<[email protected]>,Ingo Molnar <[email protected]>,Paolo Bonzini <[email protected]>,Dmitry Safonov <[email protected]>,Borislav Petkov <[email protected]>,Josh Poimboeuf <[email protected]>,Brian Gerst <[email protected]>,Jan Beulich <[email protected]>,Christian Borntraeger <[email protected]>,Fenghua Yu <[email protected]>,He Chen <[email protected]>,Russell King <[email protected]>,Vladimir Murzin <[email protected]>,Will Deacon <[email protected]>,Catalin Marinas <[email protected]>,Mark Rutland <[email protected]>,James Morse <[email protected]>,"David A . Long" <[email protected]>,Pratyush Anand <[email protected]>,Laura Abbott <[email protected]>,Andre Przywara <[email protected]>,Chris Metcalf <[email protected]>,[email protected],[email protected],[email protected],[email protected],[email protected],[email protected] From: [email protected] Message-ID: <[email protected]>
On March 11, 2017 1:42:00 AM PST, Ingo Molnar <[email protected]> wrote: > >* Thomas Garnier <[email protected]> wrote: > >> Implement specific usage of verify_pre_usermode_state for user-mode >> returns for x86. >> --- >> Based on next-20170308 >> --- >> arch/x86/Kconfig | 1 + >> arch/x86/entry/common.c | 3 +++ >> arch/x86/entry/entry_64.S | 19 +++++++++++++++++++ >> arch/x86/include/asm/pgtable_64_types.h | 11 +++++++++++ >> arch/x86/include/asm/processor.h | 11 ----------- >> 5 files changed, 34 insertions(+), 11 deletions(-) >> >> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig >> index 005df7c825f5..6d48e18e6f09 100644 >> --- a/arch/x86/Kconfig >> +++ b/arch/x86/Kconfig >> @@ -63,6 +63,7 @@ config X86 >> select ARCH_MIGHT_HAVE_ACPI_PDC if ACPI >> select ARCH_MIGHT_HAVE_PC_PARPORT >> select ARCH_MIGHT_HAVE_PC_SERIO >> + select ARCH_NO_SYSCALL_VERIFY_PRE_USERMODE_STATE >> select ARCH_SUPPORTS_ATOMIC_RMW >> select ARCH_SUPPORTS_DEFERRED_STRUCT_PAGE_INIT >> select ARCH_SUPPORTS_NUMA_BALANCING if X86_64 >> diff --git a/arch/x86/entry/common.c b/arch/x86/entry/common.c >> index 370c42c7f046..525edbb77f03 100644 >> --- a/arch/x86/entry/common.c >> +++ b/arch/x86/entry/common.c >> @@ -22,6 +22,7 @@ >> #include <linux/context_tracking.h> >> #include <linux/user-return-notifier.h> >> #include <linux/uprobes.h> >> +#include <linux/syscalls.h> >> >> #include <asm/desc.h> >> #include <asm/traps.h> >> @@ -180,6 +181,8 @@ __visible inline void >prepare_exit_to_usermode(struct pt_regs *regs) >> struct thread_info *ti = current_thread_info(); >> u32 cached_flags; >> >> + verify_pre_usermode_state(); >> + >> if (IS_ENABLED(CONFIG_PROVE_LOCKING) && WARN_ON(!irqs_disabled())) >> local_irq_disable(); >> >> diff --git a/arch/x86/entry/entry_64.S b/arch/x86/entry/entry_64.S >> index d2b2a2948ffe..04db589be466 100644 >> --- a/arch/x86/entry/entry_64.S >> +++ b/arch/x86/entry/entry_64.S >> @@ -218,6 +218,25 @@ entry_SYSCALL_64_fastpath: >> testl $_TIF_ALLWORK_MASK, TASK_TI_flags(%r11) >> jnz 1f >> >> + /* >> + * Check user-mode state on fast path return, the same check is >done >> + * under the slow path through syscall_return_slowpath. >> + */ >> +#ifdef CONFIG_BUG_ON_DATA_CORRUPTION >> + call verify_pre_usermode_state >> +#else >> + /* >> + * Similar to set_fs(USER_DS) in verify_pre_usermode_state without >a >> + * warning. >> + */ >> + movq PER_CPU_VAR(current_task), %rax >> + movq $TASK_SIZE_MAX, %rcx >> + cmp %rcx, TASK_addr_limit(%rax) >> + jz 1f >> + movq %rcx, TASK_addr_limit(%rax) >> +1: >> +#endif >> + >> LOCKDEP_SYS_EXIT >> TRACE_IRQS_ON /* user mode is traced as IRQs on */ >> movq RIP(%rsp), %rcx > >Ugh, so you call an assembly function just to ... call another >function. > >Plus why is it in assembly to begin with? Is this some older code that >got >written when the x86 entry code was in assembly, and never properly >converted to C? > >Thanks, > > Ingo The code does a compare to jump around a store. It would be much cleaner and faster to simply clobber the value unconditionally. If there is a test it should be to avoid the function call, not (only) the assignment. -- Sent from my Android device with K-9 Mail. Please excuse my brevity.
