linux/arch/s390/kernel/nmi.c

367 lines
9.6 KiB
C
Raw Normal View History

/*
* Machine check handler
*
* Copyright IBM Corp. 2000, 2009
* Author(s): Ingo Adlung <adlung@de.ibm.com>,
* Martin Schwidefsky <schwidefsky@de.ibm.com>,
* Cornelia Huck <cornelia.huck@de.ibm.com>,
* Heiko Carstens <heiko.carstens@de.ibm.com>,
*/
#include <linux/kernel_stat.h>
#include <linux/init.h>
#include <linux/errno.h>
#include <linux/hardirq.h>
#include <linux/time.h>
#include <linux/module.h>
#include <asm/lowcore.h>
#include <asm/smp.h>
#include <asm/etr.h>
#include <asm/cputime.h>
#include <asm/nmi.h>
#include <asm/crw.h>
#include <asm/switch_to.h>
#include <asm/fpu-internal.h>
s390/nmi: fix vector register corruption If a machine check happens, the machine has the vector facility installed and the extended save area exists, the cpu will save vector register contents into the extended save area. This is regardless of control register 0 contents, which enables and disables the vector facility during runtime. On each machine check we should validate the vector registers. The current code however tries to validate the registers only if the running task is using vector registers in user space. However even the current code is broken and causes vector register corruption on machine checks, if user space uses them: the prefix area contains a pointer (absolute address) to the machine check extended save area. In order to save some space the save area was put into an unused area of the second prefix page. When validating vector register contents the code uses the absolute address of the extended save area, which is wrong. Due to prefixing the vector instructions will then access contents using absolute addresses instead of real addresses, where the machine stored the contents. If the above would work there is still the problem that register validition would only happen if user space uses vector registers. If kernel space uses them also, this may also lead to vector register content corruption: if the kernel makes use of vector instructions, but the current running user space context does not, the machine check handler will validate floating point registers instead of vector registers. Given the fact that writing to a floating point register may change the upper halve of the corresponding vector register, we also experience vector register corruption in this case. Fix all of these issues, and always validate vector registers on each machine check, if the machine has the vector facility installed and the extended save area is defined. Cc: <stable@vger.kernel.org> # 4.1+ Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-07-07 14:40:49 +08:00
#include <asm/ctl_reg.h>
struct mcck_struct {
int kill_task;
int channel_report;
int warning;
unsigned long long mcck_code;
};
static DEFINE_PER_CPU(struct mcck_struct, cpu_mcck);
static void s390_handle_damage(char *msg)
{
smp_send_stop();
disabled_wait((unsigned long) __builtin_return_address(0));
while (1);
}
/*
* Main machine check handler function. Will be called with interrupts enabled
* or disabled and machine checks enabled or disabled.
*/
void s390_handle_mcck(void)
{
unsigned long flags;
struct mcck_struct mcck;
/*
* Disable machine checks and get the current state of accumulated
* machine checks. Afterwards delete the old state and enable machine
* checks again.
*/
local_irq_save(flags);
local_mcck_disable();
mcck = *this_cpu_ptr(&cpu_mcck);
memset(this_cpu_ptr(&cpu_mcck), 0, sizeof(mcck));
clear_cpu_flag(CIF_MCCK_PENDING);
local_mcck_enable();
local_irq_restore(flags);
if (mcck.channel_report)
crw_handle_channel_report();
/*
* A warning may remain for a prolonged period on the bare iron.
* (actually until the machine is powered off, or the problem is gone)
* So we just stop listening for the WARNING MCH and avoid continuously
* being interrupted. One caveat is however, that we must do this per
* processor and cannot use the smp version of ctl_clear_bit().
* On VM we only get one interrupt per virtally presented machinecheck.
* Though one suffices, we may get one interrupt per (virtual) cpu.
