diff options
author | Peter Zijlstra <peterz@infradead.org> | 2019-07-11 13:40:59 +0200 |
---|---|---|
committer | Thomas Gleixner <tglx@linutronix.de> | 2019-07-17 23:17:38 +0200 |
commit | a0d14b8909de55139b8702fe0c7e80b69763dcfb (patch) | |
tree | a9f6b04591b9087f4a64086c27ac46c09a2f82aa /arch/x86/mm/fault.c | |
parent | 4234653e882740cbf6625eeee294e388b3176583 (diff) | |
download | linux-stable-a0d14b8909de55139b8702fe0c7e80b69763dcfb.tar.gz linux-stable-a0d14b8909de55139b8702fe0c7e80b69763dcfb.tar.bz2 linux-stable-a0d14b8909de55139b8702fe0c7e80b69763dcfb.zip |
x86/mm, tracing: Fix CR2 corruption
Despite the current efforts to read CR2 before tracing happens there still
exist a number of possible holes:
idtentry page_fault do_page_fault has_error_code=1
call error_entry
TRACE_IRQS_OFF
call trace_hardirqs_off*
#PF // modifies CR2
CALL_enter_from_user_mode
__context_tracking_exit()
trace_user_exit(0)
#PF // modifies CR2
call do_page_fault
address = read_cr2(); /* whoopsie */
And similar for i386.
Fix it by pulling the CR2 read into the entry code, before any of that
stuff gets a chance to run and ruin things.
Reported-by: He Zhe <zhe.he@windriver.com>
Reported-by: Eiichi Tsukata <devel@etsukata.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andy Lutomirski <luto@kernel.org>
Cc: bp@alien8.de
Cc: rostedt@goodmis.org
Cc: torvalds@linux-foundation.org
Cc: hpa@zytor.com
Cc: dave.hansen@linux.intel.com
Cc: jgross@suse.com
Cc: joel@joelfernandes.org
Link: https://lkml.kernel.org/r/20190711114336.116812491@infradead.org
Debugged-by: Steven Rostedt <rostedt@goodmis.org>
Diffstat (limited to 'arch/x86/mm/fault.c')
-rw-r--r-- | arch/x86/mm/fault.c | 30 |
1 files changed, 11 insertions, 19 deletions
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 794f364cb882..0799cc79efd3 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -1507,9 +1507,8 @@ good_area: NOKPROBE_SYMBOL(do_user_addr_fault); /* - * This routine handles page faults. It determines the address, - * and the problem, and then passes it off to one of the appropriate - * routines. + * Explicitly marked noinline such that the function tracer sees this as the + * page_fault entry point. */ static noinline void __do_page_fault(struct pt_regs *regs, unsigned long hw_error_code, @@ -1528,33 +1527,26 @@ __do_page_fault(struct pt_regs *regs, unsigned long hw_error_code, } NOKPROBE_SYMBOL(__do_page_fault); -static nokprobe_inline void -trace_page_fault_entries(unsigned long address, struct pt_regs *regs, - unsigned long error_code) +static __always_inline void +trace_page_fault_entries(struct pt_regs *regs, unsigned long error_code, + unsigned long address) { + if (!trace_pagefault_enabled()) + return; + if (user_mode(regs)) trace_page_fault_user(address, regs, error_code); else trace_page_fault_kernel(address, regs, error_code); } -/* - * We must have this function blacklisted from kprobes, tagged with notrace - * and call read_cr2() before calling anything else. To avoid calling any - * kind of tracing machinery before we've observed the CR2 value. - * - * exception_{enter,exit}() contains all sorts of tracepoints. - */ -dotraplinkage void notrace -do_page_fault(struct pt_regs *regs, unsigned long error_code) +dotraplinkage void +do_page_fault(struct pt_regs *regs, unsigned long error_code, unsigned long address) { - unsigned long address = read_cr2(); /* Get the faulting address */ enum ctx_state prev_state; prev_state = exception_enter(); - if (trace_pagefault_enabled()) - trace_page_fault_entries(address, regs, error_code); - + trace_page_fault_entries(regs, error_code, address); __do_page_fault(regs, error_code, address); exception_exit(prev_state); } |