From patchwork Tue Apr 26 01:57:50 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Li Huafei X-Patchwork-Id: 12826414 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id BD2C1C433F5 for ; Tue, 26 Apr 2022 01:59:25 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender: Content-Transfer-Encoding:Content-Type:List-Subscribe:List-Help:List-Post: List-Archive:List-Unsubscribe:List-Id:MIME-Version:Message-ID:Date:Subject:CC :To:From:Reply-To:Content-ID:Content-Description:Resent-Date:Resent-From: Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:In-Reply-To:References: List-Owner; bh=Nil/bGbGA+sVni3GjyTBmdyyZxRE1gDZge7iTklNnXE=; b=Po4FFnavprwghO W1zLBpZupDv3ijK4gAdgbr7JJ54uBVA9CcpMCJ2O3jbujyZiv8JHx/sNS/ySBLSq+kf+9HZ3+HIkF YDduQKpZPFevu1/I8P7VTVe0PGMTO8aRAmQEEyERm23re/YfviVDOfxjJo9xrNceYhwljbvhPpv0X EU2drb4H1FHz1PmDwyk4BvZG8lcjbCGf7tGyAT87CoIznTRkScNxnbblZj2CUsb0RfRKCVo7tHPwE YOGf9BKTJVytrLO84D3PDu8F5AOHIvNfXEW4xIMhZ9LoI1SGFi4kfnrFKpzeZMbeft0okXwMBvoLj +8+10CX04geoWGt+0QlA==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.94.2 #2 (Red Hat Linux)) id 1njATv-00C8Az-Vc; Tue, 26 Apr 2022 01:59:11 +0000 Received: from szxga02-in.huawei.com ([45.249.212.188]) by bombadil.infradead.org with esmtps (Exim 4.94.2 #2 (Red Hat Linux)) id 1njATs-00C87j-6R for linux-riscv@lists.infradead.org; Tue, 26 Apr 2022 01:59:10 +0000 Received: from kwepemi100020.china.huawei.com (unknown [172.30.72.55]) by szxga02-in.huawei.com (SkyGuard) with ESMTP id 4KnQ5v1kBJzQjBn; Tue, 26 Apr 2022 09:58:43 +0800 (CST) Received: from kwepemm600010.china.huawei.com (7.193.23.86) by kwepemi100020.china.huawei.com (7.221.188.48) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2375.24; Tue, 26 Apr 2022 09:58:59 +0800 Received: from ubuntu1804.huawei.com (10.67.174.174) by kwepemm600010.china.huawei.com (7.193.23.86) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2375.24; Tue, 26 Apr 2022 09:58:58 +0800 From: Li Huafei To: , , CC: , , , , , , , , , , , , , , Subject: [PATCH 1/2] riscv: ftrace: Fix the comments about the number of ftrace instruction Date: Tue, 26 Apr 2022 09:57:50 +0800 Message-ID: <20220426015751.88582-1-lihuafei1@huawei.com> X-Mailer: git-send-email 2.17.1 MIME-Version: 1.0 X-Originating-IP: [10.67.174.174] X-ClientProxiedBy: dggems704-chm.china.huawei.com (10.3.19.181) To kwepemm600010.china.huawei.com (7.193.23.86) X-CFilter-Loop: Reflected X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20220425_185908_442431_F8DCCB27 X-CRM114-Status: UNSURE ( 8.25 ) X-CRM114-Notice: Please train this message. X-BeenThere: linux-riscv@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-riscv" Errors-To: linux-riscv-bounces+linux-riscv=archiver.kernel.org@lists.infradead.org When DYNAMIC_FTRACE is enabled, we put four instructions in front of the function for ftrace use, not five. Signed-off-by: Li Huafei --- arch/riscv/kernel/ftrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/riscv/kernel/ftrace.c b/arch/riscv/kernel/ftrace.c index 4716f4cdc038..63f457650fa4 100644 --- a/arch/riscv/kernel/ftrace.c +++ b/arch/riscv/kernel/ftrace.c @@ -73,7 +73,7 @@ static int __ftrace_modify_call(unsigned long hook_pos, unsigned long target, } /* - * Put 5 instructions with 16 bytes at the front of function within + * Put 4 instructions with 16 bytes at the front of function within * patchable function entry nops' area. * * 0: REG_S ra, -SZREG(sp) From patchwork Tue Apr 26 01:57:51 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Li Huafei X-Patchwork-Id: 12826413 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 756A4C4332F for ; Tue, 26 Apr 2022 01:59:23 