*/
if (mcck.warning) { /* WARNING pending ? */
static int mchchk_wng_posted = 0;
/* Use single cpu clear, as we cannot handle smp here. */
__ctl_clear_bit(14, 24); /* Disable WARNING MCH */
if (xchg(&mchchk_wng_posted, 1) == 0)
kill_cad_pid(SIGPWR, 1);
}
if (mcck.kill_task) {
local_irq_enable();
printk(KERN_EMERG "mcck: Terminating task because of machine "
"malfunction (code 0x%016llx).\n", mcck.mcck_code);
printk(KERN_EMERG "mcck: task: %s, pid: %d.\n",
current->comm, current->pid);
do_exit(SIGSEGV);
}
}
EXPORT_SYMBOL_GPL(s390_handle_mcck);
/*
* returns 0 if all registers could be validated
* returns 1 otherwise
*/
static int notrace s390_revalidate_registers(struct mci *mci)
{
int kill_task;
u64 zero;
void *fpt_save_area, *fpt_creg_save_area;
kill_task = 0;
zero = 0;
if (!mci->gr) {
/*
* General purpose registers couldn't be restored and have
* unknown contents. Process needs to be terminated.
*/
kill_task = 1;
}
if (!mci->fp) {
/*
* Floating point registers can't be restored and
* therefore the process needs to be terminated.
*/
kill_task = 1;
}
fpt_save_area = &S390_lowcore.floating_pt_save_area;
fpt_creg_save_area = &S390_lowcore.fpt_creg_save_area;
if (!mci->fc) {
/*
* Floating point control register can't be restored.
* Task will be terminated.
*/
asm volatile("lfpc 0(%0)" : : "a" (&zero), "m" (zero));
kill_task = 1;
} else
asm volatile("lfpc 0(%0)" : : "a" (fpt_creg_save_area));
s390/nmi: fix vector register corruption If a machine check happens, the machine has the vector facility installed and the extended save area exists, the cpu will save vector register contents into the extended save area. This is regardless of control register 0 contents, which enables and disables the vector facility during runtime. On each machine check we should validate the vector registers. The current code however tries to validate the registers only if the running task is using vector registers in user space. However even the current code is broken and causes vector register corruption on machine checks, if user space uses them: the prefix area contains a pointer (absolute address) to the machine check extended save area. In order to save some space the save area was put into an unused area of the second prefix page. When validating vector register contents the code uses the absolute address of the extended save area, which is wrong. Due to prefixing the vector instructions will then access contents using absolute addresses instead of real addresses, where the machine stored the contents. If the above would work there is still the problem that register validition would only happen if user space uses vector registers. If kernel space uses them also, this may also lead to vector register content corruption: if the kernel makes use of vector instructions, but the current running user space context does not, the machine check handler will validate floating point registers instead of vector registers. Given the fact that writing to a floating point register may change the upper halve of the corresponding vector register, we also experience vector register corruption in this case. Fix all of these issues, and always validate vector registers on each machine check, if the machine has the vector facility installed and the extended save area is defined. Cc: <stable@vger.kernel.org> # 4.1+ Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-07-07 14:40:49 +08:00
if (!MACHINE_HAS_VX) {
/* Revalidate floating point registers */
asm volatile(
" ld 0,0(%0)\n"
" ld 1,8(%0)\n"
" ld 2,16(%0)\n"
" ld 3,24(%0)\n"
" ld 4,32(%0)\n"
" ld 5,40(%0)\n"
" ld 6,48(%0)\n"
" ld 7,56(%0)\n"
" ld 8,64(%0)\n"
" ld 9,72(%0)\n"
" ld 10,80(%0)\n"
" ld 11,88(%0)\n"
" ld 12,96(%0)\n"
" ld 13,104(%0)\n"
" ld 14,112(%0)\n"
" ld 15,120(%0)\n"
: : "a" (fpt_save_area));
} else {
/* Revalidate vector registers */
union ctlreg0 cr0;
if (!mci->vr) {
/*
* Vector registers can't be restored and therefore
* the process needs to be terminated.
