2019-05-27 14:55:01 +08:00
|
|
|
// SPDX-License-Identifier: GPL-2.0-or-later
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* PowerPC version
|
|
|
|
* Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org)
|
|
|
|
*
|
|
|
|
* Derived from "arch/i386/kernel/signal.c"
|
|
|
|
* Copyright (C) 1991, 1992 Linus Torvalds
|
|
|
|
* 1997-11-28 Modified for POSIX.1b signals by Richard Henderson
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/sched.h>
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/smp.h>
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/signal.h>
|
|
|
|
#include <linux/errno.h>
|
|
|
|
#include <linux/wait.h>
|
|
|
|
#include <linux/unistd.h>
|
|
|
|
#include <linux/stddef.h>
|
|
|
|
#include <linux/elf.h>
|
|
|
|
#include <linux/ptrace.h>
|
2011-06-04 13:36:54 +08:00
|
|
|
#include <linux/ratelimit.h>
|
2018-05-02 21:20:47 +08:00
|
|
|
#include <linux/syscalls.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
#include <asm/sigcontext.h>
|
|
|
|
#include <asm/ucontext.h>
|
2016-12-25 03:46:01 +08:00
|
|
|
#include <linux/uaccess.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <asm/pgtable.h>
|
|
|
|
#include <asm/unistd.h>
|
|
|
|
#include <asm/cacheflush.h>
|
2006-03-23 07:00:08 +08:00
|
|
|
#include <asm/syscalls.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <asm/vdso.h>
|
2012-03-29 01:30:02 +08:00
|
|
|
#include <asm/switch_to.h>
|
2013-02-14 00:21:41 +08:00
|
|
|
#include <asm/tm.h>
|
2016-09-06 13:32:43 +08:00
|
|
|
#include <asm/asm-prototypes.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2007-06-04 13:15:49 +08:00
|
|
|
#include "signal.h"
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
|
2005-05-06 10:10:04 +08:00
|
|
|
#define GP_REGS_SIZE min(sizeof(elf_gregset_t), sizeof(struct pt_regs))
|
2005-04-17 06:20:36 +08:00
|
|
|
#define FP_REGS_SIZE sizeof(elf_fpregset_t)
|
|
|
|
|
|
|
|
#define TRAMP_TRACEBACK 3
|
|
|
|
#define TRAMP_SIZE 6
|
|
|
|
|
|
|
|
/*
|
|
|
|
* When we have signals to deliver, we set up on the user stack,
|
|
|
|
* going down from the original stack pointer:
|
|
|
|
* 1) a rt_sigframe struct which contains the ucontext
|
|
|
|
* 2) a gap of __SIGNAL_FRAMESIZE bytes which acts as a dummy caller
|
|
|
|
* frame for the signal handler.
|
|
|
|
*/
|
|
|
|
|
|
|
|
struct rt_sigframe {
|
|
|
|
/* sys_rt_sigreturn requires the ucontext be the first field */
|
|
|
|
struct ucontext uc;
|
2013-02-14 00:21:41 +08:00
|
|
|
#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
|
|
|
|
struct ucontext uc_transact;
|
|
|
|
#endif
|
2005-04-17 06:20:36 +08:00
|
|
|
unsigned long _unused[2];
|
|
|
|
unsigned int tramp[TRAMP_SIZE];
|
2006-02-01 18:28:09 +08:00
|
|
|
struct siginfo __user *pinfo;
|
|
|
|
void __user *puc;
|
2005-04-17 06:20:36 +08:00
|
|
|
struct siginfo info;
|
2014-02-26 14:07:38 +08:00
|
|
|
/* New 64 bit little-endian ABI allows redzone of 512 bytes below sp */
|
|
|
|
char abigap[USER_REDZONE_SIZE];
|
2005-04-17 06:20:36 +08:00
|
|
|
} __attribute__ ((aligned (16)));
|
|
|
|
|
2007-10-12 08:20:07 +08:00
|
|
|
static const char fmt32[] = KERN_INFO \
|
|
|
|
"%s[%d]: bad frame in %s: %08lx nip %08lx lr %08lx\n";
|
|
|
|
static const char fmt64[] = KERN_INFO \
|
|
|
|
"%s[%d]: bad frame in %s: %016lx nip %016lx lr %016lx\n";
|
|
|
|
|
2015-07-20 10:58:43 +08:00
|
|
|
/*
|
|
|
|
* This computes a quad word aligned pointer inside the vmx_reserve array
|
|
|
|
* element. For historical reasons sigcontext might not be quad word aligned,
|
|
|
|
* but the location we write the VMX regs to must be. See the comment in
|
|
|
|
* sigcontext for more detail.
|
|
|
|
*/
|
|
|
|
#ifdef CONFIG_ALTIVEC
|
|
|
|
static elf_vrreg_t __user *sigcontext_vmx_regs(struct sigcontext __user *sc)
|
|
|
|
{
|
|
|
|
return (elf_vrreg_t __user *) (((unsigned long)sc->vmx_reserve + 15) & ~0xful);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* Set up the sigcontext for the signal frame.
|
|
|
|
*/
|
|
|
|
|
2016-09-23 14:18:12 +08:00
|
|
|
static long setup_sigcontext(struct sigcontext __user *sc,
|
|
|
|
struct task_struct *tsk, int signr, sigset_t *set,
|
|
|
|
unsigned long handler, int ctx_has_vsx_region)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
/* When CONFIG_ALTIVEC is set, we _always_ setup v_regs even if the
|
|
|
|
* process never used altivec yet (MSR_VEC is zero in pt_regs of
|
|
|
|
* the context). This is very important because we must ensure we
|
|
|
|
* don't lose the VRSAVE content that may have been set prior to
|
|
|
|
* the process doing its first vector operation
|
2012-09-20 09:48:00 +08:00
|
|
|
* Userland shall check AT_HWCAP to know whether it can rely on the
|
2005-04-17 06:20:36 +08:00
|
|
|
* v_regs pointer or not
|
|
|
|
*/
|
|
|
|
#ifdef CONFIG_ALTIVEC
|
2015-07-20 10:58:43 +08:00
|
|
|
elf_vrreg_t __user *v_regs = sigcontext_vmx_regs(sc);
|
2016-05-29 20:03:51 +08:00
|
|
|
unsigned long vrsave;
|
2005-04-17 06:20:36 +08:00
|
|
|
#endif
|
2016-09-23 14:18:12 +08:00
|
|
|
struct pt_regs *regs = tsk->thread.regs;
|
2008-06-02 14:22:59 +08:00
|
|
|
unsigned long msr = regs->msr;
|
2005-04-17 06:20:36 +08:00
|
|
|
long err = 0;
|
2017-08-21 01:58:24 +08:00
|
|
|
/* Force usr to alway see softe as 1 (interrupts enabled) */
|
|
|
|
unsigned long softe = 0x1;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2016-09-23 14:18:12 +08:00
|
|
|
BUG_ON(tsk != current);
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
#ifdef CONFIG_ALTIVEC
|
|
|
|
err |= __put_user(v_regs, &sc->v_regs);
|
|
|
|
|
|
|
|
/* save altivec registers */
|
2016-09-23 14:18:12 +08:00
|
|
|
if (tsk->thread.used_vr) {
|
|
|
|
flush_altivec_to_thread(tsk);
|
2005-04-17 06:20:36 +08:00
|
|
|
/* Copy 33 vec registers (vr0..31 and vscr) to the stack */
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= __copy_to_user(v_regs, &tsk->thread.vr_state,
|
2013-09-10 18:20:42 +08:00
|
|
|
33 * sizeof(vector128));
|
2005-04-17 06:20:36 +08:00
|
|
|
/* set MSR_VEC in the MSR value in the frame to indicate that sc->v_reg)
|
|
|
|
* contains valid data.
|
|
|
|
*/
|
2008-06-02 14:22:59 +08:00
|
|
|
msr |= MSR_VEC;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
/* We always copy to/from vrsave, it's 0 if we don't have or don't
|
|
|
|
* use altivec.
|
|
|
|
*/
|
2016-05-29 20:03:51 +08:00
|
|
|
vrsave = 0;
|
|
|
|
if (cpu_has_feature(CPU_FTR_ALTIVEC)) {
|
|
|
|
vrsave = mfspr(SPRN_VRSAVE);
|
2016-09-23 14:18:12 +08:00
|
|
|
tsk->thread.vrsave = vrsave;
|
2016-05-29 20:03:51 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
err |= __put_user(vrsave, (u32 __user *)&v_regs[33]);
|
2005-04-17 06:20:36 +08:00
|
|
|
#else /* CONFIG_ALTIVEC */
|
|
|
|
err |= __put_user(0, &sc->v_regs);
|
|
|
|
#endif /* CONFIG_ALTIVEC */
|
2016-09-23 14:18:12 +08:00
|
|
|
flush_fp_to_thread(tsk);
|
2008-07-02 12:06:37 +08:00
|
|
|
/* copy fpr regs and fpscr */
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= copy_fpr_to_user(&sc->fp_regs, tsk);
|
2013-11-25 08:12:20 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Clear the MSR VSX bit to indicate there is no valid state attached
|
|
|
|
* to this context, except in the specific case below where we set it.
|
|
|
|
*/
|
|
|
|
msr &= ~MSR_VSX;
|
2008-06-25 12:07:18 +08:00
|
|
|
#ifdef CONFIG_VSX
|
2008-06-25 12:07:18 +08:00
|
|
|
/*
|
|
|
|
* Copy VSX low doubleword to local buffer for formatting,
|
|
|
|
* then out to userspace. Update v_regs to point after the
|
|
|
|
* VMX data.
|
|
|
|
*/
|
2016-09-23 14:18:12 +08:00
|
|
|
if (tsk->thread.used_vsr && ctx_has_vsx_region) {
|
|
|
|
flush_vsx_to_thread(tsk);
|
2008-06-25 12:07:18 +08:00
|
|
|
v_regs += ELF_NVRREG;
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= copy_vsx_to_user(v_regs, tsk);
|
2008-06-25 12:07:18 +08:00
|
|
|
/* set MSR_VSX in the MSR value in the frame to
|
|
|
|
* indicate that sc->vs_reg) contains valid data.
|
|
|
|
*/
|
|
|
|
msr |= MSR_VSX;
|
|
|
|
}
|
2008-06-25 12:07:18 +08:00
|
|
|
#endif /* CONFIG_VSX */
|
2005-04-17 06:20:36 +08:00
|
|
|
err |= __put_user(&sc->gp_regs, &sc->regs);
|
2006-03-08 10:24:22 +08:00
|
|
|
WARN_ON(!FULL_REGS(regs));
|
2005-04-17 06:20:36 +08:00
|
|
|
err |= __copy_to_user(&sc->gp_regs, regs, GP_REGS_SIZE);
|
2008-06-02 14:22:59 +08:00
|
|
|
err |= __put_user(msr, &sc->gp_regs[PT_MSR]);
|
2017-08-21 01:58:24 +08:00
|
|
|
err |= __put_user(softe, &sc->gp_regs[PT_SOFTE]);
|
2005-04-17 06:20:36 +08:00
|
|
|
err |= __put_user(signr, &sc->signal);
|
|
|
|
err |= __put_user(handler, &sc->handler);
|
|
|
|
if (set != NULL)
|
|
|
|
err |= __put_user(set->sig[0], &sc->oldmask);
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2013-02-14 00:21:41 +08:00
|
|
|
#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
|
|
|
|
/*
|
|
|
|
* As above, but Transactional Memory is in use, so deliver sigcontexts
|
|
|
|
* containing checkpointed and transactional register states.
|
|
|
|
*
|
2013-05-27 02:09:41 +08:00
|
|
|
* To do this, we treclaim (done before entering here) to gather both sets of
|
|
|
|
* registers and set up the 'normal' sigcontext registers with rolled-back
|
|
|
|
* register values such that a simple signal handler sees a correct
|
|
|
|
* checkpointed register state. If interested, a TM-aware sighandler can
|
|
|
|
* examine the transactional registers in the 2nd sigcontext to determine the
|
|
|
|
* real origin of the signal.
|
2013-02-14 00:21:41 +08:00
|
|
|
*/
|
|
|
|
static long setup_tm_sigcontexts(struct sigcontext __user *sc,
|
|
|
|
struct sigcontext __user *tm_sc,
|
2016-09-23 14:18:12 +08:00
|
|
|
struct task_struct *tsk,
|
2013-02-14 00:21:41 +08:00
|
|
|
int signr, sigset_t *set, unsigned long handler)
|
|
|
|
{
|
|
|
|
/* When CONFIG_ALTIVEC is set, we _always_ setup v_regs even if the
|
|
|
|
* process never used altivec yet (MSR_VEC is zero in pt_regs of
|
|
|
|
* the context). This is very important because we must ensure we
|
|
|
|
* don't lose the VRSAVE content that may have been set prior to
|
|
|
|
* the process doing its first vector operation
|
|
|
|
* Userland shall check AT_HWCAP to know wether it can rely on the
|
|
|
|
* v_regs pointer or not.
|
|
|
|
*/
|
|
|
|
#ifdef CONFIG_ALTIVEC
|
2015-07-20 10:58:43 +08:00
|
|
|
elf_vrreg_t __user *v_regs = sigcontext_vmx_regs(sc);
|
|
|
|
elf_vrreg_t __user *tm_v_regs = sigcontext_vmx_regs(tm_sc);
|
2013-02-14 00:21:41 +08:00
|
|
|
#endif
|
2016-09-23 14:18:12 +08:00
|
|
|
struct pt_regs *regs = tsk->thread.regs;
|
2018-01-01 07:20:45 +08:00
|
|
|
unsigned long msr = tsk->thread.regs->msr;
|
2013-02-14 00:21:41 +08:00
|
|
|
long err = 0;
|
|
|
|
|
2016-09-23 14:18:12 +08:00
|
|
|
BUG_ON(tsk != current);
|
|
|
|
|
2013-02-14 00:21:41 +08:00
|
|
|
BUG_ON(!MSR_TM_ACTIVE(regs->msr));
|
|
|
|
|
2017-10-12 18:17:19 +08:00
|
|
|
WARN_ON(tm_suspend_disabled);
|
|
|
|
|
2018-01-01 07:20:45 +08:00
|
|
|
/* Restore checkpointed FP, VEC, and VSX bits from ckpt_regs as
|
|
|
|
* it contains the correct FP, VEC, VSX state after we treclaimed
|
|
|
|
* the transaction and giveup_all() was called on reclaiming.
|
|
|
|
*/
|
|
|
|
msr |= tsk->thread.ckpt_regs.msr & (MSR_FP | MSR_VEC | MSR_VSX);
|
|
|
|
|
powerpc: Don't corrupt transactional state when using FP/VMX in kernel
Currently, when we have a process using the transactional memory
facilities on POWER8 (that is, the processor is in transactional
or suspended state), and the process enters the kernel and the
kernel then uses the floating-point or vector (VMX/Altivec) facility,
we end up corrupting the user-visible FP/VMX/VSX state. This
happens, for example, if a page fault causes a copy-on-write
operation, because the copy_page function will use VMX to do the
copy on POWER8. The test program below demonstrates the bug.