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender: Content-Transfer-Encoding:Content-Type:List-Subscribe:List-Help:List-Post: List-Archive:List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To: Message-ID:Date:Subject:CC:To:From:Reply-To:Content-ID:Content-Description: Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID: List-Owner; bh=cHHdvffnT4rEwhA8ee98yU5g85MlVm/smQM/ONT4h7s=; b=DCFRGS/uAfawY1 c+zWb2/pbn4i0DvIoGdncAM4qW4Mo6kvMxmLLdRnibZ2tA4/GBr5BLvJnfI2g2Y7dKxDvCt0nifSk RgQ4l0/3x6hZlLHFh3PshB3sCYmPaLnsKN3/BtT3xYi7qD1ndmRVPgzpg+dUmlXDT5+dyy+XurYk8 zIYN7MnQXCVe6r22GS7NHfg0mn89hjY/ObVOt5md3T+wehJIRTMTD6iRSX4PSL/AGBtSr0Gnc+duY Y2kcw0lt9dTN3OeWJDdvkCr5JLOfo3nulMWUGoQWksjw8ElJuBh4PlHB83JH8DesOqljIs9NMkfA8 KQgCVWTCOhONPpzraV2Q==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.94.2 #2 (Red Hat Linux)) id 1njATx-00C8Bi-Gc; Tue, 26 Apr 2022 01:59:13 +0000 Received: from szxga08-in.huawei.com ([45.249.212.255]) by bombadil.infradead.org with esmtps (Exim 4.94.2 #2 (Red Hat Linux)) id 1njATs-00C87i-AF for linux-riscv@lists.infradead.org; Tue, 26 Apr 2022 01:59:11 +0000 Received: from kwepemi100024.china.huawei.com (unknown [172.30.72.55]) by szxga08-in.huawei.com (SkyGuard) with ESMTP id 4KnQ5D0q66z1JBmX; Tue, 26 Apr 2022 09:58:08 +0800 (CST) Received: from kwepemm600010.china.huawei.com (7.193.23.86) by kwepemi100024.china.huawei.com (7.221.188.87) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2375.24; Tue, 26 Apr 2022 09:59:00 +0800 Received: from ubuntu1804.huawei.com (10.67.174.174) by kwepemm600010.china.huawei.com (7.193.23.86) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2375.24; Tue, 26 Apr 2022 09:58:59 +0800 From: Li Huafei To: , , CC: , , , , , , , , , , , , , , Subject: [PATCH 2/2] riscv: kprobe: Allow coexistence of ftrace and kprobe Date: Tue, 26 Apr 2022 09:57:51 +0800 Message-ID: <20220426015751.88582-2-lihuafei1@huawei.com> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20220426015751.88582-1-lihuafei1@huawei.com> References: <20220426015751.88582-1-lihuafei1@huawei.com> MIME-Version: 1.0 X-Originating-IP: [10.67.174.174] X-ClientProxiedBy: dggems704-chm.china.huawei.com (10.3.19.181) To kwepemm600010.china.huawei.com (7.193.23.86) X-CFilter-Loop: Reflected X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20220425_185908_717951_E5CA6547 X-CRM114-Status: GOOD ( 19.94 ) X-BeenThere: linux-riscv@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-riscv" Errors-To: linux-riscv-bounces+linux-riscv=archiver.kernel.org@lists.infradead.org When using ftrace and kprobe at the same time, it was found that one might cause the other to be unavailable. This can be reproduced by the following steps. # cd /sys/kernel/debug/tracing/ # echo cmdline_proc_show > set_ftrace_filter # echo function > current_tracer # echo 'p cmdline_proc_show' > kprobe_events # echo 'p cmdline_proc_show+4' >> kprobe_events # ls events/kprobes/ enable p_cmdline_proc_show_0 filter p_cmdline_proc_show_4 # echo 1 > events/kprobes/p_cmdline_proc_show_4/enable # echo 1 > events/kprobes/p_cmdline_proc_show_0/enable [ 129.830108] 00000000ebed457d: expected (ffdb0097 4f0080e7) but got (00100073 4f0080e7) [ 129.835990] ------------[ ftrace bug ]------------ [ 129.839510] ftrace failed to modify [ 129.839536] [] cmdline_proc_show+0x0/0x46 [ 129.845831] actual: 23:3c:11:fe:73:00:10:00 [ 129.849299] Updating ftrace call site to call a different ftrace function [ 129.853998] ftrace record flags: e0000002 [ 129.856771] (2) R [ 129.856771] expected tramp: ffffffff80008e60 [ 129.861688] ------------[ cut here ]------------ [ 129.865092] WARNING: CPU: 