*/
kill_task = 1;
}
s390/nmi: fix vector register corruption If a machine check happens, the machine has the vector facility installed and the extended save area exists, the cpu will save vector register contents into the extended save area. This is regardless of control register 0 contents, which enables and disables the vector facility during runtime. On each machine check we should validate the vector registers. The current code however tries to validate the registers only if the running task is using vector registers in user space. However even the current code is broken and causes vector register corruption on machine checks, if user space uses them: the prefix area contains a pointer (absolute address) to the machine check extended save area. In order to save some space the save area was put into an unused area of the second prefix page. When validating vector register contents the code uses the absolute address of the extended save area, which is wrong. Due to prefixing the vector instructions will then access contents using absolute addresses instead of real addresses, where the machine stored the contents. If the above would work there is still the problem that register validition would only happen if user space uses vector registers. If kernel space uses them also, this may also lead to vector register content corruption: if the kernel makes use of vector instructions, but the current running user space context does not, the machine check handler will validate floating point registers instead of vector registers. Given the fact that writing to a floating point register may change the upper halve of the corresponding vector register, we also experience vector register corruption in this case. Fix all of these issues, and always validate vector registers on each machine check, if the machine has the vector facility installed and the extended save area is defined. Cc: <stable@vger.kernel.org> # 4.1+ Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-07-07 14:40:49 +08:00
cr0.val = S390_lowcore.cregs_save_area[0];
cr0.afp = cr0.vx = 1;
__ctl_load(cr0.val, 0, 0);
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 18:53:42 +08:00
asm volatile(
" la 1,%0\n"
" .word 0xe70f,0x1000,0x0036\n" /* vlm 0,15,0(1) */
" .word 0xe70f,0x1100,0x0c36\n" /* vlm 16,31,256(1) */
: : "Q" (*(struct vx_array *)
&S390_lowcore.vector_save_area) : "1");
s390/nmi: fix vector register corruption If a machine check happens, the machine has the vector facility installed and the extended save area exists, the cpu will save vector register contents into the extended save area. This is regardless of control register 0 contents, which enables and disables the vector facility during runtime. On each machine check we should validate the vector registers. The current code however tries to validate the registers only if the running task is using vector registers in user space. However even the current code is broken and causes vector register corruption on machine checks, if user space uses them: the prefix area contains a pointer (absolute address) to the machine check extended save area. In order to save some space the save area was put into an unused area of the second prefix page. When validating vector register contents the code uses the absolute address of the extended save area, which is wrong. Due to prefixing the vector instructions will then access contents using absolute addresses instead of real addresses, where the machine stored the contents. If the above would work there is still the problem that register validition would only happen if user space uses vector registers. If kernel space uses them also, this may also lead to vector register content corruption: if the kernel makes use of vector instructions, but the current running user space context does not, the machine check handler will validate floating point registers instead of vector registers. Given the fact that writing to a floating point register may change the upper halve of the corresponding vector register, we also experience vector register corruption in this case. Fix all of these issues, and always validate vector registers on each machine check, if the machine has the vector facility installed and the extended save area is defined. Cc: <stable@vger.kernel.org> # 4.1+ Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-07-07 14:40:49 +08:00
__ctl_load(S390_lowcore.cregs_save_area[0], 0, 0);
}
/* Revalidate access registers */
asm volatile(
" lam 0,15,0(%0)"
: : "a" (&S390_lowcore.access_regs_save_area));
if (!mci->ar) {
/*
* Access registers have unknown contents.
* Terminating task.
*/
kill_task = 1;
}
/* Revalidate control registers */
if (!mci->cr) {
/*
* Control registers have unknown contents.
* Can't recover and therefore stopping machine.
*/
s390_handle_damage("invalid control registers.");
} else {
asm volatile(
" lctlg 0,15,0(%0)"
: : "a" (&S390_lowcore.cregs_save_area));
}
/*
* We don't even try to revalidate the TOD register, since we simply
* can't write something sensible into that register.
