linux-kernel - Re: [PATCH 5.15 000/923] 5.15.3-rc3 review

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
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YZYfYOcqNqOyZ8Yo@hirez.programming.kicks-ass.net>
Date:   Thu, 18 Nov 2021 10:39:44 +0100
From:   Peter Zijlstra <peterz@...radead.org>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     Holger Hoffstätte 
        <holger@...lied-asynchrony.com>,
        Qi Zheng <zhengqi.arch@...edance.com>,
        Kees Cook <keescook@...omium.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Justin Forbes <jmforbes@...uxtx.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Guenter Roeck <linux@...ck-us.net>,
        Shuah Khan <shuah@...nel.org>, patches@...nelci.org,
        lkft-triage@...ts.linaro.org, Pavel Machek <pavel@...x.de>,
        Jon Hunter <jonathanh@...dia.com>,
        Florian Fainelli <f.fainelli@...il.com>,
        stable <stable@...r.kernel.org>,
        Josh Poimboeuf <jpoimboe@...hat.com>
Subject: Re: [PATCH 5.15 000/923] 5.15.3-rc3 review

On Thu, Nov 18, 2021 at 09:18:52AM +0100, Peter Zijlstra wrote:
> On Thu, Nov 18, 2021 at 09:06:27AM +0100, Peter Zijlstra wrote:
> > On Wed, Nov 17, 2021 at 03:50:17PM -0800, Linus Torvalds wrote:
> > 
> > > I really don't think the WCHAN code should use unwinders at all. It's
> > > too damn fragile, and it's too easily triggered from user space.
> > 
> > On x86, esp. with ORC, it pretty much has to. The thing is, the ORC
> > unwinder has been very stable so far. I'm guessing there's some really
> > stupid thing going on, like for example trying to unwind a freed stack.
> > 
> > I *just* managed to reproduce, so let me go have a poke.
> 
> Confirmed, with the below it no longer reproduces. Now, let me go undo
> that and fix the unwinder to not explode while trying to unwind nothing.

OK, so the bug is firmly with 5d1ceb3969b6 ("x86: Fix __get_wchan() for
!STACKTRACE") which lost the try_get_task_stack() that stack_trace_*()
does.

We can ofc trivially re-instate that, but I'm now running with the
below which I suppose is a better fix, hmm?

(obv I still need to look a the other two unwinders)

---
diff --git a/arch/x86/kernel/unwind_orc.c b/arch/x86/kernel/unwind_orc.c
index e6f7592790af..9261ff1343cf 100644
--- a/arch/x86/kernel/unwind_orc.c
+++ b/arch/x86/kernel/unwind_orc.c
@@ -352,8 +352,14 @@ static bool deref_stack_reg(struct unwind_state *state, unsigned long addr,
 	if (!stack_access_ok(state, addr, sizeof(long)))
 		return false;
 
-	*val = READ_ONCE_NOCHECK(*(unsigned long *)addr);
+	pagefault_disable();
+	__get_kernel_nofault(val, addr, unsigned long, Efault);
+	pagefault_enable();
 	return true;
+
+Efault:
+	pagefault_enable();
+	return false;
 }
 
 static bool deref_stack_regs(struct unwind_state *state, unsigned long addr,
@@ -367,9 +373,16 @@ static bool deref_stack_regs(struct unwind_state *state, unsigned long addr,
 	if (!stack_access_ok(state, addr, sizeof(struct pt_regs)))
 		return false;
 
-	*ip = READ_ONCE_NOCHECK(regs->ip);
-	*sp = READ_ONCE_NOCHECK(regs->sp);
+	pagefault_disable();
+	__get_kernel_nofault(ip, &regs->ip, unsigned long, Efault);
+	__get_kernel_nofault(sp, &regs->sp, unsigned long, Efault);
+	pagefault_enable();
+
 	return true;
+
+Efault:
+	pagefault_enable();
+	return false;
 }
 
 static bool deref_stack_iret_regs(struct unwind_state *state, unsigned long addr,
@@ -380,9 +393,16 @@ static bool deref_stack_iret_regs(struct unwind_state *state, unsigned long addr
 	if (!stack_access_ok(state, addr, IRET_FRAME_SIZE))
 		return false;
 
-	*ip = READ_ONCE_NOCHECK(regs->ip);
-	*sp = READ_ONCE_NOCHECK(regs->sp);
+	pagefault_disable();
+	__get_kernel_nofault(ip, &regs->ip, unsigned long, Efault);
+	__get_kernel_nofault(sp, &regs->sp, unsigned long, Efault);
+	pagefault_enable();
+
 	return true;
+
+Efault:
+	pagefault_enable();
+	return false;
 }
 
 /*
@@ -396,22 +416,27 @@ static bool deref_stack_iret_regs(struct unwind_state *state, unsigned long addr
 static bool get_reg(struct unwind_state *state, unsigned int reg_off,
 		    unsigned long *val)
 {
-	unsigned int reg = reg_off/8;
-
 	if (!state->regs)
 		return false;
 
+	pagefault_disable();
 	if (state->full_regs) {
-		*val = READ_ONCE_NOCHECK(((unsigned long *)state->regs)[reg]);
+		__get_kernel_nofault(val, (void *)state->regs + reg_off, unsigned long, Efault);
+		pagefault_enable();
 		return true;
 	}
 
 	if (state->prev_regs) {
-		*val = READ_ONCE_NOCHECK(((unsigned long *)state->prev_regs)[reg]);
+		__get_kernel_nofault(val, (void *)state->regs + reg_off, unsigned long, Efault);
+		pagefault_enable();
 		return true;
 	}
 
 	return false;
+
+Efault:
+	pagefault_enable();
+	return false;
 }
 
 bool unwind_next_frame(struct unwind_state *state)
@@ -673,8 +698,12 @@ void __unwind_start(struct unwind_state *state, struct task_struct *task,
 		struct inactive_task_frame *frame = (void *)task->thread.sp;
 
 		state->sp = task->thread.sp + sizeof(*frame);
-		state->bp = READ_ONCE_NOCHECK(frame->bp);
-		state->ip = READ_ONCE_NOCHECK(frame->ret_addr);
+
+		pagefault_disable();
+		__get_kernel_nofault(&state->bp, &frame->bp, unsigned long, Efault);
+		__get_kernel_nofault(&state->ip, &frame->ret_addr, unsigned long, Efault);
+		pagefault_enable();
+
 		state->signal = (void *)state->ip == ret_from_fork;
 	}
 
@@ -713,6 +742,8 @@ void __unwind_start(struct unwind_state *state, struct task_struct *task,
 
 	return;
 
+Efault:
+	pagefault_enable();
 err:
 	state->error = true;
 the_end: