| Message ID | 20190502171139.mqtegctsg35cir2e@linutronix.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] x86/fpu: Fault-in user stack if copy_fpstate_to_sigframe() fails | expand |
On Thu, May 02, 2019 at 07:11:39PM +0200, Sebastian Andrzej Siewior wrote: > In the compacted form, XSAVES may save only the XMM+SSE state but skip > FP (x87 state). > > This is denoted by header->xfeatures = 6. The fastpath > (copy_fpregs_to_sigframe()) does that but _also_ initialises the FP > state (cwd to 0x37f, mxcsr as we do, remaining fields to 0). > > The slowpath (copy_xstate_to_user()) leaves most of the FP > state untouched. Only mxcsr and mxcsr_flags are set due to > xfeatures_mxcsr_quirk(). Now that XFEATURE_MASK_FP is set > unconditionally, see > > 04944b793e18 ("x86: xsave: set FP, SSE bits in the xsave header in the user sigcontext"), > > on return from the signal, random garbage is loaded as the FP state. > > Instead of utilizing copy_xstate_to_user(), fault-in the user memory > and retry the fast path. Ideally, the fast path succeeds on the second > attempt but may be retried again if the memory is swapped out due > to memory pressure. If the user memory can not be faulted-in then > get_user_pages() returns an error so we don't loop forever. > > Fault in memory via get_user_pages_unlocked() so > copy_fpregs_to_sigframe() succeeds without a fault. > > Fixes: 69277c98f5eef ("x86/fpu: Always store the registers in copy_fpstate_to_sigframe()") > Reported-by: Kurt Kanzenbach <kurt.kanzenbach@linutronix.de> > Suggested-by: Dave Hansen <dave.hansen@intel.com> > Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> > --- > v1…v2: > - s/get_user_pages()/get_user_pages_unlocked()/ > - merge cleanups > > I'm posting this all-in-one fix up replacing the original patch so we > don't have a merge window with known bugs (that is the one that the > patch was going the fix and the KASAN fallout that it introduced). > > arch/x86/kernel/fpu/signal.c | 31 +++++++++++++++---------------- > 1 file changed, 15 insertions(+), 16 deletions(-) Queued to tip:WIP.x86/fpu for some hammering ontop ...
diff --git a/arch/x86/kernel/fpu/signal.c b/arch/x86/kernel/fpu/signal.c index 7026f1c4e5e30..5a8d118bc423e 100644 --- a/arch/x86/kernel/fpu/signal.c +++ b/arch/x86/kernel/fpu/signal.c @@ -157,11 +157,9 @@ static inline int copy_fpregs_to_sigframe(struct xregs_state __user *buf) */ int copy_fpstate_to_sigframe(void __user *buf, void __user *buf_fx, int size) { - struct fpu *fpu = ¤t->thread.fpu; - struct xregs_state *xsave = &fpu->state.xsave; struct task_struct *tsk = current; int ia32_fxstate = (buf != buf_fx); - int ret = -EFAULT; + int ret; ia32_fxstate &= (IS_ENABLED(CONFIG_X86_32) || IS_ENABLED(CONFIG_IA32_EMULATION)); @@ -174,11 +172,12 @@ int copy_fpstate_to_sigframe(void __user *buf, void __user *buf_fx, int size) sizeof(struct user_i387_ia32_struct), NULL, (struct _fpstate_32 __user *) buf) ? -1 : 1; +retry: /* * Load the FPU registers if they are not valid for the current task. * With a valid FPU state we can attempt to save the state directly to - * userland's stack frame which will likely succeed. If it does not, do - * the slowpath. + * userland's stack frame which will likely succeed. If it does not, + * resolve the fault in the user memory and try again. */ fpregs_lock(); if (test_thread_flag(TIF_NEED_FPU_LOAD)) @@ -187,20 +186,20 @@ int copy_fpstate_to_sigframe(void __user *buf, void __user *buf_fx, int size) pagefault_disable(); ret = copy_fpregs_to_sigframe(buf_fx); pagefault_enable(); - if (ret && !test_thread_flag(TIF_NEED_FPU_LOAD)) - copy_fpregs_to_fpstate(fpu); - set_thread_flag(TIF_NEED_FPU_LOAD); fpregs_unlock(); if (ret) { - if (using_compacted_format()) { - if (copy_xstate_to_user(buf_fx, xsave, 0, size)) - return -1; - } else { - fpstate_sanitize_xstate(fpu); - if (__copy_to_user(buf_fx, xsave, fpu_user_xstate_size)) - return -1; - } + int aligned_size; + int nr_pages; + + aligned_size = offset_in_page(buf_fx) + fpu_user_xstate_size; + nr_pages = DIV_ROUND_UP(aligned_size, PAGE_SIZE); + + ret = get_user_pages_unlocked((unsigned long)buf_fx, nr_pages, + NULL, FOLL_WRITE); + if (ret == nr_pages) + goto retry; + return -EFAULT; } /* Save the fsave header for the 32-bit frames. */
In the compacted form, XSAVES may save only the XMM+SSE state but skip FP (x87 state). This is denoted by header->xfeatures = 6. The fastpath (copy_fpregs_to_sigframe()) does that but _also_ initialises the FP state (cwd to 0x37f, mxcsr as we do, remaining fields to 0). The slowpath (copy_xstate_to_user()) leaves most of the FP state untouched. Only mxcsr and mxcsr_flags are set due to xfeatures_mxcsr_quirk(). Now that XFEATURE_MASK_FP is set unconditionally, see 04944b793e18 ("x86: xsave: set FP, SSE bits in the xsave header in the user sigcontext"), on return from the signal, random garbage is loaded as the FP state. Instead of utilizing copy_xstate_to_user(), fault-in the user memory and retry the fast path. Ideally, the fast path succeeds on the second attempt but may be retried again if the memory is swapped out due to memory pressure. If the user memory can not be faulted-in then get_user_pages() returns an error so we don't loop forever. Fault in memory via get_user_pages_unlocked() so copy_fpregs_to_sigframe() succeeds without a fault. Fixes: 69277c98f5eef ("x86/fpu: Always store the registers in copy_fpstate_to_sigframe()") Reported-by: Kurt Kanzenbach <kurt.kanzenbach@linutronix.de> Suggested-by: Dave Hansen <dave.hansen@intel.com> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> --- v1…v2: - s/get_user_pages()/get_user_pages_unlocked()/ - merge cleanups I'm posting this all-in-one fix up replacing the original patch so we don't have a merge window with known bugs (that is the one that the patch was going the fix and the KASAN fallout that it introduced). arch/x86/kernel/fpu/signal.c | 31 +++++++++++++++---------------- 1 file changed, 15 insertions(+), 16 deletions(-)