0 PID: 14 at kernel/trace/ftrace.c:2085 ftrace_bug+0x21a/0x24c [ 129.870949] Modules linked in: [ 129.873301] CPU: 0 PID: 14 Comm: migration/0 Not tainted 5.18.0-rc3-00002-gd8bfcd250f58 #6 [ 129.879217] Hardware name: riscv-virtio,qemu (DT) [ 129.882666] Stopper: multi_cpu_stop+0x0/0x168 <- stop_machine_cpuslocked+0xfa/0x12e [ 129.888430] epc : ftrace_bug+0x21a/0x24c [ 129.891254] ra : ftrace_bug+0x21a/0x24c [ 129.894057] epc : ffffffff807c3bee ra : ffffffff807c3bee sp : ff20000000283c80 [ 129.899144] gp : ffffffff813a83b8 tp : ff60000080021600 t0 : ffffffff804155c0 [ 129.904257] t1 : 0720072007200720 t2 : 7420646574636570 s0 : ff20000000283cb0 [ 129.909402] s1 : ff6000007fe622a0 a0 : 0000000000000022 a1 : c0000000ffffefff [ 129.914472] a2 : 0000000000000001 a3 : 0000000000000000 a4 : 341adec112294700 [ 129.919487] a5 : 341adec112294700 a6 : 0000000000000730 a7 : 0000000000000030 [ 129.924595] s2 : ffffffff80258910 s3 : ffffffffffffffea s4 : 0000000000000000 [ 129.929776] s5 : 0000000000000a35 s6 : ffffffff80d667c8 s7 : ff6000007fe04000 [ 129.934892] s8 : 0000000000000004 s9 : 0000000000000002 s10: 0000000000000001 [ 129.939992] s11: 0000000000000003 t3 : ff6000007ff20f00 t4 : ff6000007ff20f00 [ 129.945134] t5 : ff6000007ff20000 t6 : ff200000002839c8 [ 129.948838] status: 0000000000000100 badaddr: 0000000000000000 cause: 0000000000000003 [ 129.954580] [] ftrace_replace_code+0xce/0xd0 [ 129.958771] [] ftrace_modify_all_code+0xb4/0x12c [ 129.964420] [] __ftrace_modify_code+0x12/0x1c [ 129.969163] [] multi_cpu_stop+0x9a/0x168 [ 129.973035] [] cpu_stopper_thread+0xb4/0x15e [ 129.977205] [] smpboot_thread_fn+0x106/0x1e4 [ 129.981277] [] kthread+0xee/0x108 [ 129.984763] [] ret_from_exception+0x0/0xc [ 129.988809] ---[ end trace 0000000000000000 ]--- # cat trace # tracer: function # # WARNING: FUNCTION TRACING IS CORRUPTED # MAY BE MISSING FUNCTION EVENTS # entries-in-buffer/entries-written: 0/0 #P:4 # # _-----=> irqs-off/BH-disabled # / _----=> need-resched # | / _---=> hardirq/softirq # || / _--=> preempt-depth # ||| / _-=> migrate-disable # |||| / delay # TASK-PID CPU# ||||| TIMESTAMP FUNCTION # | | | ||||| | | As you can see, the ftrace functionality is broken. This is because DYNAMIC_FTRACE uses 4 instructions to make the jump from the function entry to the ftrace trampoline. After 'echo function > current_tracer': : 0xffffffff80258910 <+0>: sd ra,-8(sp) 0xffffffff80258914 <+4>: auipc ra,0xffdb0 0xffffffff80258918 <+8>: jalr 1264(ra) # 0xffffffff80008e04 0xffffffff8025891c <+12>: ld ra,-8(sp) After 'echo 1 > events/kprobes/p_cmdline_proc_show_4/enable': : 0xffffffff80258910 <+0>: sd ra,-8(sp) 0xffffffff80258914 <+4>: ebreak 0xffffffff80258918 <+8>: jalr 1264(ra) 0xffffffff8025891c <+12>: ld ra,-8(sp) This time the second instruction is replaced with a breakpoint instruction and the ftrace-related instructions are broken. Then, when enabling 'p_cmdline_proc_show_0', a regs ftrace handler is registered based on KPROBES_ON_FTRACE, ftrace_modify_call() changes 'ftrace_caller' to 'ftrace_regs_caller', and checks for ftrace-related instructions is modified or not. Here it has been modified, causing ftrace to report a warning and disable itself. In turn, the breakpoints inserted by kprobe may be overwritten by ftrace. The problem is that we think that only the first instruction address at the function entry is needed for ftrace, but it is actually the first 4. As Masami mentioned in [1], we can treat the first 4 instructions as a 16-byte virtual instruction, and we can probe func+0 based on KPROBES_ON_FTRACE, but not func+4, func+8 and func+12. This way, ftrace