[PATCH for 4.18.y] x86/dumpstack: Don't dump kernel memory based on usermode RIP

[PATCH for 4.18.y] x86/dumpstack: Don't dump kernel memory based on usermode RIP

[Jann Horn]

commit 342db04ae71273322f0011384a9ed414df8bdae4 upstream.

show_opcodes() is used both for dumping kernel instructions and for dumping
user instructions. If userspace causes #PF by jumping to a kernel address,
show_opcodes() can be reached with regs->ip controlled by the user,
pointing to kernel code. Make sure that userspace can't trick us into
dumping kernel memory into dmesg.

Manually backported: show_opcodes() has changed a bit in the meantime.
I have manually tested the backport.

Fixes: 7cccf0725cf7 ("x86/dumpstack: Add a show_ip() function")
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20180828154901.112726-1-jannh@google.com
Signed-off-by: Jann Horn <jannh@google.com>
---
Since I manually backported this, I have removed all other
sign-off/reviewed-by lines. I hope that's correct?

 arch/x86/include/asm/stacktrace.h |  2 +-
 arch/x86/kernel/dumpstack.c       | 21 +++++++++++++++------
 arch/x86/mm/fault.c               |  2 +-
 3 files changed, 17 insertions(+), 8 deletions(-)

diff --git a/arch/x86/include/asm/stacktrace.h b/arch/x86/include/asm/stacktrace.h
index b6dc698f992a..f335aad404a4 100644
--- a/arch/x86/include/asm/stacktrace.h
+++ b/arch/x86/include/asm/stacktrace.h
@@ -111,6 +111,6 @@ static inline unsigned long caller_frame_pointer(void)
 	return (unsigned long)frame;
 }
 
-void show_opcodes(u8 *rip, const char *loglvl);
+void show_opcodes(struct pt_regs *regs, const char *loglvl);
 void show_ip(struct pt_regs *regs, const char *loglvl);
 #endif /* _ASM_X86_STACKTRACE_H */
diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c
index 666a284116ac..1c65c648cacc 100644
--- a/arch/x86/kernel/dumpstack.c
+++ b/arch/x86/kernel/dumpstack.c
@@ -91,23 +91,32 @@ static void printk_stack_address(unsigned long address, int reliable,
  * Thus, the 2/3rds prologue and 64 byte OPCODE_BUFSIZE is just a random
  * guesstimate in attempt to achieve all of the above.
  */
-void show_opcodes(u8 *rip, const char *loglvl)
+void show_opcodes(struct pt_regs *regs, const char *loglvl)
 {
 	unsigned int code_prologue = OPCODE_BUFSIZE * 2 / 3;
 	u8 opcodes[OPCODE_BUFSIZE];
-	u8 *ip;
+	unsigned long ip;
 	int i;
+	bool bad_ip;
 
 	printk("%sCode: ", loglvl);
 
-	ip = (u8 *)rip - code_prologue;
-	if (probe_kernel_read(opcodes, ip, OPCODE_BUFSIZE)) {
+	ip = regs->ip - code_prologue;
+
+	/*
+	 * Make sure userspace isn't trying to trick us into dumping kernel
+	 * memory by pointing the userspace instruction pointer at it.
+	 */
+	bad_ip = user_mode(regs) &&
+		 __chk_range_not_ok(ip, OPCODE_BUFSIZE, TASK_SIZE_MAX);
+
+	if (bad_ip || probe_kernel_read(opcodes, (u8 *)ip, OPCODE_BUFSIZE)) {
 		pr_cont("Bad RIP value.\n");
 		return;
 	}
 
 	for (i = 0; i < OPCODE_BUFSIZE; i++, ip++) {
-		if (ip == rip)
+		if (ip == regs->ip)
 			pr_cont("<%02x> ", opcodes[i]);
 		else
 			pr_cont("%02x ", opcodes[i]);
@@ -122,7 +131,7 @@ void show_ip(struct pt_regs *regs, const char *loglvl)
 #else
 	printk("%sRIP: %04x:%pS\n", loglvl, (int)regs->cs, (void *)regs->ip);
 #endif
-	show_opcodes((u8 *)regs->ip, loglvl);
+	show_opcodes(regs, loglvl);
 }
 
 void show_iret_regs(struct pt_regs *regs)
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 2aafa6ab6103..d1f1612672c7 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -838,7 +838,7 @@ show_signal_msg(struct pt_regs *regs, unsigned long error_code,
 
 	printk(KERN_CONT "\n");
 
-	show_opcodes((u8 *)regs->ip, loglvl);
+	show_opcodes(regs, loglvl);
 }
 
 static void
-- 
2.19.0.rc1.350.ge57e33dbd1-goog

Re: [PATCH for 4.18.y] x86/dumpstack: Don't dump kernel memory based on usermode RIP

[Greg KH]

On Mon, Sep 03, 2018 at 04:32:48PM +0200, Jann Horn wrote:
> commit 342db04ae71273322f0011384a9ed414df8bdae4 upstream.
> 
> show_opcodes() is used both for dumping kernel instructions and for dumping
> user instructions. If userspace causes #PF by jumping to a kernel address,
> show_opcodes() can be reached with regs->ip controlled by the user,
> pointing to kernel code. Make sure that userspace can't trick us into
> dumping kernel memory into dmesg.
> 
> Manually backported: show_opcodes() has changed a bit in the meantime.
> I have manually tested the backport.
> 
> Fixes: 7cccf0725cf7 ("x86/dumpstack: Add a show_ip() function")
> Cc: stable@vger.kernel.org
> Link: https://lkml.kernel.org/r/20180828154901.112726-1-jannh@google.com
> Signed-off-by: Jann Horn <jannh@google.com>
> ---
> Since I manually backported this, I have removed all other
> sign-off/reviewed-by lines. I hope that's correct?

Yes, that's fine, but I added them back as this wasn't that different of
a backport :)

THanks for the patch, now queued up.

greg k-h