*/
/*
* See if we can revalidate the TOD programmable register with its
* old contents (should be zero) otherwise set it to zero.
*/
if (!mci->pr)
asm volatile(
" sr 0,0\n"
" sckpf"
: : : "0", "cc");
else
asm volatile(
" l 0,0(%0)\n"
" sckpf"
: : "a" (&S390_lowcore.tod_progreg_save_area)
: "0", "cc");
/* Revalidate clock comparator register */
set_clock_comparator(S390_lowcore.clock_comparator);
/* Check if old PSW is valid */
if (!mci->wp)
/*
* Can't tell if we come from user or kernel mode
* -> stopping machine.
*/
s390_handle_damage("old psw invalid.");
if (!mci->ms || !mci->pm || !mci->ia)
kill_task = 1;
return kill_task;
}
#define MAX_IPD_COUNT 29
#define MAX_IPD_TIME (5 * 60 * USEC_PER_SEC) /* 5 minutes */
#define ED_STP_ISLAND 6 /* External damage STP island check */
#define ED_STP_SYNC 7 /* External damage STP sync check */
#define ED_ETR_SYNC 12 /* External damage ETR sync check */
#define ED_ETR_SWITCH 13 /* External damage ETR switch to local */
/*
* machine check handler.
*/
void notrace s390_do_machine_check(struct pt_regs *regs)
{
static int ipd_count;
static DEFINE_SPINLOCK(ipd_lock);
static unsigned long long last_ipd;
struct mcck_struct *mcck;
unsigned long long tmp;
struct mci *mci;
int umode;
nmi_enter();
inc_irq_stat(NMI_NMI);
mci = (struct mci *) &S390_lowcore.mcck_interruption_code;
s390: Replace __get_cpu_var uses __get_cpu_var() is used for multiple purposes in the kernel source. One of them is address calculation via the form &__get_cpu_var(x). This calculates the address for the instance of the percpu variable of the current processor based on an offset. Other use cases are for storing and retrieving data from the current processors percpu area. __get_cpu_var() can be used as an lvalue when writing data or on the right side of an assignment. __get_cpu_var() is defined as : #define __get_cpu_var(var) (*this_cpu_ptr(&(var))) __get_cpu_var() always only does an address determination. However, store and retrieve operations could use a segment prefix (or global register on other platforms) to avoid the address calculation. this_cpu_write() and this_cpu_read() can directly take an offset into a percpu area and use optimized assembly code to read and write per cpu variables. This patch converts __get_cpu_var into either an explicit address calculation using this_cpu_ptr() or into a use of this_cpu operations that use the offset. Thereby address calculations are avoided and less registers are used when code is generated. At the end of the patch set all uses of __get_cpu_var have been removed so the macro is removed too. The patch set includes passes over all arches as well. Once these operations are used throughout then specialized macros can be defined in non -x86 arches as well in order to optimize per cpu access by f.e. using a global register that may be set to the per cpu base. Transformations done to __get_cpu_var() 1. Determine the address of the percpu instance of the current processor. DEFINE_PER_CPU(int, y); int *x = &__get_cpu_var(y); Converts to int *x = this_cpu_ptr(&y); 2. Same as #1 but this time an array structure is involved. DEFINE_PER_CPU(int, y[20]); int *x = __get_cpu_var(y); Converts to int *x = this_cpu_ptr(y); 3. Retrieve the content of the current processors instance of a per cpu variable. DEFINE_PER_CPU(int, y); int x = __get_cpu_var(y) Converts to int x = __this_cpu_read(y); 4. Retrieve the content of a