The bug happens because when FP/VMX state for a transactional process
is stored in the thread_struct, we store the checkpointed state in
.fp_state/.vr_state and the transactional (current) state in
.transact_fp/.transact_vr. However, when the kernel wants to use
FP/VMX, it calls enable_kernel_fp() or enable_kernel_altivec(),
which saves the current state in .fp_state/.vr_state. Furthermore,
when we return to the user process we return with FP/VMX/VSX
disabled. The next time the process uses FP/VMX/VSX, we don't know
which set of state (the current register values, .fp_state/.vr_state,
or .transact_fp/.transact_vr) we should be using, since we have no
way to tell if we are still in the same transaction, and if not,
whether the previous transaction succeeded or failed.
Thus it is necessary to strictly adhere to the rule that if FP has
been enabled at any point in a transaction, we must keep FP enabled
for the user process with the current transactional state in the
FP registers, until we detect that it is no longer in a transaction.
Similarly for VMX; once enabled it must stay enabled until the
process is no longer transactional.
In order to keep this rule, we add a new thread_info flag which we
test when returning from the kernel to userspace, called TIF_RESTORE_TM.
This flag indicates that there is FP/VMX/VSX state to be restored
before entering userspace, and when it is set the .tm_orig_msr field
in the thread_struct indicates what state needs to be restored.
The restoration is done by restore_tm_state(). The TIF_RESTORE_TM
bit is set by new giveup_fpu/altivec_maybe_transactional helpers,
which are called from enable_kernel_fp/altivec, giveup_vsx, and
flush_fp/altivec_to_thread instead of giveup_fpu/altivec.
The other thing to be done is to get the transactional FP/VMX/VSX
state from .fp_state/.vr_state when doing reclaim, if that state
has been saved there by giveup_fpu/altivec_maybe_transactional.
Having done this, we set the FP/VMX bit in the thread's MSR after
reclaim to indicate that that part of the state is now valid
(having been reclaimed from the processor's checkpointed state).
Finally, in the signal handling code, we move the clearing of the
transactional state bits in the thread's MSR a bit earlier, before
calling flush_fp_to_thread(), so that we don't unnecessarily set
the TIF_RESTORE_TM bit.
This is the test program:
/* Michael Neuling 4/12/2013
*
* See if the altivec state is leaked out of an aborted transaction due to
* kernel vmx copy loops.
*
* gcc -m64 htm_vmxcopy.c -o htm_vmxcopy
*
*/
/* We don't use all of these, but for reference: */
int main(int argc, char *argv[])
{
long double vecin = 1.3;
long double vecout;
unsigned long pgsize = getpagesize();
int i;
int fd;
int size = pgsize*16;
char tmpfile[] = "/tmp/page_faultXXXXXX";
char buf[pgsize];
char *a;
uint64_t aborted = 0;
fd = mkstemp(tmpfile);
assert(fd >= 0);
memset(buf, 0, pgsize);
for (i = 0; i < size; i += pgsize)
assert(write(fd, buf, pgsize) == pgsize);
unlink(tmpfile);
a = mmap(NULL, size, PROT_READ|PROT_WRITE, MAP_PRIVATE, fd, 0);
assert(a != MAP_FAILED);
asm __volatile__(
"lxvd2x 40,0,%[vecinptr] ; " // set 40 to initial value
TBEGIN
"beq 3f ;"
TSUSPEND
"xxlxor 40,40,40 ; " // set 40 to 0
"std 5, 0(%[map]) ;" // cause kernel vmx copy page
TABORT
TRESUME
TEND
"li %[res], 0 ;"
"b 5f ;"
"3: ;" // Abort handler
"li %[res], 1 ;"
"5: ;"
"stxvd2x 40,0,%[vecoutptr] ; "
: [res]"=r"(aborted)
: [vecinptr]"r"(&vecin),
[vecoutptr]"r"(&vecout),
[map]"r"(a)
: "memory", "r0", "r3", "r4", "r5", "r6", "r7");
if (aborted && (vecin != vecout)){
printf("FAILED: vector state leaked on abort %f != %f\n",
(double)vecin, (double)vecout);
exit(1);
}
munmap(a, size);
close(fd);
printf("PASSED!\n");
return 0;
}
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2014-01-13 12:56:29 +08:00
|
|
|
/* Remove TM bits from thread's MSR. The MSR in the sigcontext
|
|
|
|
* just indicates to userland that we were doing a transaction, but we
|
|
|
|
* don't want to return in transactional state. This also ensures
|
|
|
|
* that flush_fp_to_thread won't set TIF_RESTORE_TM again.
|
|
|
|
*/
|
|
|
|
regs->msr &= ~MSR_TS_MASK;
|
|
|
|
|
2013-02-14 00:21:41 +08:00
|
|
|
#ifdef CONFIG_ALTIVEC
|
|
|
|
err |= __put_user(v_regs, &sc->v_regs);
|
|
|
|
err |= __put_user(tm_v_regs, &tm_sc->v_regs);
|
|
|
|
|
|
|
|
/* save altivec registers */
|
2016-09-23 14:18:12 +08:00
|
|
|
if (tsk->thread.used_vr) {
|
2013-02-14 00:21:41 +08:00
|
|
|
/* Copy 33 vec registers (vr0..31 and vscr) to the stack */
|
2016-09-23 14:18:25 +08:00
|
|
|
err |= __copy_to_user(v_regs, &tsk->thread.ckvr_state,
|
2013-02-14 00:21:41 +08:00
|
|
|
33 * sizeof(vector128));
|
|
|
|
/* If VEC was enabled there are transactional VRs valid too,
|
|
|
|
* else they're a copy of the checkpointed VRs.
|
|
|
|
*/
|
|
|
|
if (msr & MSR_VEC)
|
|
|
|
err |= __copy_to_user(tm_v_regs,
|
2016-09-23 14:18:24 +08:00
|
|
|
&tsk->thread.vr_state,
|
2013-02-14 00:21:41 +08:00
|
|
|
33 * sizeof(vector128));
|
|
|
|
else
|
|
|
|
err |= __copy_to_user(tm_v_regs,
|
2016-09-23 14:18:25 +08:00
|
|
|
&tsk->thread.ckvr_state,
|
2013-02-14 00:21:41 +08:00
|
|
|
33 * sizeof(vector128));
|
|
|
|
|
|
|
|
/* set MSR_VEC in the MSR value in the frame to indicate
|
|
|
|
* that sc->v_reg contains valid data.
|
|
|
|
*/
|
|
|
|
msr |= MSR_VEC;
|
|
|
|
}
|
|
|
|
/* We always copy to/from vrsave, it's 0 if we don't have or don't
|
|
|
|
* use altivec.
|
|
|
|
*/
|
2013-08-05 12:13:16 +08:00
|
|
|
if (cpu_has_feature(CPU_FTR_ALTIVEC))
|
2016-09-23 14:18:25 +08:00
|
|
|
tsk->thread.ckvrsave = mfspr(SPRN_VRSAVE);
|
|
|
|
err |= __put_user(tsk->thread.ckvrsave, (u32 __user *)&v_regs[33]);
|
2013-02-14 00:21:41 +08:00
|
|
|
if (msr & MSR_VEC)
|
2016-09-23 14:18:24 +08:00
|
|
|
err |= __put_user(tsk->thread.vrsave,
|
2013-02-14 00:21:41 +08:00
|
|
|
(u32 __user *)&tm_v_regs[33]);
|
|
|
|
else
|
2016-09-23 14:18:25 +08:00
|
|
|
err |= __put_user(tsk->thread.ckvrsave,
|
2013-02-14 00:21:41 +08:00
|
|
|
(u32 __user *)&tm_v_regs[33]);
|
|
|
|
|
|
|
|
#else /* CONFIG_ALTIVEC */
|
|
|
|
err |= __put_user(0, &sc->v_regs);
|
|
|
|
err |= __put_user(0, &tm_sc->v_regs);
|
|
|
|
#endif /* CONFIG_ALTIVEC */
|
|
|
|
|
|
|
|
/* copy fpr regs and fpscr */
|
2016-09-23 14:18:25 +08:00
|
|
|
err |= copy_ckfpr_to_user(&sc->fp_regs, tsk);
|
2013-02-14 00:21:41 +08:00
|
|
|
if (msr & MSR_FP)
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= copy_fpr_to_user(&tm_sc->fp_regs, tsk);
|
2016-09-23 14:18:24 +08:00
|
|
|
else
|
2016-09-23 14:18:25 +08:00
|
|
|
err |= copy_ckfpr_to_user(&tm_sc->fp_regs, tsk);
|
2013-02-14 00:21:41 +08:00
|
|
|
|
|
|
|
#ifdef CONFIG_VSX
|
|
|
|
/*
|
|
|
|
* Copy VSX low doubleword to local buffer for formatting,
|
|
|
|
* then out to userspace. Update v_regs to point after the
|
|
|
|
* VMX data.
|
|
|
|
*/
|
2016-09-23 14:18:12 +08:00
|
|
|
if (tsk->thread.used_vsr) {
|
2013-02-14 00:21:41 +08:00
|
|
|
v_regs += ELF_NVRREG;
|
|
|
|
tm_v_regs += ELF_NVRREG;
|
|
|
|
|
2016-09-23 14:18:25 +08:00
|
|
|
err |= copy_ckvsx_to_user(v_regs, tsk);
|
2013-02-14 00:21:41 +08:00
|
|
|
|
|
|
|
if (msr & MSR_VSX)
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= copy_vsx_to_user(tm_v_regs, tsk);
|
2016-09-23 14:18:24 +08:00
|
|
|
else
|
2016-09-23 14:18:25 +08:00
|
|
|
err |= copy_ckvsx_to_user(tm_v_regs, tsk);
|
2013-02-14 00:21:41 +08:00
|
|
|
|
|
|
|
/* set MSR_VSX in the MSR value in the frame to
|
|
|
|
* indicate that sc->vs_reg) contains valid data.
|
|
|
|
*/
|
|
|
|
msr |= MSR_VSX;
|
|
|
|
}
|
|
|
|
#endif /* CONFIG_VSX */
|
|
|
|
|
|
|
|
err |= __put_user(&sc->gp_regs, &sc->regs);
|
|
|
|
err |= __put_user(&tm_sc->gp_regs, &tm_sc->regs);
|
|
|
|
WARN_ON(!FULL_REGS(regs));
|
|
|
|
err |= __copy_to_user(&tm_sc->gp_regs, regs, GP_REGS_SIZE);
|
|
|
|
err |= __copy_to_user(&sc->gp_regs,
|
2016-09-23 14:18:12 +08:00
|
|
|
&tsk->thread.ckpt_regs, GP_REGS_SIZE);
|
2013-02-14 00:21:41 +08:00
|
|
|
err |= __put_user(msr, &tm_sc->gp_regs[PT_MSR]);
|
|
|
|
err |= __put_user(msr, &sc->gp_regs[PT_MSR]);
|
|
|
|
err |= __put_user(signr, &sc->signal);
|
|
|
|
err |= __put_user(handler, &sc->handler);
|
|
|
|
if (set != NULL)
|
|
|
|
err |= __put_user(set->sig[0], &sc->oldmask);
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* Restore the sigcontext from the signal frame.
