addr);
+
#endif /* _LINUX_BPF_H */
diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index df82d5a42b23..1bacf70e6509 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -2140,6 +2140,36 @@ int __weak skb_copy_bits(const struct sk_buff *skb, int offset, void *to,
return -EFAULT;
}
+int __weak arch_register_ftrace_hack(unsigned long ip, unsigned long addr)
+{
+ return -ENOTSUPP;
+}
+
+int __weak register_ftrace_direct(unsigned long ip, unsigned long addr)
+{
+ return arch_register_ftrace_hack(ip, addr);
+}
+
+int __weak arch_unregister_ftrace_hack(unsigned long ip, unsigned long addr)
+{
+ return -ENOTSUPP;
+}
+
+int __weak unregister_ftrace_direct(unsigned long ip, unsigned long addr)
+{
+ return arch_unregister_ftrace_hack(ip, addr);
+}
+
+int __weak arch_modify_ftrace_hack(unsigned long ip, unsigned long addr)
+{
+ return -ENOTSUPP;
+}
+
+int __weak modify_ftrace_direct(unsigned long ip, unsigned long addr)
+{
+ return arch_modify_ftrace_hack(ip, addr);
+}
+
DEFINE_STATIC_KEY_FALSE(bpf_stats_enabled_key);
EXPORT_SYMBOL(bpf_stats_enabled_key);
--
2.23.0
next prev parent reply other threads:[~2019-11-02 22:01 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-11-02 22:00 [PATCH bpf-next 0/7] Introduce BPF trampoline Alexei Starovoitov
2019-11-02 22:00 ` Alexei Starovoitov [this message]
2019-11-02 22:00 ` [PATCH bpf-next 2/7] bpf: refactor x86 JIT into helpers Alexei Starovoitov
2019-11-02 22:00 ` [PATCH bpf-next 3/7] bpf: Introduce BPF trampoline Alexei Starovoitov
2019-11-05 19:51 ` Andrii Nakryiko
2019-11-02 22:00 ` [PATCH bpf-next 4/7] libbpf: Add support to attach to fentry/fexit tracing progs Alexei Starovoitov
2019-11-05 21:17 ` Andrii Nakryiko
2019-11-05 23:17 ` Alexei Starovoitov
2019-11-02 22:00 ` [PATCH bpf-next 5/7] selftest/bpf: Simple test for fentry/fexit Alexei Starovoitov
2019-11-05 21:37 ` Andrii Nakryiko
2019-11-02 22:00 ` [PATCH bpf-next 6/7] bpf: Add kernel test functions for fentry testing Alexei Starovoitov
2019-11-02 22:00 ` [PATCH bpf-next 7/7] selftests/bpf: Add test for BPF trampoline Alexei Starovoitov
2019-11-05 21:50 ` Andrii Nakryiko
2019-11-05 14:31 ` [PATCH bpf-next 0/7] Introduce " Alexei Starovoitov
2019-11-05 15:40 ` Steven Rostedt
2019-11-05 15:47 ` Alexei Starovoitov
2019-11-05 16:00 ` Steven Rostedt
2019-11-05 16:28 ` Alexei Starovoitov
2019-11-05 17:26 ` Steven Rostedt
2019-11-05 17:59 ` Alexei Starovoitov
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=20191102220025.2475981-2-ast@kernel.org \
--to=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=davem@davemloft.net \
--cc=kernel-team@fb.com \
--cc=netdev@vger.kernel.org \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=x86@kernel.org \
/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).