* [PATCH] ARC: MM: fix UB and kernel resourse leak in do_page_fault
@ 2018-11-07 12:12 Eugeniy Paltsev
2018-11-09 17:32 ` Vineet Gupta
0 siblings, 1 reply; 2+ messages in thread
From: Eugeniy Paltsev @ 2018-11-07 12:12 UTC (permalink / raw)
To: linux-snps-arc, Vineet Gupta
Cc: linux-kernel, Alexey Brodkin, Eric W . Biederman, linux-arch,
Eugeniy Paltsev
Commit 15773ae938d8 ("signal/arc: Use force_sig_fault where
appropriate") adds undefined behaviour and kernel resource leak
to userspace in ARC do_page_fault() implementation.
This happens because we don't initialize si_code variable after we
switch to force_sig_fault using. si_code (as a part of siginfo_t
structure) was previously initialized by clear_siginfo(&info) call
which was removed.
Undefined behaviour path:
-------------------->8---------------------------
}}} a/arch/arc/mm/fault.c
!! -67,6 +67,7 @@ void do_page_fault(unsigned long address, struct pt_regs *regs)
struct task_struct *tsk = current;
struct mm_struct *mm = tsk->mm;
+ /// >>> si_code - uninitialized
int si_code;
int ret;
vm_fault_t fault;
int write = regs->ecr_cause & ECR_C_PROTV_STORE; /* ST/EX */
!! -81,8 +82,10 @@ void do_page_fault(unsigned long address, struct pt_regs *regs)
* only copy the information from the master page table,
* nothing more.
*/
+ /// >>> take true branch
if (address >= VMALLOC_START) {
ret = handle_kernel_vaddr_fault(address);
+ /// >>> take true branch
if (unlikely(ret))
+ /// >>> jump to label "bad_area_nosemaphore"
goto bad_area_nosemaphore;
else
!! -193,10 +196,13 @@ void do_page_fault(unsigned long address, struct pt_regs *regs)
bad_area:
up_read(&mm->mmap_sem);
+ /// >>> reach label "bad_area_nosemaphore"
bad_area_nosemaphore:
/* User mode accesses just cause a SIGSEGV */
+ /// >>> take true branch
if (user_mode(regs)) {
tsk->thread.fault_address = address;
+ /// >>> Ooops:
+ /// >>> use uninitialized value "si_code"
+ /// >>> when calling "force_sig_fault"
force_sig_fault(SIGSEGV, si_code, (void __user *)address, tsk);
return;
}
-------------------->8---------------------------
Fixes: 15773ae938d8 ("signal/arc: Use force_sig_fault where appropriate")
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
---
arch/arc/mm/fault.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/arc/mm/fault.c b/arch/arc/mm/fault.c
index c9da6102eb4f..e2d9fc3fea01 100644
--- a/arch/arc/mm/fault.c
+++ b/arch/arc/mm/fault.c
@@ -66,7 +66,7 @@ void do_page_fault(unsigned long address, struct pt_regs *regs)
struct vm_area_struct *vma = NULL;
struct task_struct *tsk = current;
struct mm_struct *mm = tsk->mm;
- int si_code;
+ int si_code = 0;
int ret;
vm_fault_t fault;
int write = regs->ecr_cause & ECR_C_PROTV_STORE; /* ST/EX */
--
2.14.5
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] ARC: MM: fix UB and kernel resourse leak in do_page_fault
2018-11-07 12:12 [PATCH] ARC: MM: fix UB and kernel resourse leak in do_page_fault Eugeniy Paltsev
@ 2018-11-09 17:32 ` Vineet Gupta
0 siblings, 0 replies; 2+ messages in thread
From: Vineet Gupta @ 2018-11-09 17:32 UTC (permalink / raw)
To: Eugeniy Paltsev, linux-snps-arc
Cc: linux-kernel, Alexey Brodkin, Eric W . Biederman, linux-arch
On 11/7/18 4:12 AM, Eugeniy Paltsev wrote:
> Commit 15773ae938d8 ("signal/arc: Use force_sig_fault where
> appropriate") adds undefined behaviour and kernel resource leak
> to userspace in ARC do_page_fault() implementation.
>
> This happens because we don't initialize si_code variable after we
> switch to force_sig_fault using. si_code (as a part of siginfo_t
> structure) was previously initialized by clear_siginfo(&info) call
> which was removed.
>
> Undefined behaviour path:
> -------------------->8---------------------------
> }}} a/arch/arc/mm/fault.c
> !! -67,6 +67,7 @@ void do_page_fault(unsigned long address, struct pt_regs *regs)
> struct task_struct *tsk = current;
> struct mm_struct *mm = tsk->mm;
> + /// >>> si_code - uninitialized
> int si_code;
> int ret;
> vm_fault_t fault;
> int write = regs->ecr_cause & ECR_C_PROTV_STORE; /* ST/EX */
> !! -81,8 +82,10 @@ void do_page_fault(unsigned long address, struct pt_regs *regs)
> * only copy the information from the master page table,
> * nothing more.
> */
> + /// >>> take true branch
> if (address >= VMALLOC_START) {
> ret = handle_kernel_vaddr_fault(address);
> + /// >>> take true branch
> if (unlikely(ret))
While the fix is legit, are you sure this code path was taken. VMALLOC_START is
memory for kernel modules and/or vmalloc. So the fault was induced in kernel space
in which case user_mode)r) will not be true.
> + /// >>> jump to label "bad_area_nosemaphore"
> goto bad_area_nosemaphore;
> else
> !! -193,10 +196,13 @@ void do_page_fault(unsigned long address, struct pt_regs *regs)
> bad_area:
> up_read(&mm->mmap_sem);
>
> + /// >>> reach label "bad_area_nosemaphore"
> bad_area_nosemaphore:
> /* User mode accesses just cause a SIGSEGV */
> + /// >>> take true branch
> if (user_mode(regs)) {
> tsk->thread.fault_address = address;
> + /// >>> Ooops:
> + /// >>> use uninitialized value "si_code"
> + /// >>> when calling "force_sig_fault"
> force_sig_fault(SIGSEGV, si_code, (void __user *)address, tsk);
> return;
> }
> -------------------->8---------------------------
>
> Fixes: 15773ae938d8 ("signal/arc: Use force_sig_fault where appropriate")
> Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Added to for-curr after trimming the changelog
Thx,
-Vineet
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2018-11-09 17:32 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-11-07 12:12 [PATCH] ARC: MM: fix UB and kernel resourse leak in do_page_fault Eugeniy Paltsev
2018-11-09 17:32 ` Vineet Gupta
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).