|
|
|
|
*/
|
|
|
|
|
2016-09-23 14:18:12 +08:00
|
|
|
static long restore_sigcontext(struct task_struct *tsk, sigset_t *set, int sig,
|
2005-04-17 06:20:36 +08:00
|
|
|
struct sigcontext __user *sc)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_ALTIVEC
|
|
|
|
elf_vrreg_t __user *v_regs;
|
|
|
|
#endif
|
|
|
|
unsigned long err = 0;
|
|
|
|
unsigned long save_r13 = 0;
|
|
|
|
unsigned long msr;
|
2016-09-23 14:18:12 +08:00
|
|
|
struct pt_regs *regs = tsk->thread.regs;
|
2008-07-02 12:06:37 +08:00
|
|
|
#ifdef CONFIG_VSX
|
|
|
|
int i;
|
|
|
|
#endif
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2016-09-23 14:18:12 +08:00
|
|
|
BUG_ON(tsk != current);
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/* If this is not a signal return, we preserve the TLS in r13 */
|
|
|
|
if (!sig)
|
|
|
|
save_r13 = regs->gpr[13];
|
|
|
|
|
2008-06-27 14:18:27 +08:00
|
|
|
/* copy the GPRs */
|
|
|
|
err |= __copy_from_user(regs->gpr, sc->gp_regs, sizeof(regs->gpr));
|
|
|
|
err |= __get_user(regs->nip, &sc->gp_regs[PT_NIP]);
|
2006-06-07 14:14:40 +08:00
|
|
|
/* get MSR separately, transfer the LE bit if doing signal return */
|
|
|
|
err |= __get_user(msr, &sc->gp_regs[PT_MSR]);
|
|
|
|
if (sig)
|
|
|
|
regs->msr = (regs->msr & ~MSR_LE) | (msr & MSR_LE);
|
2008-06-27 14:18:27 +08:00
|
|
|
err |= __get_user(regs->orig_gpr3, &sc->gp_regs[PT_ORIG_R3]);
|
|
|
|
err |= __get_user(regs->ctr, &sc->gp_regs[PT_CTR]);
|
|
|
|
err |= __get_user(regs->link, &sc->gp_regs[PT_LNK]);
|
|
|
|
err |= __get_user(regs->xer, &sc->gp_regs[PT_XER]);
|
|
|
|
err |= __get_user(regs->ccr, &sc->gp_regs[PT_CCR]);
|
2006-06-07 14:14:40 +08:00
|
|
|
/* skip SOFTE */
|
2010-09-21 04:48:57 +08:00
|
|
|
regs->trap = 0;
|
2008-06-27 14:18:27 +08:00
|
|
|
err |= __get_user(regs->dar, &sc->gp_regs[PT_DAR]);
|
|
|
|
err |= __get_user(regs->dsisr, &sc->gp_regs[PT_DSISR]);
|
|
|
|
err |= __get_user(regs->result, &sc->gp_regs[PT_RESULT]);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
if (!sig)
|
|
|
|
regs->gpr[13] = save_r13;
|
|
|
|
if (set != NULL)
|
|
|
|
err |= __get_user(set->sig[0], &sc->oldmask);
|
|
|
|
|
2007-06-26 12:49:11 +08:00
|
|
|
/*
|
|
|
|
* Force reload of FP/VEC.
|
2016-09-23 14:18:12 +08:00
|
|
|
* This has to be done before copying stuff into tsk->thread.fpr/vr
|
2007-06-26 12:49:11 +08:00
|
|
|
* for the reasons explained in the previous comment.
|
|
|
|
*/
|
2008-06-25 12:07:18 +08:00
|
|
|
regs->msr &= ~(MSR_FP | MSR_FE0 | MSR_FE1 | MSR_VEC | MSR_VSX);
|
2007-06-26 12:49:11 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
#ifdef CONFIG_ALTIVEC
|
|
|
|
err |= __get_user(v_regs, &sc->v_regs);
|
|
|
|
if (err)
|
|
|
|
return err;
|
Remove 'type' argument from access_ok() function
Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument
of the user address range verification function since we got rid of the
old racy i386-only code to walk page tables by hand.
It existed because the original 80386 would not honor the write protect
bit when in kernel mode, so you had to do COW by hand before doing any
user access. But we haven't supported that in a long time, and these
days the 'type' argument is a purely historical artifact.
A discussion about extending 'user_access_begin()' to do the range
checking resulted this patch, because there is no way we're going to
move the old VERIFY_xyz interface to that model. And it's best done at
the end of the merge window when I've done most of my merges, so let's
just get this done once and for all.
This patch was mostly done with a sed-script, with manual fix-ups for
the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form.
There were a couple of notable cases:
- csky still had the old "verify_area()" name as an alias.
- the iter_iov code had magical hardcoded knowledge of the actual
values of VERIFY_{READ,WRITE} (not that they mattered, since nothing
really used it)
- microblaze used the type argument for a debug printout
but other than those oddities this should be a total no-op patch.
I tried to fix up all architectures, did fairly extensive grepping for
access_ok() uses, and the changes are trivial, but I may have missed
something. Any missed conversion should be trivially fixable, though.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-01-04 10:57:57 +08:00
|
|
|
if (v_regs && !access_ok(v_regs, 34 * sizeof(vector128)))
|
2006-06-09 11:02:59 +08:00
|
|
|
return -EFAULT;
|
2005-04-17 06:20:36 +08:00
|
|
|
/* Copy 33 vec registers (vr0..31 and vscr) from the stack */
|
2016-07-26 16:06:01 +08:00
|
|
|
if (v_regs != NULL && (msr & MSR_VEC) != 0) {
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= __copy_from_user(&tsk->thread.vr_state, v_regs,
|
2005-04-17 06:20:36 +08:00
|
|
|
33 * sizeof(vector128));
|
2016-09-23 14:18:12 +08:00
|
|
|
tsk->thread.used_vr = true;
|
|
|
|
} else if (tsk->thread.used_vr) {
|
|
|
|
memset(&tsk->thread.vr_state, 0, 33 * sizeof(vector128));
|
2016-07-26 16:06:01 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
/* Always get VRSAVE back */
|
2013-08-07 00:01:24 +08:00
|
|
|
if (v_regs != NULL)
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= __get_user(tsk->thread.vrsave, (u32 __user *)&v_regs[33]);
|
2005-04-17 06:20:36 +08:00
|
|
|
else
|
2016-09-23 14:18:12 +08:00
|
|
|
tsk->thread.vrsave = 0;
|
2013-08-05 12:13:16 +08:00
|
|
|
if (cpu_has_feature(CPU_FTR_ALTIVEC))
|
2016-09-23 14:18:12 +08:00
|
|
|
mtspr(SPRN_VRSAVE, tsk->thread.vrsave);
|
2005-04-17 06:20:36 +08:00
|
|
|
#endif /* CONFIG_ALTIVEC */
|
2008-06-25 12:07:18 +08:00
|
|
|
/* restore floating point */
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= copy_fpr_from_user(tsk, &sc->fp_regs);
|
2008-07-02 12:06:37 +08:00
|
|
|
#ifdef CONFIG_VSX
|
2008-06-25 12:07:18 +08:00
|
|
|
/*
|
|
|
|
* Get additional VSX data. Update v_regs to point after the
|
|
|
|
* VMX data. Copy VSX low doubleword from userspace to local
|
|
|
|
* buffer for formatting, then into the taskstruct.
|
|
|
|
*/
|
|
|
|
v_regs += ELF_NVRREG;
|
2016-07-26 16:06:01 +08:00
|
|
|
if ((msr & MSR_VSX) != 0) {
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= copy_vsx_from_user(tsk, v_regs);
|
|
|
|
tsk->thread.used_vsr = true;
|
|
|
|
} else {
|
2008-07-02 12:06:37 +08:00
|
|
|
for (i = 0; i < 32 ; i++)
|
2016-09-23 14:18:12 +08:00
|
|
|
tsk->thread.fp_state.fpr[i][TS_VSRLOWOFFSET] = 0;
|
|
|
|
}
|
2008-06-25 12:07:18 +08:00
|
|
|
#endif
|
2005-04-17 06:20:36 +08:00
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2013-02-14 00:21:41 +08:00
|
|
|
#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
|
|
|
|
/*
|
|
|
|
* Restore the two sigcontexts from the frame of a transactional processes.
|
|
|
|
*/
|
|
|
|
|
2016-09-23 14:18:12 +08:00
|
|
|
static long restore_tm_sigcontexts(struct task_struct *tsk,
|
2013-02-14 00:21:41 +08:00
|
|
|
struct sigcontext __user *sc,
|
|
|
|
struct sigcontext __user *tm_sc)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_ALTIVEC
|
|
|
|
elf_vrreg_t __user *v_regs, *tm_v_regs;
|
|
|
|
#endif
|
|
|
|
unsigned long err = 0;
|
|
|
|
unsigned long msr;
|
2016-09-23 14:18:12 +08:00
|
|
|
struct pt_regs *regs = tsk->thread.regs;
|
2013-02-14 00:21:41 +08:00
|
|
|
#ifdef CONFIG_VSX
|
|
|
|
int i;
|
|
|
|
#endif
|
2016-09-23 14:18:12 +08:00
|
|
|
|
|
|
|
BUG_ON(tsk != current);
|
|
|
|
|
2017-10-12 18:17:19 +08:00
|
|
|
if (tm_suspend_disabled)
|
|
|
|
return -EINVAL;
|
|
|
|
|
2013-02-14 00:21:41 +08:00
|
|
|
/* copy the GPRs */
|
|
|
|
err |= __copy_from_user(regs->gpr, tm_sc->gp_regs, sizeof(regs->gpr));
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= __copy_from_user(&tsk->thread.ckpt_regs, sc->gp_regs,
|
2013-02-14 00:21:41 +08:00
|
|
|
sizeof(regs->gpr));
|
|
|
|
|
|
|
|
/*
|
|
|
|
* TFHAR is restored from the checkpointed 'wound-back' ucontext's NIP.
|
|
|
|
* TEXASR was set by the signal delivery reclaim, as was TFIAR.
|
|
|
|
* Users doing anything abhorrent like thread-switching w/ signals for
|
|
|
|
* TM-Suspended code will have to back TEXASR/TFIAR up themselves.
|
|
|
|
* For the case of getting a signal and simply returning from it,
|
|
|
|
* we don't need to re-copy them here.
|
|
|
|
*/
|
|
|
|
err |= __get_user(regs->nip, &tm_sc->gp_regs[PT_NIP]);
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= __get_user(tsk->thread.tm_tfhar, &sc->gp_regs[PT_NIP]);
|
2013-02-14 00:21:41 +08:00
|
|
|
|
|
|
|
/* get MSR separately, transfer the LE bit if doing signal return */
|
|
|
|
err |= __get_user(msr, &sc->gp_regs[PT_MSR]);
|
2015-11-19 12:44:44 +08:00
|
|
|
/* Don't allow reserved mode. */
|
|
|
|
if (MSR_TM_RESV(msr))
|
|
|
|
return -EINVAL;
|
|
|
|
|
2013-06-09 19:23:19 +08:00
|
|
|
/* pull in MSR LE from user context */
|
2013-02-14 00:21:41 +08:00
|
|
|
regs->msr = (regs->msr & ~MSR_LE) | (msr & MSR_LE);
|
|
|
|
|
|
|
|
/* The following non-GPR non-FPR non-VR state is also checkpointed: */
|
|
|
|
err |= __get_user(regs->ctr, &tm_sc->gp_regs[PT_CTR]);
|
|
|
|
err |= __get_user(regs->link, &tm_sc->gp_regs[PT_LNK]);
|
|
|
|
err |= __get_user(regs->xer, &tm_sc->gp_regs[PT_XER]);
|
|
|
|
err |= __get_user(regs->ccr, &tm_sc->gp_regs[PT_CCR]);
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= __get_user(tsk->thread.ckpt_regs.ctr,
|
2013-02-14 00:21:41 +08:00
|
|
|
&sc->gp_regs[PT_CTR]);
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= __get_user(tsk->thread.ckpt_regs.link,
|
2013-02-14 00:21:41 +08:00
|
|
|
&sc->gp_regs[PT_LNK]);
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= __get_user(tsk->thread.ckpt_regs.xer,
|
2013-02-14 00:21:41 +08:00
|
|
|
&sc->gp_regs[PT_XER]);
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= __get_user(tsk->thread.ckpt_regs.ccr,
|
2013-02-14 00:21:41 +08:00
|
|
|
&sc->gp_regs[PT_CCR]);
|
|
|
|
|
|
|
|
/* These regs are not checkpointed; they can go in 'regs'. */
|
|
|
|
err |= __get_user(regs->trap, &sc->gp_regs[PT_TRAP]);
|
|
|
|
err |= __get_user(regs->dar, &sc->gp_regs[PT_DAR]);
|
|
|
|
err |= __get_user(regs->dsisr, &sc->gp_regs[PT_DSISR]);
|
|
|
|
err |= __get_user(regs->result, &sc->gp_regs[PT_RESULT]);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Force reload of FP/VEC.
|
2016-09-23 14:18:12 +08:00
|
|
|
* This has to be done before copying stuff into tsk->thread.fpr/vr
|
2013-02-14 00:21:41 +08:00
|
|
|
* for the reasons explained in the previous comment.
|
|
|
|
*/
|
|
|
|
regs->msr &= ~(MSR_FP | MSR_FE0 | MSR_FE1 | MSR_VEC | MSR_VSX);
|
|
|
|
|
|
|
|
#ifdef CONFIG_ALTIVEC
|
|
|
|
err |= __get_user(v_regs, &sc->v_regs);
|
|
|
|
err |= __get_user(tm_v_regs, &tm_sc->v_regs);
|
|
|
|
if (err)
|
|
|
|
return err;
|
Remove 'type' argument from access_ok() function
Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument
of the user address range verification function since we got rid of the
old racy i386-only code to walk page tables by hand.
It existed because the original 80386 would not honor the write protect
bit when in kernel mode, so you had to do COW by hand before doing any
user access. But we haven't supported that in a long time, and these
days the 'type' argument is a purely historical artifact.
A discussion about extending 'user_access_begin()' to do the range
checking resulted this patch, because there is no way we're going to
move the old VERIFY_xyz interface to that model. And it's best done at
the end of the merge window when I've done most of my merges, so let's
just get this done once and for all.
This patch was mostly done with a sed-script, with manual fix-ups for
the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form.
There were a couple of notable cases:
- csky still had the old "verify_area()" name as an alias.
- the iter_iov code had magical hardcoded knowledge of the actual
values of VERIFY_{READ,WRITE} (not that they mattered, since nothing
really used it)
- microblaze used the type argument for a debug printout
but other than those oddities this should be a total no-op patch.