and kprobe do not bother each other and can coexist. After this patch: # cd /sys/kernel/debug/tracing/ # echo cmdline_proc_show > set_ftrace_filter # echo function > current_tracer # echo 'p cmdline_proc_show' > kprobe_events # echo 'p cmdline_proc_show' > kprobe_events # echo 'p cmdline_proc_show+4' >> kprobe_events sh: write error: Invalid or incomplete multibyte or wide character # echo 'p cmdline_proc_show+8' >> kprobe_events sh: write error: Invalid or incomplete multibyte or wide character # echo 'p cmdline_proc_show+12' >> kprobe_events sh: write error: Invalid or incomplete multibyte or wide character # echo 'p cmdline_proc_show+16' >> kprobe_events # ls events/kprobes/ enable p_cmdline_proc_show_0 filter p_cmdline_proc_show_16 # echo 1 > events/kprobes/enable # cat ../kprobes/list ffffffff80258b14 k cmdline_proc_show+0x0 [FTRACE] ffffffff80258b24 k cmdline_proc_show+0x10 # cat /proc/cmdline nokaslr rootwait root=/dev/vda rw # cat trace # tracer: function # # entries-in-buffer/entries-written: 3/3 #P:4 # # _-----=> irqs-off/BH-disabled # / _----=> need-resched # | / _---=> hardirq/softirq # || / _--=> preempt-depth # ||| / _-=> migrate-disable # |||| / delay # TASK-PID CPU# ||||| TIMESTAMP FUNCTION # | | | ||||| | | cat-142 [000] ...2. 181.493233: p_cmdline_proc_show_0: (cmdline_proc_show+0x0/0x46) cat-142 [000] ...2. 181.494271: cmdline_proc_show <-seq_read_iter cat-142 [000] d.... 181.494573: p_cmdline_proc_show_16: (cmdline_proc_show+0x10/0x46) [1] https://patchwork.kernel.org/project/linux-arm-kernel/patch/20191218140622.57bbaca5@xhacker.debian/ Fixes: c22b0bcb1dd0 ("riscv: Add kprobes supported") Signed-off-by: Li Huafei --- arch/riscv/kernel/probes/kprobes.c | 31 ++++++++++++++++++++++++++++++ kernel/kprobes.c | 2 +- 2 files changed, 32 insertions(+), 1 deletion(-) diff --git a/arch/riscv/kernel/probes/kprobes.c b/arch/riscv/kernel/probes/kprobes.c index e6e950b7cf32..ebe32eb6c349 100644 --- a/arch/riscv/kernel/probes/kprobes.c +++ b/arch/riscv/kernel/probes/kprobes.c @@ -21,6 +21,37 @@ DEFINE_PER_CPU(struct kprobe_ctlblk, kprobe_ctlblk); static void __kprobes post_kprobe_handler(struct kprobe *, struct kprobe_ctlblk *, struct pt_regs *); +#ifdef CONFIG_DYNAMIC_FTRACE +kprobe_opcode_t *arch_adjust_kprobe_addr(unsigned long addr, unsigned long offset, + bool *on_func_entry) +{ + unsigned long faddr = ftrace_location(addr); + + /* + * The first 4 instructions at the beginning of the ftraced function + * are used as a jump to ftrace trampoline, which we can think of as + * a 16-byte virtual instruction. This allows the function entry to be + * probed with the help of KPROBES_ON_FTRACE and does not break ftrace + * functionality. + */ + if (faddr && offset <= 12) { + /* + * As mentioned above, we treat the first 4 instructions of + * the ftraced function as a virtual instruction, so that we + * can only probe func+0, not func+4, func+8 and func+12. + */ + if (offset) + return ERR_PTR(-EILSEQ); + + *on_func_entry = true; + return (kprobe_opcode_t *)faddr; + } + + *on_func_entry = !offset; + return (kprobe_opcode_t *)(addr + offset); +} +#endif + static void __kprobes arch_prepare_ss_slot(struct kprobe *p) { unsigned long offset = GET_INSN_LENGTH(p->opcode); diff --git a/kernel/kprobes.c b/kernel/kprobes.c index dbe57df2e199..b41de7ce8bff 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -1553,7 +1553,7 @@ _kprobe_addr(kprobe_opcode_t *addr, const char *symbol_name, * at the start of the function. */ addr = arch_adjust_kprobe_addr((unsigned long)addr, offset, on_func_entry); - if (addr) + if (IS_ERR(addr) || addr) return addr; invalid: