Skip to contentSkip to conversation

x86/dumpstack: Don't dump kernel memory based on usermode RIP

PATCH In review Jann Horn 2 participants 2 messages Sep 3, 2018 Stable branch

lore
Share & tools
lore

Jann Horn

author
#0 lore raw

PATCH 1/1 x86/dumpstack: Don't dump kernel memory based on usermode RIP

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")

---
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(-)
arch/x86/include/asm/stacktrace.h +1 -1
index b6dc698f992a..f335aad404a4 100644
@@ -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 */
arch/x86/kernel/dumpstack.c +15 -6
index 666a284116ac..1c65c648cacc 100644
@@ -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)
arch/x86/mm/fault.c +1 -1
index 2aafa6ab6103..d1f1612672c7 100644
@@ -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
Signature · 1 line
2.19.0.rc1.350.ge57e33dbd1-goog
maintainer
#1 lore raw
Quoted 11 lines · 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.

Quoted 7 lines
Fixes 7cccf0725cf7 ("x86/dumpstack: Add a show_ip() function")

---
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