I tried to fix up all architectures, did fairly extensive grepping for
access_ok() uses, and the changes are trivial, but I may have missed
something. Any missed conversion should be trivially fixable, though.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-01-04 10:57:57 +08:00
|
|
|
if (v_regs && !access_ok(v_regs, 34 * sizeof(vector128)))
|
2013-02-14 00:21:41 +08:00
|
|
|
return -EFAULT;
|
Remove 'type' argument from access_ok() function
Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument
of the user address range verification function since we got rid of the
old racy i386-only code to walk page tables by hand.
It existed because the original 80386 would not honor the write protect
bit when in kernel mode, so you had to do COW by hand before doing any
user access. But we haven't supported that in a long time, and these
days the 'type' argument is a purely historical artifact.
A discussion about extending 'user_access_begin()' to do the range
checking resulted this patch, because there is no way we're going to
move the old VERIFY_xyz interface to that model. And it's best done at
the end of the merge window when I've done most of my merges, so let's
just get this done once and for all.
This patch was mostly done with a sed-script, with manual fix-ups for
the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form.
There were a couple of notable cases:
- csky still had the old "verify_area()" name as an alias.
- the iter_iov code had magical hardcoded knowledge of the actual
values of VERIFY_{READ,WRITE} (not that they mattered, since nothing
really used it)
- microblaze used the type argument for a debug printout
but other than those oddities this should be a total no-op patch.
I tried to fix up all architectures, did fairly extensive grepping for
access_ok() uses, and the changes are trivial, but I may have missed
something. Any missed conversion should be trivially fixable, though.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-01-04 10:57:57 +08:00
|
|
|
if (tm_v_regs && !access_ok(tm_v_regs, 34 * sizeof(vector128)))
|
2013-02-14 00:21:41 +08:00
|
|
|
return -EFAULT;
|
|
|
|
/* Copy 33 vec registers (vr0..31 and vscr) from the stack */
|
2013-08-07 00:01:24 +08:00
|
|
|
if (v_regs != NULL && tm_v_regs != NULL && (msr & MSR_VEC) != 0) {
|
2016-09-23 14:18:25 +08:00
|
|
|
err |= __copy_from_user(&tsk->thread.ckvr_state, v_regs,
|
2013-02-14 00:21:41 +08:00
|
|
|
33 * sizeof(vector128));
|
2016-09-23 14:18:24 +08:00
|
|
|
err |= __copy_from_user(&tsk->thread.vr_state, tm_v_regs,
|
2013-02-14 00:21:41 +08:00
|
|
|
33 * sizeof(vector128));
|
2016-07-26 16:06:01 +08:00
|
|
|
current->thread.used_vr = true;
|
2013-02-14 00:21:41 +08:00
|
|
|
}
|
2016-09-23 14:18:12 +08:00
|
|
|
else if (tsk->thread.used_vr) {
|
|
|
|
memset(&tsk->thread.vr_state, 0, 33 * sizeof(vector128));
|
2016-09-23 14:18:25 +08:00
|
|
|
memset(&tsk->thread.ckvr_state, 0, 33 * sizeof(vector128));
|
2013-02-14 00:21:41 +08:00
|
|
|
}
|
|
|
|
/* Always get VRSAVE back */
|
2013-08-07 00:01:24 +08:00
|
|
|
if (v_regs != NULL && tm_v_regs != NULL) {
|
2016-09-23 14:18:25 +08:00
|
|
|
err |= __get_user(tsk->thread.ckvrsave,
|
2016-09-23 14:18:24 +08:00
|
|
|
(u32 __user *)&v_regs[33]);
|
|
|
|
err |= __get_user(tsk->thread.vrsave,
|
2013-02-14 00:21:41 +08:00
|
|
|
(u32 __user *)&tm_v_regs[33]);
|
|
|
|
}
|
|
|
|
else {
|
2016-09-23 14:18:12 +08:00
|
|
|
tsk->thread.vrsave = 0;
|
2016-09-23 14:18:25 +08:00
|
|
|
tsk->thread.ckvrsave = 0;
|
2013-02-14 00:21:41 +08:00
|
|
|
}
|
2013-08-05 12:13:16 +08:00
|
|
|
if (cpu_has_feature(CPU_FTR_ALTIVEC))
|
2016-09-23 14:18:12 +08:00
|
|
|
mtspr(SPRN_VRSAVE, tsk->thread.vrsave);
|
2013-02-14 00:21:41 +08:00
|
|
|
#endif /* CONFIG_ALTIVEC */
|
|
|
|
/* restore floating point */
|
2016-09-23 14:18:24 +08:00
|
|
|
err |= copy_fpr_from_user(tsk, &tm_sc->fp_regs);
|
2016-09-23 14:18:25 +08:00
|
|
|
err |= copy_ckfpr_from_user(tsk, &sc->fp_regs);
|
2013-02-14 00:21:41 +08:00
|
|
|
#ifdef CONFIG_VSX
|
|
|
|
/*
|
|
|
|
* Get additional VSX data. Update v_regs to point after the
|
|
|
|
* VMX data. Copy VSX low doubleword from userspace to local
|
|
|
|
* buffer for formatting, then into the taskstruct.
|
|
|
|
*/
|
|
|
|
if (v_regs && ((msr & MSR_VSX) != 0)) {
|
|
|
|
v_regs += ELF_NVRREG;
|
|
|
|
tm_v_regs += ELF_NVRREG;
|
2016-09-23 14:18:24 +08:00
|
|
|
err |= copy_vsx_from_user(tsk, tm_v_regs);
|
2016-09-23 14:18:25 +08:00
|
|
|
err |= copy_ckvsx_from_user(tsk, v_regs);
|
2016-09-23 14:18:12 +08:00
|
|
|
tsk->thread.used_vsr = true;
|
2013-02-14 00:21:41 +08:00
|
|
|
} else {
|
|
|
|
for (i = 0; i < 32 ; i++) {
|
2016-09-23 14:18:12 +08:00
|
|
|
tsk->thread.fp_state.fpr[i][TS_VSRLOWOFFSET] = 0;
|
2016-09-23 14:18:25 +08:00
|
|
|
tsk->thread.ckfp_state.fpr[i][TS_VSRLOWOFFSET] = 0;
|
2013-02-14 00:21:41 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
tm_enable();
|
2014-04-04 17:19:48 +08:00
|
|
|
/* Make sure the transaction is marked as failed */
|
2016-09-23 14:18:12 +08:00
|
|
|
tsk->thread.tm_texasr |= TEXASR_FS;
|
powerpc/tm: Set MSR[TS] just prior to recheckpoint
On a signal handler return, the user could set a context with MSR[TS] bits
set, and these bits would be copied to task regs->msr.
At restore_tm_sigcontexts(), after current task regs->msr[TS] bits are set,
several __get_user() are called and then a recheckpoint is executed.
This is a problem since a page fault (in kernel space) could happen when
calling __get_user(). If it happens, the process MSR[TS] bits were
already set, but recheckpoint was not executed, and SPRs are still invalid.
The page fault can cause the current process to be de-scheduled, with
MSR[TS] active and without tm_recheckpoint() being called. More
importantly, without TEXASR[FS] bit set also.
Since TEXASR might not have the FS bit set, and when the process is
scheduled back, it will try to reclaim, which will be aborted because of
the CPU is not in the suspended state, and, then, recheckpoint. This
recheckpoint will restore thread->texasr into TEXASR SPR, which might be
zero, hitting a BUG_ON().
kernel BUG at /build/linux-sf3Co9/linux-4.9.30/arch/powerpc/kernel/tm.S:434!
cpu 0xb: Vector: 700 (Program Check) at [c00000041f1576d0]
pc: c000000000054550: restore_gprs+0xb0/0x180
lr: 0000000000000000
sp: c00000041f157950
msr: 8000000100021033
current = 0xc00000041f143000
paca = 0xc00000000fb86300 softe: 0 irq_happened: 0x01
pid = 1021, comm = kworker/11:1
kernel BUG at /build/linux-sf3Co9/linux-4.9.30/arch/powerpc/kernel/tm.S:434!
Linux version 4.9.0-3-powerpc64le (debian-kernel@lists.debian.org) (gcc version 6.3.0 20170516 (Debian 6.3.0-18) ) #1 SMP Debian 4.9.30-2+deb9u2 (2017-06-26)
enter ? for help
[c00000041f157b30] c00000000001bc3c tm_recheckpoint.part.11+0x6c/0xa0
[c00000041f157b70] c00000000001d184 __switch_to+0x1e4/0x4c0
[c00000041f157bd0] c00000000082eeb8 __schedule+0x2f8/0x990
[c00000041f157cb0] c00000000082f598 schedule+0x48/0xc0
[c00000041f157ce0] c0000000000f0d28 worker_thread+0x148/0x610
[c00000041f157d80] c0000000000f96b0 kthread+0x120/0x140
[c00000041f157e30] c00000000000c0e0 ret_from_kernel_thread+0x5c/0x7c
This patch simply delays the MSR[TS] set, so, if there is any page fault in
the __get_user() section, it does not have regs->msr[TS] set, since the TM
structures are still invalid, thus avoiding doing TM operations for
in-kernel exceptions and possible process reschedule.
With this patch, the MSR[TS] will only be set just before recheckpointing
and setting TEXASR[FS] = 1, thus avoiding an interrupt with TM registers in
invalid state.
Other than that, if CONFIG_PREEMPT is set, there might be a preemption just
after setting MSR[TS] and before tm_recheckpoint(), thus, this block must
be atomic from a preemption perspective, thus, calling
preempt_disable/enable() on this code.
It is not possible to move tm_recheckpoint to happen earlier, because it is
required to get the checkpointed registers from userspace, with
__get_user(), thus, the only way to avoid this undesired behavior is
delaying the MSR[TS] set.
The 32-bits signal handler seems to be safe this current issue, but, it
might be exposed to the preemption issue, thus, disabling preemption in
this chunk of code.
Changes from v2:
* Run the critical section with preempt_disable.
Fixes: 87b4e5393af7 ("powerpc/tm: Fix return of active 64bit signals")
Cc: stable@vger.kernel.org (v3.9+)
Signed-off-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-11-22 03:21:09 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Disabling preemption, since it is unsafe to be preempted
|
|
|
|
* with MSR[TS] set without recheckpointing.
|
|
|
|
*/
|
|
|
|
preempt_disable();
|
|
|
|
|
|
|
|
/* pull in MSR TS bits from user context */
|
powerpc/tm: Avoid machine crash on rt_sigreturn()
There is a kernel crash that happens if rt_sigreturn() is called inside
a transactional block.
This crash happens if the kernel hits an in-kernel page fault when
accessing userspace memory, usually through copy_ckvsx_to_user(). A
major page fault calls might_sleep() function, which can cause a task
reschedule. A task reschedule (switch_to()) reclaim and recheckpoint
the TM states, but, in the signal return path, the checkpointed memory
was already reclaimed, thus the exception stack has MSR that points to
MSR[TS]=0.
When the code returns from might_sleep() and a task reschedule
happened, then this task is returned with the memory recheckpointed,
and CPU MSR[TS] = suspended.
This means that there is a side effect at might_sleep() if it is
called with CPU MSR[TS] = 0 and the task has regs->msr[TS] != 0.
This side effect can cause a TM bad thing, since at the exception
entrance, the stack saves MSR[TS]=0, and this is what will be used at
RFID, but, the processor has MSR[TS] = Suspended, and this transition
will be invalid and a TM Bad thing will be raised, causing the
following crash:
Unexpected TM Bad Thing exception at c00000000000e9ec (msr 0x8000000302a03031) tm_scratch=800000010280b033
cpu 0xc: Vector: 700 (Program Check) at [c00000003ff1fd70]
pc: c00000000000e9ec: fast_exception_return+0x100/0x1bc
lr: c000000000032948: handle_rt_signal64+0xb8/0xaf0
sp: c0000004263ebc40
msr: 8000000302a03031
current = 0xc000000415050300
paca = 0xc00000003ffc4080 irqmask: 0x03 irq_happened: 0x01
pid = 25006, comm = sigfuz
Linux version 5.0.0-rc1-00001-g3bd6e94bec12 (breno@debian) (gcc version 8.2.0 (Debian 8.2.0-3)) #899 SMP Mon Jan 7 11:30:07 EST 2019
WARNING: exception is not recoverable, can't continue
enter ? for help
[c0000004263ebc40] c000000000032948 handle_rt_signal64+0xb8/0xaf0 (unreliable)
[c0000004263ebd30] c000000000022780 do_notify_resume+0x2f0/0x430
[c0000004263ebe20] c00000000000e844 ret_from_except_lite+0x70/0x74
--- Exception: c00 (System Call) at 00007fffbaac400c
SP (7fffeca90f40) is in userspace
The solution for this problem is running the sigreturn code with
regs->msr[TS] disabled, thus, avoiding hitting the side effect above.
This does not seem to be a problem since regs->msr will be replaced by
the ucontext value, so, it is being flushed already. In this case, it
is flushed earlier.
Signed-off-by: Breno Leitao <leitao@debian.org>
Acked-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2019-01-17 00:47:44 +08:00
|
|
|
regs->msr |= msr & MSR_TS_MASK;
|
powerpc/tm: Set MSR[TS] just prior to recheckpoint
On a signal handler return, the user could set a context with MSR[TS] bits
set, and these bits would be copied to task regs->msr.
At restore_tm_sigcontexts(), after current task regs->msr[TS] bits are set,
several __get_user() are called and then a recheckpoint is executed.