percpu struct DEFINE_PER_CPU(struct mystruct, y); struct mystruct x = __get_cpu_var(y); Converts to memcpy(&x, this_cpu_ptr(&y), sizeof(x)); 5. Assignment to a per cpu variable DEFINE_PER_CPU(int, y) __get_cpu_var(y) = x; Converts to this_cpu_write(y, x); 6. Increment/Decrement etc of a per cpu variable DEFINE_PER_CPU(int, y); __get_cpu_var(y)++ Converts to this_cpu_inc(y) Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> CC: linux390@de.ibm.com Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Christoph Lameter <cl@linux.com> Signed-off-by: Tejun Heo <tj@kernel.org>
2014-08-18 01:30:45 +08:00
mcck = this_cpu_ptr(&cpu_mcck);
umode = user_mode(regs);
if (mci->sd) {
/* System damage -> stopping machine */
s390_handle_damage("received system damage machine check.");
}
if (mci->pd) {
if (mci->b) {
/* Processing backup -> verify if we can survive this */
u64 z_mcic, o_mcic, t_mcic;
z_mcic = (1ULL<<63 | 1ULL<<59 | 1ULL<<29);
o_mcic = (1ULL<<43 | 1ULL<<42 | 1ULL<<41 | 1ULL<<40 |
1ULL<<36 | 1ULL<<35 | 1ULL<<34 | 1ULL<<32 |
1ULL<<30 | 1ULL<<21 | 1ULL<<20 | 1ULL<<17 |
1ULL<<16);
t_mcic = *(u64 *)mci;
if (((t_mcic & z_mcic) != 0) ||
((t_mcic & o_mcic) != o_mcic)) {
s390_handle_damage("processing backup machine "
"check with damage.");
}
/*
* Nullifying exigent condition, therefore we might
* retry this instruction.
*/
spin_lock(&ipd_lock);
tmp = get_tod_clock();
if (((tmp - last_ipd) >> 12) < MAX_IPD_TIME)
ipd_count++;
else
ipd_count = 1;
last_ipd = tmp;
if (ipd_count == MAX_IPD_COUNT)
s390_handle_damage("too many ipd retries.");
spin_unlock(&ipd_lock);
} else {
/* Processing damage -> stopping machine */
s390_handle_damage("received instruction processing "
"damage machine check.");
}
}
if (s390_revalidate_registers(mci)) {
if (umode) {
/*
* Couldn't restore all register contents while in
* user mode -> mark task for termination.
*/
mcck->kill_task = 1;
mcck->mcck_code = *(unsigned long long *) mci;
set_cpu_flag(CIF_MCCK_PENDING);
} else {
/*
* Couldn't restore all register contents while in
* kernel mode -> stopping machine.
*/
s390_handle_damage("unable to revalidate registers.");
}
}
if (mci->cd) {
/* Timing facility damage */
s390_handle_damage("TOD clock damaged");
}
if (mci->ed && mci->ec) {
/* External damage */
if (S390_lowcore.external_damage_code & (1U << ED_ETR_SYNC))
etr_sync_check();
if (S390_lowcore.external_damage_code & (1U << ED_ETR_SWITCH))
etr_switch_to_local();
if (S390_lowcore.external_damage_code & (1U << ED_STP_SYNC))
stp_sync_check();
if (S390_lowcore.external_damage_code & (1U << ED_STP_ISLAND))
stp_island_check();
}
if (mci->se)
/* Storage error uncorrected */
s390_handle_damage("received storage error uncorrected "
"machine check.");
if (mci->ke)
/* Storage key-error uncorrected */
s390_handle_damage("received storage key-error uncorrected "
"machine check.");
if (mci->ds && mci->fa)
/* Storage degradation */
s390_handle_damage("received storage degradation machine "
"check.");
if (mci->cp) {
/* Channel report word pending */
mcck->channel_report = 1;
set_cpu_flag(CIF_MCCK_PENDING);
}
if (mci->w) {
/* Warning pending */
mcck->warning = 1;
set_cpu_flag(CIF_MCCK_PENDING);
}
nmi_exit();
}
static int __init machine_check_init(void)
{
ctl_set_bit(14, 25); /* enable external damage MCH */
ctl_set_bit(14, 27); /* enable system recovery MCH */
ctl_set_bit(14, 24); /* enable warning MCH */
return 0;
}
early_initcall(machine_check_init);