This is a problem since a page fault (in kernel space) could happen when
calling __get_user(). If it happens, the process MSR[TS] bits were
already set, but recheckpoint was not executed, and SPRs are still invalid.
The page fault can cause the current process to be de-scheduled, with
MSR[TS] active and without tm_recheckpoint() being called. More
importantly, without TEXASR[FS] bit set also.
Since TEXASR might not have the FS bit set, and when the process is
scheduled back, it will try to reclaim, which will be aborted because of
the CPU is not in the suspended state, and, then, recheckpoint. This
recheckpoint will restore thread->texasr into TEXASR SPR, which might be
zero, hitting a BUG_ON().
kernel BUG at /build/linux-sf3Co9/linux-4.9.30/arch/powerpc/kernel/tm.S:434!
cpu 0xb: Vector: 700 (Program Check) at [c00000041f1576d0]
pc: c000000000054550: restore_gprs+0xb0/0x180
lr: 0000000000000000
sp: c00000041f157950
msr: 8000000100021033
current = 0xc00000041f143000
paca = 0xc00000000fb86300 softe: 0 irq_happened: 0x01
pid = 1021, comm = kworker/11:1
kernel BUG at /build/linux-sf3Co9/linux-4.9.30/arch/powerpc/kernel/tm.S:434!
Linux version 4.9.0-3-powerpc64le (debian-kernel@lists.debian.org) (gcc version 6.3.0 20170516 (Debian 6.3.0-18) ) #1 SMP Debian 4.9.30-2+deb9u2 (2017-06-26)
enter ? for help
[c00000041f157b30] c00000000001bc3c tm_recheckpoint.part.11+0x6c/0xa0
[c00000041f157b70] c00000000001d184 __switch_to+0x1e4/0x4c0
[c00000041f157bd0] c00000000082eeb8 __schedule+0x2f8/0x990
[c00000041f157cb0] c00000000082f598 schedule+0x48/0xc0
[c00000041f157ce0] c0000000000f0d28 worker_thread+0x148/0x610
[c00000041f157d80] c0000000000f96b0 kthread+0x120/0x140
[c00000041f157e30] c00000000000c0e0 ret_from_kernel_thread+0x5c/0x7c
This patch simply delays the MSR[TS] set, so, if there is any page fault in
the __get_user() section, it does not have regs->msr[TS] set, since the TM
structures are still invalid, thus avoiding doing TM operations for
in-kernel exceptions and possible process reschedule.
With this patch, the MSR[TS] will only be set just before recheckpointing
and setting TEXASR[FS] = 1, thus avoiding an interrupt with TM registers in
invalid state.
Other than that, if CONFIG_PREEMPT is set, there might be a preemption just
after setting MSR[TS] and before tm_recheckpoint(), thus, this block must
be atomic from a preemption perspective, thus, calling
preempt_disable/enable() on this code.
It is not possible to move tm_recheckpoint to happen earlier, because it is
required to get the checkpointed registers from userspace, with
__get_user(), thus, the only way to avoid this undesired behavior is
delaying the MSR[TS] set.
The 32-bits signal handler seems to be safe this current issue, but, it
might be exposed to the preemption issue, thus, disabling preemption in
this chunk of code.
Changes from v2:
* Run the critical section with preempt_disable.
Fixes: 87b4e5393af7 ("powerpc/tm: Fix return of active 64bit signals")
Cc: stable@vger.kernel.org (v3.9+)
Signed-off-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-11-22 03:21:09 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Ensure that TM is enabled in regs->msr before we leave the signal
|
|
|
|
* handler. It could be the case that (a) user disabled the TM bit
|
|
|
|
* through the manipulation of the MSR bits in uc_mcontext or (b) the
|
|
|
|
* TM bit was disabled because a sufficient number of context switches
|
|
|
|
* happened whilst in the signal handler and load_tm overflowed,
|
|
|
|
* disabling the TM bit. In either case we can end up with an illegal
|
|
|
|
* TM state leading to a TM Bad Thing when we return to userspace.
|
|
|
|
*
|
|
|
|
* CAUTION:
|
|
|
|
* After regs->MSR[TS] being updated, make sure that get_user(),
|
|
|
|
* put_user() or similar functions are *not* called. These
|
|
|
|
* functions can generate page faults which will cause the process
|
|
|
|
* to be de-scheduled with MSR[TS] set but without calling
|
|
|
|
* tm_recheckpoint(). This can cause a bug.
|
|
|
|
*/
|
|
|
|
regs->msr |= MSR_TM;
|
|
|
|
|
2013-02-14 00:21:41 +08:00
|
|
|
/* This loads the checkpointed FP/VEC state, if used */
|
2017-11-02 11:09:05 +08:00
|
|
|
tm_recheckpoint(&tsk->thread);
|
2013-02-14 00:21:41 +08:00
|
|
|
|
2016-09-23 14:18:24 +08:00
|
|
|
msr_check_and_set(msr & (MSR_FP | MSR_VEC));
|
2013-02-14 00:21:41 +08:00
|
|
|
if (msr & MSR_FP) {
|
2016-09-23 14:18:24 +08:00
|
|
|
load_fp_state(&tsk->thread.fp_state);
|
2016-09-23 14:18:12 +08:00
|
|
|
regs->msr |= (MSR_FP | tsk->thread.fpexc_mode);
|
2013-02-14 00:21:41 +08:00
|
|
|
}
|
|
|
|
if (msr & MSR_VEC) {
|
2016-09-23 14:18:24 +08:00
|
|
|
load_vr_state(&tsk->thread.vr_state);
|
2013-02-14 00:21:41 +08:00
|
|
|
regs->msr |= MSR_VEC;
|
|
|
|
}
|
|
|
|
|
powerpc/tm: Set MSR[TS] just prior to recheckpoint
On a signal handler return, the user could set a context with MSR[TS] bits
set, and these bits would be copied to task regs->msr.
At restore_tm_sigcontexts(), after current task regs->msr[TS] bits are set,
several __get_user() are called and then a recheckpoint is executed.
This is a problem since a page fault (in kernel space) could happen when
calling __get_user(). If it happens, the process MSR[TS] bits were
already set, but recheckpoint was not executed, and SPRs are still invalid.
The page fault can cause the current process to be de-scheduled, with
MSR[TS] active and without tm_recheckpoint() being called. More
importantly, without TEXASR[FS] bit set also.
Since TEXASR might not have the FS bit set, and when the process is
scheduled back, it will try to reclaim, which will be aborted because of
the CPU is not in the suspended state, and, then, recheckpoint. This
recheckpoint will restore thread->texasr into TEXASR SPR, which might be
zero, hitting a BUG_ON().
kernel BUG at /build/linux-sf3Co9/linux-4.9.30/arch/powerpc/kernel/tm.S:434!
cpu 0xb: Vector: 700 (Program Check) at [c00000041f1576d0]
pc: c000000000054550: restore_gprs+0xb0/0x180
lr: 0000000000000000
sp: c00000041f157950
msr: 8000000100021033
current = 0xc00000041f143000
paca = 0xc00000000fb86300 softe: 0 irq_happened: 0x01
pid = 1021, comm = kworker/11:1
kernel BUG at /build/linux-sf3Co9/linux-4.9.30/arch/powerpc/kernel/tm.S:434!
Linux version 4.9.0-3-powerpc64le (debian-kernel@lists.debian.org) (gcc version 6.3.0 20170516 (Debian 6.3.0-18) ) #1 SMP Debian 4.9.30-2+deb9u2 (2017-06-26)
enter ? for help
[c00000041f157b30] c00000000001bc3c tm_recheckpoint.part.11+0x6c/0xa0
[c00000041f157b70] c00000000001d184 __switch_to+0x1e4/0x4c0
[c00000041f157bd0] c00000000082eeb8 __schedule+0x2f8/0x990
[c00000041f157cb0] c00000000082f598 schedule+0x48/0xc0
[c00000041f157ce0] c0000000000f0d28 worker_thread+0x148/0x610
[c00000041f157d80] c0000000000f96b0 kthread+0x120/0x140
[c00000041f157e30] c00000000000c0e0 ret_from_kernel_thread+0x5c/0x7c
This patch simply delays the MSR[TS] set, so, if there is any page fault in
the __get_user() section, it does not have regs->msr[TS] set, since the TM
structures are still invalid, thus avoiding doing TM operations for
in-kernel exceptions and possible process reschedule.
With this patch, the MSR[TS] will only be set just before recheckpointing
and setting TEXASR[FS] = 1, thus avoiding an interrupt with TM registers in
invalid state.
Other than that, if CONFIG_PREEMPT is set, there might be a preemption just
after setting MSR[TS] and before tm_recheckpoint(), thus, this block must
be atomic from a preemption perspective, thus, calling
preempt_disable/enable() on this code.
It is not possible to move tm_recheckpoint to happen earlier, because it is
required to get the checkpointed registers from userspace, with
__get_user(), thus, the only way to avoid this undesired behavior is
delaying the MSR[TS] set.
The 32-bits signal handler seems to be safe this current issue, but, it
might be exposed to the preemption issue, thus, disabling preemption in
this chunk of code.
Changes from v2:
* Run the critical section with preempt_disable.
Fixes: 87b4e5393af7 ("powerpc/tm: Fix return of active 64bit signals")
Cc: stable@vger.kernel.org (v3.9+)
Signed-off-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-11-22 03:21:09 +08:00
|
|
|
preempt_enable();
|
|
|
|
|
2013-02-14 00:21:41 +08:00
|
|
|
return err;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* Setup the trampoline code on the stack
|
|
|
|
*/
|
|
|
|
static long setup_trampoline(unsigned int syscall, unsigned int __user *tramp)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
long err = 0;
|
|
|
|
|
|
|
|
/* addi r1, r1, __SIGNAL_FRAMESIZE # Pop the dummy stackframe */
|
2018-11-10 01:33:28 +08:00
|
|
|
err |= __put_user(PPC_INST_ADDI | __PPC_RT(R1) | __PPC_RA(R1) |
|
|
|
|
(__SIGNAL_FRAMESIZE & 0xffff), &tramp[0]);
|
2005-04-17 06:20:36 +08:00
|
|
|
/* li r0, __NR_[rt_]sigreturn| */
|
2018-11-10 01:33:28 +08:00
|
|
|
err |= __put_user(PPC_INST_ADDI | (syscall & 0xffff), &tramp[1]);
|
2005-04-17 06:20:36 +08:00
|
|
|
/* sc */
|
2018-11-10 01:33:28 +08:00
|
|
|
err |= __put_user(PPC_INST_SC, &tramp[2]);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/* Minimal traceback info */
|
|
|
|
for (i=TRAMP_TRACEBACK; i < TRAMP_SIZE ;i++)
|
|
|
|
err |= __put_user(0, &tramp[i]);
|
|
|
|
|
|
|
|
if (!err)
|
|
|
|
flush_icache_range((unsigned long) &tramp[0],
|
|
|
|
(unsigned long) &tramp[TRAMP_SIZE]);
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2008-07-08 16:43:41 +08:00
|
|
|
/*
|
|
|
|
* Userspace code may pass a ucontext which doesn't include VSX added
|
|
|
|
* at the end. We need to check for this case.
|
|
|
|
*/
|
|
|
|
#define UCONTEXTSIZEWITHOUTVSX \
|
|
|
|
(sizeof(struct ucontext) - 32*sizeof(long))
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* Handle {get,set,swap}_context operations
|
|
|
|
*/
|
2018-05-02 21:20:47 +08:00
|
|
|
SYSCALL_DEFINE3(swapcontext, struct ucontext __user *, old_ctx,
|
|
|
|
struct ucontext __user *, new_ctx, long, ctx_size)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
unsigned char tmp;
|
|
|
|
sigset_t set;
|
2008-07-08 16:43:41 +08:00
|
|
|
unsigned long new_msr = 0;
|
2008-10-23 08:42:36 +08:00
|
|
|
int ctx_has_vsx_region = 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-07-08 16:43:41 +08:00
|
|
|
if (new_ctx &&
|
2008-10-23 08:42:36 +08:00
|
|
|
get_user(new_msr, &new_ctx->uc_mcontext.gp_regs[PT_MSR]))
|
2008-07-08 16:43:41 +08:00
|
|
|
return -EFAULT;
|
|
|
|
/*
|
|
|
|
* Check that the context is not smaller than the original
|
|
|
|
* size (with VMX but without VSX)
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
2008-07-08 16:43:41 +08:00
|
|
|
if (ctx_size < UCONTEXTSIZEWITHOUTVSX)
|
2005-04-17 06:20:36 +08:00
|
|
|
return -EINVAL;
|
2008-07-08 16:43:41 +08:00
|
|
|
/*
|
|
|
|
* If the new context state sets the MSR VSX bits but
|
|
|
|
* it doesn't provide VSX state.
|
|
|
|
*/
|
|
|
|
if ((ctx_size < sizeof(struct ucontext)) &&
|
|
|
|
(new_msr & MSR_VSX))
|
|
|
|
return -EINVAL;
|
2008-10-23 08:42:36 +08:00
|
|
|
/* Does the context have enough room to store VSX data? */
|
|
|
|
if (ctx_size >= sizeof(struct ucontext))
|
|
|
|
ctx_has_vsx_region = 1;
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
if (old_ctx != NULL) {
|
Remove 'type' argument from access_ok() function
Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument
of the user address range verification function since we got rid of the
old racy i386-only code to walk page tables by hand.
It existed because the original 80386 would not honor the write protect
bit when in kernel mode, so you had to do COW by hand before doing any
user access. But we haven't supported that in a long time, and these
days the 'type' argument is a purely historical artifact.
A discussion about extending 'user_access_begin()' to do the range
checking resulted this patch, because there is no way we're going to
move the old VERIFY_xyz interface to that model. And it's best done at
the end of the merge window when I've done most of my merges, so let's
just get this done once and for all.
This patch was mostly done with a sed-script, with manual fix-ups for
the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form.
There were a couple of notable cases:
- csky still had the old "verify_area()" name as an alias.
- the iter_iov code had magical hardcoded knowledge of the actual
values of VERIFY_{READ,WRITE} (not that they mattered, since nothing
really used it)
- microblaze used the type argument for a debug printout
but other than those oddities this should be a total no-op patch.
I tried to fix up all architectures, did fairly extensive grepping for
access_ok() uses, and the changes are trivial, but I may have missed
something. Any missed conversion should be trivially fixable, though.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-01-04 10:57:57 +08:00
|
|
|
if (!access_ok(old_ctx, ctx_size)
|
2016-09-23 14:18:12 +08:00
|
|
|
|| setup_sigcontext(&old_ctx->uc_mcontext, current, 0, NULL, 0,
|
2008-10-23 08:42:36 +08:00
|
|
|
ctx_has_vsx_region)
|
2005-04-17 06:20:36 +08:00
|
|
|
|| __copy_to_user(&old_ctx->uc_sigmask,
|
|
|
|
¤t->blocked, sizeof(sigset_t)))
|
|
|
|
return -EFAULT;
|
|
|
|
}
|
|
|
|
if (new_ctx == NULL)
|
|
|
|
return 0;
|
Remove 'type' argument from access_ok() function
Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument
of the user address range verification function since we got rid of the
old racy i386-only code to walk page tables by hand.
It existed because the original 80386 would not honor the write protect
bit when in kernel mode, so you had to do COW by hand before doing any
user access. But we haven't supported that in a long time, and these
days the 'type' argument is a purely historical artifact.
A discussion about extending 'user_access_begin()' to do the range
checking resulted this patch, because there is no way we're going to
move the old VERIFY_xyz interface to that model. And it's best done at
the end of the merge window when I've done most of my merges, so let's
just get this done once and for all.
This patch was mostly done with a sed-script, with manual fix-ups for
the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form.
There were a couple of notable cases:
- csky still had the old "verify_area()" name as an alias.
- the iter_iov code had magical hardcoded knowledge of the actual
values of VERIFY_{READ,WRITE} (not that they mattered, since nothing
really used it)
- microblaze used the type argument for a debug printout
but other than those oddities this should be a total no-op patch.
I tried to fix up all architectures, did fairly extensive grepping for
access_ok() uses, and the changes are trivial, but I may have missed
something. Any missed conversion should be trivially fixable, though.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-01-04 10:57:57 +08:00
|
|
|
if (!access_ok(new_ctx, ctx_size)
|
2005-04-17 06:20:36 +08:00
|
|
|
|| __get_user(tmp, (u8 __user *) new_ctx)
|
2008-10-23 08:42:36 +08:00
|
|
|
|| __get_user(tmp, (u8 __user *) new_ctx + ctx_size - 1))
|
2005-04-17 06:20:36 +08:00
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If we get a fault copying the context into the kernel's
|
|
|
|
* image of the user's registers, we can't just return -EFAULT
|
|
|
|
* because the user's registers will be corrupted. For instance
|
|
|
|
* the NIP value may have been updated but not some of the
|
|
|
|
* other registers. Given that we have done the access_ok
|
|
|
|
* and successfully read the first and last bytes of the region
|
|
|
|
* above, this should only happen in an out-of-memory situation
|
|
|
|
* or if another thread unmaps the region containing the context.
|
|
|
|
* We kill the task with a SIGSEGV in this situation.
|
|
|
|
*/
|
|
|
|
|
|
|
|
if (__copy_from_user(&set, &new_ctx->uc_sigmask, sizeof(set)))
|
|
|
|
do_exit(SIGSEGV);
|
2012-04-28 02:09:19 +08:00
|
|
|
set_current_blocked(&set);
|
2016-09-23 14:18:12 +08:00
|
|
|
if (restore_sigcontext(current, NULL, 0, &new_ctx->uc_mcontext))
|
2005-04-17 06:20:36 +08:00
|
|
|
do_exit(SIGSEGV);
|
|
|
|
|
|
|
|
/* This returns like rt_sigreturn */
|
[PATCH] syscall entry/exit revamp
This cleanup patch speeds up the null syscall path on ppc64 by about 3%,
and brings the ppc32 and ppc64 code slightly closer together.
The ppc64 code was checking current_thread_info()->flags twice in the
syscall exit path; once for TIF_SYSCALL_T_OR_A before disabling
interrupts, and then again for TIF_SIGPENDING|TIF_NEED_RESCHED etc after
disabling interrupts. Now we do the same as ppc32 -- check the flags
only once in the fast path, and re-enable interrupts if necessary in the
ptrace case.
The patch abolishes the 'syscall_noerror' member of struct thread_info
and replaces it with a TIF_NOERROR bit in the flags, which is handled in
the slow path. This shortens the syscall entry code, which no longer
needs to clear syscall_noerror.
The patch adds a TIF_SAVE_NVGPRS flag which causes the syscall exit slow
path to save the non-volatile GPRs into a signal frame. This removes the
need for the assembly wrappers around sys_sigsuspend(),
sys_rt_sigsuspend(), et al which existed solely to save those registers
in advance. It also means I don't have to add new wrappers for ppoll()
and pselect(), which is what I was supposed to be doing when I got
distracted into this...
Finally, it unifies the ppc64 and ppc32 methods of handling syscall exit
directly into a signal handler (as required by sigsuspend et al) by
introducing a TIF_RESTOREALL flag which causes _all_ the registers to be
reloaded from the pt_regs by taking the ret_from_exception path, instead
of the normal syscall exit path which stomps on the callee-saved GPRs.
It appears to pass an LTP test run on ppc64, and passes basic testing on
ppc32 too. Brief tests of ptrace functionality with strace and gdb also
appear OK. I wouldn't send it to Linus for 2.6.15 just yet though :)
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-11-16 02:52:18 +08:00
|
|
|
set_thread_flag(TIF_RESTOREALL);
|
2005-04-17 06:20:36 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Do a signal return; undo the signal stack.
|
|
|
|
*/
|
|
|
|
|
2018-05-02 21:20:47 +08:00
|
|
|
SYSCALL_DEFINE0(rt_sigreturn)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2018-05-02 21:20:47 +08:00
|
|
|
struct pt_regs *regs = current_pt_regs();
|
2005-04-17 06:20:36 +08:00
|
|
|
struct ucontext __user *uc = (struct ucontext __user *)regs->gpr[1];
|
|
|
|
sigset_t set;
|
2013-02-14 00:21:41 +08:00
|
|
|
#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
|
|
|
|
unsigned long msr;
|
|
|
|
#endif
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/* Always make any pending restarted system calls return -EINTR */
|
2015-02-13 07:01:14 +08:00
|
|
|
current->restart_block.fn = do_no_restart_syscall;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
Remove 'type' argument from access_ok() function
Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument
of the user address range verification function since we got rid of the
old racy i386-only code to walk page tables by hand.
It existed because the original 80386 would not honor the write protect
bit when in kernel mode, so you had to do COW by hand before doing any
user access. But we haven't supported that in a long time, and these
days the 'type' argument is a purely historical artifact.
A discussion about extending 'user_access_begin()' to do the range
checking resulted this patch, because there is no way we're going to
move the old VERIFY_xyz interface to that model. And it's best done at
the end of the merge window when I've done most of my merges, so let's
just get this done once and for all.
This patch was mostly done with a sed-script, with manual fix-ups for
the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form.
There were a couple of notable cases:
- csky still had the old "verify_area()" name as an alias.
- the iter_iov code had magical hardcoded knowledge of the actual
values of VERIFY_{READ,WRITE} (not that they mattered, since nothing
really used it)
- microblaze used the type argument for a debug printout
but other than those oddities this should be a total no-op patch.
I tried to fix up all architectures, did fairly extensive grepping for
access_ok() uses, and the changes are trivial, but I may have missed
something. Any missed conversion should be trivially fixable, though.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-01-04 10:57:57 +08:00
|
|
|
if (!access_ok(uc, sizeof(*uc)))
|
2005-04-17 06:20:36 +08:00
|
|
|
goto badframe;
|
|
|
|
|
|
|
|
if (__copy_from_user(&set, &uc->uc_sigmask, sizeof(set)))
|
|
|
|
goto badframe;
|
2012-04-28 02:09:19 +08:00
|
|
|
set_current_blocked(&set);
|
powerpc: signals: Discard transaction state from signal frames
Userspace can begin and suspend a transaction within the signal
handler which means they might enter sys_rt_sigreturn() with the
processor in suspended state.
sys_rt_sigreturn() wants to restore process context (which may have
been in a transaction before signal delivery). To do this it must
restore TM SPRS. To achieve this, any transaction initiated within the
signal frame must be discarded in order to be able to restore TM SPRs
as TM SPRs can only be manipulated non-transactionally..
>From the PowerPC ISA:
TM Bad Thing Exception [Category: Transactional Memory]
An attempt is made to execute a mtspr targeting a TM register in
other than Non-transactional state.
Not doing so results in a TM Bad Thing:
[12045.221359] Kernel BUG at c000000000050a40 [verbose debug info unavailable]
[12045.221470] Unexpected TM Bad Thing exception at c000000000050a40 (msr 0x201033)
[12045.221540] Oops: Unrecoverable exception, sig: 6 [#1]
[12045.221586] SMP NR_CPUS=2048 NUMA PowerNV
[12045.221634] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE
nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4
xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 xt_tcpudp bridge stp llc ebtable_filter
ebtables ip6table_filter ip6_tables iptable_filter ip_tables x_tables kvm_hv kvm
uio_pdrv_genirq ipmi_powernv uio powernv_rng ipmi_msghandler autofs4 ses enclosure
scsi_transport_sas bnx2x ipr mdio libcrc32c
[12045.222167] CPU: 68 PID: 6178 Comm: sigreturnpanic Not tainted 4.7.0 #34
[12045.222224] task: c0000000fce38600 ti: c0000000fceb4000 task.ti: c0000000fceb4000
[12045.222293] NIP: c000000000050a40 LR: c0000000000163bc CTR: 0000000000000000
[12045.222361] REGS: c0000000fceb7ac0 TRAP: 0700 Not tainted (4.7.0)
[12045.222418] MSR: 9000000300201033 <SF,HV,ME,IR,DR,RI,LE,TM[SE]> CR: 28444280 XER: 20000000
[12045.222625] CFAR: c0000000000163b8 SOFTE: 0 PACATMSCRATCH: 900000014280f033
GPR00: 01100000b8000001 c0000000fceb7d40 c00000000139c100 c0000000fce390d0
GPR04: 900000034280f033 0000000000000000 0000000000000000 0000000000000000
GPR08: 0000000000000000 b000000000001033 0000000000000001 0000000000000000
GPR12: 0000000000000000 c000000002926400 0000000000000000 0000000000000000
GPR16: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
GPR20: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
GPR24: 0000000000000000 00003ffff98cadd0 00003ffff98cb470 0000000000000000
GPR28: 900000034280f033 c0000000fceb7ea0 0000000000000001 c0000000fce390d0
[12045.223535] NIP [c000000000050a40] tm_restore_sprs+0xc/0x1c
[12045.223584] LR [c0000000000163bc] tm_recheckpoint+0x5c/0xa0
[12045.223630] Call Trace:
[12045.223655] [c0000000fceb7d80] [c000000000026e74] sys_rt_sigreturn+0x494/0x6c0
[12045.223738] [c0000000fceb7e30] [c0000000000092e0] system_call+0x38/0x108
[12045.223806] Instruction dump:
[12045.223841] 7c800164 4e800020 7c0022a6 f80304a8 7c0222a6 f80304b0 7c0122a6 f80304b8
[12045.223955] 4e800020 e80304a8 7c0023a6 e80304b0 <7c0223a6> e80304b8 7c0123a6 4e800020
[12045.224074] ---[ end trace cb8002ee240bae76 ]---
It isn't clear exactly if there is really a use case for userspace
returning with a suspended transaction, however, doing so doesn't (on
its own) constitute a bad frame. As such, this patch simply discards
the transactional state of the context calling the sigreturn and
continues.
Reported-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Signed-off-by: Cyril Bur <cyrilbur@gmail.com>
Tested-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Reviewed-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Acked-by: Simon Guo <wei.guo.simon@gmail.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2016-08-23 08:46:17 +08:00
|
|
|
|
2013-02-14 00:21:41 +08:00
|
|
|
#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
|
powerpc: signals: Discard transaction state from signal frames
Userspace can begin and suspend a transaction within the signal
handler which means they might enter sys_rt_sigreturn() with the
processor in suspended state.
sys_rt_sigreturn() wants to restore process context (which may have
been in a transaction before signal delivery). To do this it must
restore TM SPRS. To achieve this, any transaction initiated within the
signal frame must be discarded in order to be able to restore TM SPRs
as TM SPRs can only be manipulated non-transactionally..
>From the PowerPC ISA:
TM Bad Thing Exception [Category: Transactional Memory]
An attempt is made to execute a mtspr targeting a TM register in
other than Non-transactional state.
Not doing so results in a TM Bad Thing:
[12045.221359] Kernel BUG at c000000000050a40 [verbose debug info unavailable]
[12045.221470] Unexpected TM Bad Thing exception at c000000000050a40 (msr 0x201033)
[12045.221540] Oops: Unrecoverable exception, sig: 6 [#1]
[12045.221586] SMP NR_CPUS=2048 NUMA PowerNV
[12045.221634] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE
nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4
xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 xt_tcpudp bridge stp llc ebtable_filter
ebtables ip6table_filter ip6_tables iptable_filter ip_tables x_tables kvm_hv kvm
uio_pdrv_genirq ipmi_powernv uio powernv_rng ipmi_msghandler autofs4 ses enclosure
scsi_transport_sas bnx2x ipr mdio libcrc32c
[12045.222167] CPU: 68 PID: 6178 Comm: sigreturnpanic Not tainted 4.7.0 #34
[12045.222224] task: c0000000fce38600 ti: c0000000fceb4000 task.ti: c0000000fceb4000
[12045.222293] NIP: c000000000050a40 LR: c0000000000163bc CTR: 0000000000000000
[12045.222361] REGS: c0000000fceb7ac0 TRAP: 0700 Not tainted (4.7.0)
[12045.222418] MSR: 9000000300201033 <SF,HV,ME,IR,DR,RI,LE,TM[SE]> CR: 28444280 XER: 20000000
[12045.222625] CFAR: c0000000000163b8 SOFTE: 0 PACATMSCRATCH: 900000014280f033
GPR00: 01100000b8000001 c0000000fceb7d40 c00000000139c100 c0000000fce390d0
GPR04: 900000034280f033 0000000000000000 0000000000000000 0000000000000000
GPR08: 0000000000000000 b000000000001033 0000000000000001 0000000000000000
GPR12: 0000000000000000 c000000002926400 0000000000000000 0000000000000000
GPR16: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
GPR20: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
GPR24: 0000000000000000 00003ffff98cadd0 00003ffff98cb470 0000000000000000
GPR28: 900000034280f033 c0000000fceb7ea0 0000000000000001 c0000000fce390d0
[12045.223535] NIP [c000000000050a40] tm_restore_sprs+0xc/0x1c
[12045.223584] LR [c0000000000163bc] tm_recheckpoint+0x5c/0xa0
[12045.223630] Call Trace:
[12045.223655] [c0000000fceb7d80] [c000000000026e74] sys_rt_sigreturn+0x494/0x6c0
[12045.223738] [c0000000fceb7e30] [c0000000000092e0] system_call+0x38/0x108
[12045.223806] Instruction dump:
[12045.223841] 7c800164 4e800020 7c0022a6 f80304a8 7c0222a6 f80304b0 7c0122a6 f80304b8
[12045.223955] 4e800020 e80304a8 7c0023a6 e80304b0 <7c0223a6> e80304b8 7c0123a6 4e800020
[12045.224074] ---[ end trace cb8002ee240bae76 ]---
It isn't clear exactly if there is really a use case for userspace
returning with a suspended transaction, however, doing so doesn't (on
its own) constitute a bad frame. As such, this patch simply discards
the transactional state of the context calling the sigreturn and
continues.
Reported-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Signed-off-by: Cyril Bur <cyrilbur@gmail.com>
Tested-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Reviewed-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Acked-by: Simon Guo <wei.guo.simon@gmail.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2016-08-23 08:46:17 +08:00
|
|
|
/*
|
|
|
|
* If there is a transactional state then throw it away.
|
|
|
|
* The purpose of a sigreturn is to destroy all traces of the
|
|
|
|
* signal frame, this includes any transactional state created
|
|
|
|
* within in. We only check for suspended as we can never be
|
|
|
|
* active in the kernel, we are active, there is nothing better to
|
|
|
|
* do than go ahead and Bad Thing later.
|
|
|
|
* The cause is not important as there will never be a
|
|
|
|
* recheckpoint so it's not user visible.
|
|
|
|
*/
|
|
|
|
if (MSR_TM_SUSPENDED(mfmsr()))
|
|
|
|
tm_reclaim_current(0);
|
|
|
|
|
powerpc/tm: Avoid machine crash on rt_sigreturn()
There is a kernel crash that happens if rt_sigreturn() is called inside
a transactional block.
This crash happens if the kernel hits an in-kernel page fault when
accessing userspace memory, usually through copy_ckvsx_to_user(). A
major page fault calls might_sleep() function, which can cause a task
reschedule. A task reschedule (switch_to()) reclaim and recheckpoint
the TM states, but, in the signal return path, the checkpointed memory
was already reclaimed, thus the exception stack has MSR that points to
MSR[TS]=0.
When the code returns from might_sleep() and a task reschedule
happened, then this task is returned with the memory recheckpointed,
and CPU MSR[TS] = suspended.
This means that there is a side effect at might_sleep() if it is
called with CPU MSR[TS] = 0 and the task has regs->msr[TS] != 0.
This side effect can cause a TM bad thing, since at the exception
entrance, the stack saves MSR[TS]=0, and this is what will be used at
RFID, but, the processor has MSR[TS] = Suspended, and this transition
will be invalid and a TM Bad thing will be raised, causing the
following crash:
Unexpected TM Bad Thing exception at c00000000000e9ec (msr 0x8000000302a03031) tm_scratch=800000010280b033
cpu 0xc: Vector: 700 (Program Check) at [c00000003ff1fd70]
pc: c00000000000e9ec: fast_exception_return+0x100/0x1bc
lr: c000000000032948: handle_rt_signal64+0xb8/0xaf0
sp: c0000004263ebc40
msr: 8000000302a03031
current = 0xc000000415050300
paca = 0xc00000003ffc4080 irqmask: 0x03 irq_happened: 0x01
pid = 25006, comm = sigfuz
Linux version 5.0.0-rc1-00001-g3bd6e94bec12 (breno@debian) (gcc version 8.2.0 (Debian 8.2.0-3)) #899 SMP Mon Jan 7 11:30:07 EST 2019
WARNING: exception is not recoverable, can't continue
enter ? for help
[c0000004263ebc40] c000000000032948 handle_rt_signal64+0xb8/0xaf0 (unreliable)
[c0000004263ebd30] c000000000022780 do_notify_resume+0x2f0/0x430
[c0000004263ebe20] c00000000000e844 ret_from_except_lite+0x70/0x74
--- Exception: c00 (System Call) at 00007fffbaac400c
SP (7fffeca90f40) is in userspace
The solution for this problem is running the sigreturn code with
regs->msr[TS] disabled, thus, avoiding hitting the side effect above.
This does not seem to be a problem since regs->msr will be replaced by
the ucontext value, so, it is being flushed already. In this case, it
is flushed earlier.
Signed-off-by: Breno Leitao <leitao@debian.org>
Acked-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2019-01-17 00:47:44 +08:00
|
|
|
/*
|
|
|
|
* Disable MSR[TS] bit also, so, if there is an exception in the
|
|
|
|
* code below (as a page fault in copy_ckvsx_to_user()), it does
|
|
|
|
* not recheckpoint this task if there was a context switch inside
|
|
|
|
* the exception.
|
|
|
|
*
|
|
|
|
* A major page fault can indirectly call schedule(). A reschedule
|
|
|
|
* process in the middle of an exception can have a side effect
|
|
|
|
* (Changing the CPU MSR[TS] state), since schedule() is called
|
|
|
|
* with the CPU MSR[TS] disable and returns with MSR[TS]=Suspended
|
|
|
|
* (switch_to() calls tm_recheckpoint() for the 'new' process). In
|
|
|
|
* this case, the process continues to be the same in the CPU, but
|
|
|
|
* the CPU state just changed.
|
|
|
|
*
|
|
|
|
* This can cause a TM Bad Thing, since the MSR in the stack will
|
|
|
|
* have the MSR[TS]=0, and this is what will be used to RFID.
|
|
|
|
*
|
|
|
|
* Clearing MSR[TS] state here will avoid a recheckpoint if there
|
|
|
|
* is any process reschedule in kernel space. The MSR[TS] state
|
|
|
|
* does not need to be saved also, since it will be replaced with
|
|
|
|
* the MSR[TS] that came from user context later, at
|
|
|
|
* restore_tm_sigcontexts.
|
|
|
|
*/
|
|
|
|
regs->msr &= ~MSR_TS_MASK;
|
|
|
|
|
2013-02-14 00:21:41 +08:00
|
|
|
if (__get_user(msr, &uc->uc_mcontext.gp_regs[PT_MSR]))
|
|
|
|
goto badframe;
|
2013-06-09 19:23:19 +08:00
|
|
|
if (MSR_TM_ACTIVE(msr)) {
|
2013-02-14 00:21:41 +08:00
|
|
|
/* We recheckpoint on return. */
|
|
|
|
struct ucontext __user *uc_transact;
|
powerpc/tm: Fix oops on sigreturn on systems without TM
On systems like P9 powernv where we have no TM (or P8 booted with
ppc_tm=off), userspace can construct a signal context which still has
the MSR TS bits set. The kernel tries to restore this context which
results in the following crash:
Unexpected TM Bad Thing exception at c0000000000022fc (msr 0x8000000102a03031) tm_scratch=800000020280f033
Oops: Unrecoverable exception, sig: 6 [#1]
LE PAGE_SIZE=64K MMU=Hash SMP NR_CPUS=2048 NUMA pSeries
Modules linked in:
CPU: 0 PID: 1636 Comm: sigfuz Not tainted 5.2.0-11043-g0a8ad0ffa4 #69
NIP: c0000000000022fc LR: 00007fffb2d67e48 CTR: 0000000000000000
REGS: c00000003fffbd70 TRAP: 0700 Not tainted (5.2.0-11045-g7142b497d8)
MSR: 8000000102a03031 <SF,VEC,VSX,FP,ME,IR,DR,LE,TM[E]> CR: 42004242 XER: 00000000
CFAR: c0000000000022e0 IRQMASK: 0
GPR00: 0000000000000072 00007fffb2b6e560 00007fffb2d87f00 0000000000000669
GPR04: 00007fffb2b6e728 0000000000000000 0000000000000000 00007fffb2b6f2a8
GPR08: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
GPR12: 0000000000000000 00007fffb2b76900 0000000000000000 0000000000000000
GPR16: 00007fffb2370000 00007fffb2d84390 00007fffea3a15ac 000001000a250420
GPR20: 00007fffb2b6f260 0000000010001770 0000000000000000 0000000000000000
GPR24: 00007fffb2d843a0 00007fffea3a14a0 0000000000010000 0000000000800000
GPR28: 00007fffea3a14d8 00000000003d0f00 0000000000000000 00007fffb2b6e728
NIP [c0000000000022fc] rfi_flush_fallback+0x7c/0x80
LR [00007fffb2d67e48] 0x7fffb2d67e48
Call Trace:
Instruction dump:
e96a0220 e96a02a8 e96a0330 e96a03b8 394a0400 4200ffdc 7d2903a6 e92d0c00
e94d0c08 e96d0c10 e82d0c18 7db242a6 <4c000024> 7db243a6 7db142a6 f82d0c18
The problem is the signal code assumes TM is enabled when
CONFIG_PPC_TRANSACTIONAL_MEM is enabled. This may not be the case as
with P9 powernv or if `ppc_tm=off` is used on P8.
This means any local user can crash the system.
Fix the problem by returning a bad stack frame to the user if they try
to set the MSR TS bits with sigreturn() on systems where TM is not
supported.
Found with sigfuz kernel selftest on P9.
This fixes CVE-2019-13648.
Fixes: 2b0a576d15e0 ("powerpc: Add new transactional memory state to the signal context")
Cc: stable@vger.kernel.org # v3.9
Reported-by: Praveen Pandey <Praveen.Pandey@in.ibm.com>
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190719050502.405-1-mikey@neuling.org
2019-07-19 13:05:02 +08:00
|
|
|
|
|
|
|
/* Trying to start TM on non TM system */
|
|
|
|
if (!cpu_has_feature(CPU_FTR_TM))
|
|
|
|
goto badframe;
|
|
|
|
|
2013-02-14 00:21:41 +08:00
|
|
|
if (__get_user(uc_transact, &uc->uc_link))
|
|
|
|
goto badframe;
|
2016-09-23 14:18:12 +08:00
|
|
|
if (restore_tm_sigcontexts(current, &uc->uc_mcontext,
|
2013-02-14 00:21:41 +08:00
|
|
|
&uc_transact->uc_mcontext))
|
|
|
|
goto badframe;
|
2019-01-09 21:16:45 +08:00
|
|
|
} else
|
2013-02-14 00:21:41 +08:00
|
|
|
#endif
|
2019-01-09 21:16:45 +08:00
|
|
|
{
|
powerpc/tm: Unset MSR[TS] if not recheckpointing
There is a TM Bad Thing bug that can be caused when you return from a
signal context in a suspended transaction but with ucontext MSR[TS] unset.
This forces regs->msr[TS] to be set at syscall entrance (since the CPU
state is transactional). It also calls treclaim() to flush the transaction
state, which is done based on the live (mfmsr) MSR state.
Since user context MSR[TS] is not set, then restore_tm_sigcontexts() is not
called, thus, not executing recheckpoint, keeping the CPU state as not
transactional. When calling rfid, SRR1 will have MSR[TS] set, but the CPU
state is non transactional, causing the TM Bad Thing with the following
stack:
[ 33.862316] Bad kernel stack pointer 3fffd9dce3e0 at c00000000000c47c
cpu 0x8: Vector: 700 (Program Check) at [c00000003ff7fd40]
pc: c00000000000c47c: fast_exception_return+0xac/0xb4
lr: 00003fff865f442c
sp: 3fffd9dce3e0
msr: 8000000102a03031
current = 0xc00000041f68b700
paca = 0xc00000000fb84800 softe: 0 irq_happened: 0x01
pid = 1721, comm = tm-signal-sigre
Linux version 4.9.0-3-powerpc64le (debian-kernel@lists.debian.org) (gcc version 6.3.0 20170516 (Debian 6.3.0-18) ) #1 SMP Debian 4.9.30-2+deb9u2 (2017-06-26)
WARNING: exception is not recoverable, can't continue
The same problem happens on 32-bits signal handler, and the fix is very
similar, if tm_recheckpoint() is not executed, then regs->msr[TS] should be
zeroed.
This patch also fixes a sparse warning related to lack of indentation when
CONFIG_PPC_TRANSACTIONAL_MEM is set.
Fixes: 2b0a576d15e0e ("powerpc: Add new transactional memory state to the signal context")
CC: Stable <stable@vger.kernel.org> # 3.10+
Signed-off-by: Breno Leitao <leitao@debian.org>
Tested-by: Michal Suchánek <msuchanek@suse.de>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-11-27 04:12:00 +08:00
|
|
|
/*
|
2019-01-09 21:16:45 +08:00
|
|
|
* Fall through, for non-TM restore
|
|
|
|
*
|
powerpc/tm: Unset MSR[TS] if not recheckpointing
There is a TM Bad Thing bug that can be caused when you return from a
signal context in a suspended transaction but with ucontext MSR[TS] unset.
This forces regs->msr[TS] to be set at syscall entrance (since the CPU
state is transactional). It also calls treclaim() to flush the transaction
state, which is done based on the live (mfmsr) MSR state.
Since user context MSR[TS] is not set, then restore_tm_sigcontexts() is not
called, thus, not executing recheckpoint, keeping the CPU state as not
transactional. When calling rfid, SRR1 will have MSR[TS] set, but the CPU
state is non transactional, causing the TM Bad Thing with the following
stack:
[ 33.862316] Bad kernel stack pointer 3fffd9dce3e0 at c00000000000c47c
cpu 0x8: Vector: 700 (Program Check) at [c00000003ff7fd40]
pc: c00000000000c47c: fast_exception_return+0xac/0xb4
lr: 00003fff865f442c
sp: 3fffd9dce3e0
msr: 8000000102a03031
current = 0xc00000041f68b700
paca = 0xc00000000fb84800 softe: 0 irq_happened: 0x01
pid = 1721, comm = tm-signal-sigre
Linux version 4.9.0-3-powerpc64le (debian-kernel@lists.debian.org) (gcc version 6.3.0 20170516 (Debian 6.3.0-18) ) #1 SMP Debian 4.9.30-2+deb9u2 (2017-06-26)
WARNING: exception is not recoverable, can't continue
The same problem happens on 32-bits signal handler, and the fix is very
similar, if tm_recheckpoint() is not executed, then regs->msr[TS] should be
zeroed.
This patch also fixes a sparse warning related to lack of indentation when
CONFIG_PPC_TRANSACTIONAL_MEM is set.
Fixes: 2b0a576d15e0e ("powerpc: Add new transactional memory state to the signal context")
CC: Stable <stable@vger.kernel.org> # 3.10+
Signed-off-by: Breno Leitao <leitao@debian.org>
Tested-by: Michal Suchánek <msuchanek@suse.de>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-11-27 04:12:00 +08:00
|
|
|
* Unset MSR[TS] on the thread regs since MSR from user
|
|
|
|
* context does not have MSR active, and recheckpoint was
|
|
|
|
* not called since restore_tm_sigcontexts() was not called
|
|
|
|
* also.
|
|
|
|
*
|
|
|
|
* If not unsetting it, the code can RFID to userspace with
|
|
|
|
* MSR[TS] set, but without CPU in the proper state,
|
|
|
|
* causing a TM bad thing.
|
|
|
|
*/
|
|
|
|
current->thread.regs->msr &= ~MSR_TS_MASK;
|
|
|
|
if (restore_sigcontext(current, NULL, 1, &uc->uc_mcontext))
|
|
|
|
goto badframe;
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2012-12-23 16:26:46 +08:00
|
|
|
if (restore_altstack(&uc->uc_stack))
|
|
|
|
goto badframe;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
[PATCH] syscall entry/exit revamp
This cleanup patch speeds up the null syscall path on ppc64 by about 3%,
and brings the ppc32 and ppc64 code slightly closer together.
The ppc64 code was checking current_thread_info()->flags twice in the
syscall exit path; once for TIF_SYSCALL_T_OR_A before disabling
interrupts, and then again for TIF_SIGPENDING|TIF_NEED_RESCHED etc after
disabling interrupts. Now we do the same as ppc32 -- check the flags
only once in the fast path, and re-enable interrupts if necessary in the
ptrace case.
The patch abolishes the 'syscall_noerror' member of struct thread_info
and replaces it with a TIF_NOERROR bit in the flags, which is handled in
the slow path. This shortens the syscall entry code, which no longer
needs to clear syscall_noerror.
The patch adds a TIF_SAVE_NVGPRS flag which causes the syscall exit slow
path to save the non-volatile GPRs into a signal frame. This removes the
need for the assembly wrappers around sys_sigsuspend(),
sys_rt_sigsuspend(), et al which existed solely to save those registers
in advance. It also means I don't have to add new wrappers for ppoll()
and pselect(), which is what I was supposed to be doing when I got
distracted into this...
Finally, it unifies the ppc64 and ppc32 methods of handling syscall exit
directly into a signal handler (as required by sigsuspend et al) by
introducing a TIF_RESTOREALL flag which causes _all_ the registers to be
reloaded from the pt_regs by taking the ret_from_exception path, instead
of the normal syscall exit path which stomps on the callee-saved GPRs.
It appears to pass an LTP test run on ppc64, and passes basic testing on
ppc32 too. Brief tests of ptrace functionality with strace and gdb also
appear OK. I wouldn't send it to Linus for 2.6.15 just yet though :)
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-11-16 02:52:18 +08:00
|
|
|
set_thread_flag(TIF_RESTOREALL);
|
|
|
|
return 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
badframe:
|
2011-06-04 13:36:54 +08:00
|
|
|
if (show_unhandled_signals)
|
|
|
|
printk_ratelimited(regs->msr & MSR_64BIT ? fmt64 : fmt32,
|
|
|
|
current->comm, current->pid, "rt_sigreturn",
|
|
|
|
(long)uc, regs->nip, regs->link);
|
2007-10-12 08:20:07 +08:00
|
|
|
|
2019-05-23 23:17:27 +08:00
|
|
|
force_sig(SIGSEGV);
|
2005-04-17 06:20:36 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2016-09-23 14:18:12 +08:00
|
|
|
int handle_rt_signal64(struct ksignal *ksig, sigset_t *set,
|
|
|
|
struct task_struct *tsk)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
struct rt_sigframe __user *frame;
|
|
|
|
unsigned long newsp = 0;
|
|
|
|
long err = 0;
|
2016-09-23 14:18:12 +08:00
|
|
|
struct pt_regs *regs = tsk->thread.regs;
|
|
|
|
|
|
|
|
BUG_ON(tsk != current);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2016-09-23 14:18:12 +08:00
|
|
|
frame = get_sigframe(ksig, get_tm_stackpointer(tsk), sizeof(*frame), 0);
|
2007-06-04 15:22:48 +08:00
|
|
|
if (unlikely(frame == NULL))
|
2005-04-17 06:20:36 +08:00
|
|
|
goto badframe;
|
|
|
|
|
|
|
|
err |= __put_user(&frame->info, &frame->pinfo);
|
|
|
|
err |= __put_user(&frame->uc, &frame->puc);
|
2014-03-02 21:46:11 +08:00
|
|
|
err |= copy_siginfo_to_user(&frame->info, &ksig->info);
|
2005-04-17 06:20:36 +08:00
|
|
|
if (err)
|
|
|
|
goto badframe;
|
|
|
|
|
|
|
|
/* Create the ucontext. */
|
|
|
|
err |= __put_user(0, &frame->uc.uc_flags);
|
2012-12-23 16:26:46 +08:00
|
|
|
err |= __save_altstack(&frame->uc.uc_stack, regs->gpr[1]);
|
2013-02-14 00:21:41 +08:00
|
|
|
#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
|
|
|
|
if (MSR_TM_ACTIVE(regs->msr)) {
|
|
|
|
/* The ucontext_t passed to userland points to the second
|
|
|
|
* ucontext_t (for transactional state) with its uc_link ptr.
|
|
|
|
*/
|
|
|
|
err |= __put_user(&frame->uc_transact, &frame->uc.uc_link);
|
|
|
|
err |= setup_tm_sigcontexts(&frame->uc.uc_mcontext,
|
|
|
|
&frame->uc_transact.uc_mcontext,
|
2016-09-23 14:18:12 +08:00
|
|
|
tsk, ksig->sig, NULL,
|
2014-03-02 21:46:11 +08:00
|
|
|
(unsigned long)ksig->ka.sa.sa_handler);
|
2013-02-14 00:21:41 +08:00
|
|
|
} else
|
|
|
|
#endif
|
|
|
|
{
|
|
|
|
err |= __put_user(0, &frame->uc.uc_link);
|
2016-09-23 14:18:12 +08:00
|
|
|
err |= setup_sigcontext(&frame->uc.uc_mcontext, tsk, ksig->sig,
|
2014-03-02 21:46:11 +08:00
|
|
|
NULL, (unsigned long)ksig->ka.sa.sa_handler,
|
2013-02-14 00:21:41 +08:00
|
|
|
1);
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set));
|
|
|
|
if (err)
|
|
|
|
goto badframe;
|
|
|
|
|
2005-11-14 18:55:15 +08:00
|
|
|
/* Make sure signal handler doesn't get spurious FP exceptions */
|
2016-09-23 14:18:12 +08:00
|
|
|
tsk->thread.fp_state.fpscr = 0;
|
2005-11-14 18:55:15 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/* Set up to return from userspace. */
|
2016-09-23 14:18:12 +08:00
|
|
|
if (vdso64_rt_sigtramp && tsk->mm->context.vdso_base) {
|
|
|
|
regs->link = tsk->mm->context.vdso_base + vdso64_rt_sigtramp;
|
2005-04-17 06:20:36 +08:00
|
|
|
} else {
|
|
|
|
err |= setup_trampoline(__NR_rt_sigreturn, &frame->tramp[0]);
|
|
|
|
if (err)
|
|
|
|
goto badframe;
|
|
|
|
regs->link = (unsigned long) &frame->tramp[0];
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Allocate a dummy caller frame for the signal handler. */
|
2007-06-04 15:22:48 +08:00
|
|
|
newsp = ((unsigned long)frame) - __SIGNAL_FRAMESIZE;
|
2005-04-17 06:20:36 +08:00
|
|
|
err |= put_user(regs->gpr[1], (unsigned long __user *)newsp);
|
|
|
|
|
|
|
|
/* Set up "regs" so we "return" to the signal handler. */
|
2013-11-20 19:15:03 +08:00
|
|
|
if (is_elf2_task()) {
|
2014-03-02 21:46:11 +08:00
|
|
|
regs->nip = (unsigned long) ksig->ka.sa.sa_handler;
|
2013-11-20 19:15:03 +08:00
|
|
|
regs->gpr[12] = regs->nip;
|
|
|
|
} else {
|
|
|
|
/* Handler is *really* a pointer to the function descriptor for
|
|
|
|
* the signal routine. The first entry in the function
|
|
|
|
* descriptor is the entry address of signal and the second
|
|
|
|
* entry is the TOC value we need to use.
|
|
|
|
*/
|
|
|
|
func_descr_t __user *funct_desc_ptr =
|
2014-03-02 21:46:11 +08:00
|
|
|
(func_descr_t __user *) ksig->ka.sa.sa_handler;
|
2013-11-20 19:15:03 +08:00
|
|
|
|
|
|
|
err |= get_user(regs->nip, &funct_desc_ptr->entry);
|
|
|
|
err |= get_user(regs->gpr[2], &funct_desc_ptr->toc);
|
|
|
|
}
|
|
|
|
|
2013-09-23 10:04:43 +08:00
|
|
|
/* enter the signal handler in native-endian mode */
|
2006-06-07 14:14:40 +08:00
|
|
|
regs->msr &= ~MSR_LE;
|
2013-09-23 10:04:43 +08:00
|
|
|
regs->msr |= (MSR_KERNEL & MSR_LE);
|
2005-04-17 06:20:36 +08:00
|
|
|
regs->gpr[1] = newsp;
|
2014-03-02 21:46:11 +08:00
|
|
|
regs->gpr[3] = ksig->sig;
|
2005-04-17 06:20:36 +08:00
|
|
|
regs->result = 0;
|
2014-03-02 21:46:11 +08:00
|
|
|
if (ksig->ka.sa.sa_flags & SA_SIGINFO) {
|
2005-04-17 06:20:36 +08:00
|
|
|
err |= get_user(regs->gpr[4], (unsigned long __user *)&frame->pinfo);
|
|
|
|
err |= get_user(regs->gpr[5], (unsigned long __user *)&frame->puc);
|
|
|
|
regs->gpr[6] = (unsigned long) frame;
|
|
|
|
} else {
|
|
|
|
regs->gpr[4] = (unsigned long)&frame->uc.uc_mcontext;
|
|
|
|
}
|
|
|
|
if (err)
|
|
|
|
goto badframe;
|
|
|
|
|
2014-03-02 21:46:11 +08:00
|
|
|
return 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
badframe:
|
2011-06-04 13:36:54 +08:00
|
|
|
if (show_unhandled_signals)
|
|
|
|
printk_ratelimited(regs->msr & MSR_64BIT ? fmt64 : fmt32,
|
2016-09-23 14:18:12 +08:00
|
|
|
tsk->comm, tsk->pid, "setup_rt_frame",
|
2011-06-04 13:36:54 +08:00
|
|
|
(long)frame, regs->nip, regs->link);
|
2007-10-12 08:20:07 +08:00
|
|
|
|
2014-03-02 21:46:11 +08:00
|
|
|
return 1;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|