#include <linux/sched.h>
#include <linux/errno.h>
#include <linux/workqueue.h>
+#include <linux/time.h>
+#include <linux/kthread.h>
#include <asm/lowcore.h>
static struct semaphore m_sem;
-extern int css_process_crw(int);
+extern int css_process_crw(int, int);
extern int chsc_process_crw(void);
extern int chp_process_crw(int, int);
extern void css_reiterate_subchannels(void);
static int
s390_collect_crw_info(void *param)
{
- struct crw crw;
+ struct crw crw[2];
int ccode, ret, slow;
struct semaphore *sem;
+ unsigned int chain;
sem = (struct semaphore *)param;
- /* Set a nice name. */
- daemonize("kmcheck");
repeat:
down_interruptible(sem);
slow = 0;
+ chain = 0;
while (1) {
- ccode = stcrw(&crw);
+ if (unlikely(chain > 1)) {
+ struct crw tmp_crw;
+
+ printk(KERN_WARNING"%s: Code does not support more "
+ "than two chained crws; please report to "
+ "linux390@de.ibm.com!\n", __FUNCTION__);
+ ccode = stcrw(&tmp_crw);
+ printk(KERN_WARNING"%s: crw reports slct=%d, oflw=%d, "
+ "chn=%d, rsc=%X, anc=%d, erc=%X, rsid=%X\n",
+ __FUNCTION__, tmp_crw.slct, tmp_crw.oflw,
+ tmp_crw.chn, tmp_crw.rsc, tmp_crw.anc,
+ tmp_crw.erc, tmp_crw.rsid);
+ printk(KERN_WARNING"%s: This was crw number %x in the "
+ "chain\n", __FUNCTION__, chain);
+ if (ccode != 0)
+ break;
+ chain = tmp_crw.chn ? chain + 1 : 0;
+ continue;
+ }
+ ccode = stcrw(&crw[chain]);
if (ccode != 0)
break;
DBG(KERN_DEBUG "crw_info : CRW reports slct=%d, oflw=%d, "
"chn=%d, rsc=%X, anc=%d, erc=%X, rsid=%X\n",
- crw.slct, crw.oflw, crw.chn, crw.rsc, crw.anc,
- crw.erc, crw.rsid);
+ crw[chain].slct, crw[chain].oflw, crw[chain].chn,
+ crw[chain].rsc, crw[chain].anc, crw[chain].erc,
+ crw[chain].rsid);
/* Check for overflows. */
- if (crw.oflw) {
+ if (crw[chain].oflw) {
pr_debug("%s: crw overflow detected!\n", __FUNCTION__);
css_reiterate_subchannels();
+ chain = 0;
slow = 1;
continue;
}
- switch (crw.rsc) {
+ switch (crw[chain].rsc) {
case CRW_RSC_SCH:
- pr_debug("source is subchannel %04X\n", crw.rsid);
- ret = css_process_crw (crw.rsid);
+ if (crw[0].chn && !chain)
+ break;
+ pr_debug("source is subchannel %04X\n", crw[0].rsid);
+ ret = css_process_crw (crw[0].rsid,
+ chain ? crw[1].rsid : 0);
if (ret == -EAGAIN)
slow = 1;
break;
pr_debug("source is monitoring facility\n");
break;
case CRW_RSC_CPATH:
- pr_debug("source is channel path %02X\n", crw.rsid);
- switch (crw.erc) {
+ pr_debug("source is channel path %02X\n", crw[0].rsid);
+ switch (crw[0].erc) {
case CRW_ERC_IPARM: /* Path has come. */
- ret = chp_process_crw(crw.rsid, 1);
+ ret = chp_process_crw(crw[0].rsid, 1);
break;
case CRW_ERC_PERRI: /* Path has gone. */
case CRW_ERC_PERRN:
- ret = chp_process_crw(crw.rsid, 0);
+ ret = chp_process_crw(crw[0].rsid, 0);
break;
default:
pr_debug("Don't know how to handle erc=%x\n",
- crw.erc);
+ crw[0].erc);
ret = 0;
}
if (ret == -EAGAIN)
pr_debug("unknown source\n");
break;
}
+ /* chain is always 0 or 1 here. */
+ chain = crw[chain].chn ? chain + 1 : 0;
}
if (slow)
queue_work(slow_path_wq, &slow_path_work);
*/
kill_task = 1;
-#ifndef __s390x__
+#ifndef CONFIG_64BIT
asm volatile("ld 0,0(%0)\n"
"ld 2,8(%0)\n"
"ld 4,16(%0)\n"
#endif
if (MACHINE_HAS_IEEE) {
-#ifdef __s390x__
+#ifdef CONFIG_64BIT
fpt_save_area = &S390_lowcore.floating_pt_save_area;
fpt_creg_save_area = &S390_lowcore.fpt_creg_save_area;
#else
*/
s390_handle_damage("invalid control registers.");
else
-#ifdef __s390x__
+#ifdef CONFIG_64BIT
asm volatile("lctlg 0,15,0(%0)"
: : "a" (&S390_lowcore.cregs_save_area));
#else
* can't write something sensible into that register.
*/
-#ifdef __s390x__
+#ifdef CONFIG_64BIT
/*
* See if we can revalidate the TOD programmable register with its
* old contents (should be zero) otherwise set it to zero.
return kill_task;
}
+#define MAX_IPD_COUNT 29
+#define MAX_IPD_TIME (5 * 60 * USEC_PER_SEC) /* 5 minutes */
+
/*
* machine check handler.
*/
void
s390_do_machine_check(struct pt_regs *regs)
{
+ static DEFINE_SPINLOCK(ipd_lock);
+ static unsigned long long last_ipd;
+ static int ipd_count;
+ unsigned long long tmp;
struct mci *mci;
struct mcck_struct *mcck;
int umode;
if (mci->b) {
/* Processing backup -> verify if we can survive this */
u64 z_mcic, o_mcic, t_mcic;
-#ifdef __s390x__
+#ifdef CONFIG_64BIT
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 |
s390_handle_damage("processing backup machine "
"check with damage.");
}
- if (!umode)
- s390_handle_damage("processing backup machine "
- "check in kernel mode.");
- mcck->kill_task = 1;
- mcck->mcck_code = *(unsigned long long *) mci;
+
+ /*
+ * Nullifying exigent condition, therefore we might
+ * retry this instruction.
+ */
+
+ spin_lock(&ipd_lock);
+
+ tmp = get_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 */
static int __init
machine_check_crw_init (void)
{
- kernel_thread(s390_collect_crw_info, &m_sem, CLONE_FS|CLONE_FILES);
+ kthread_run(s390_collect_crw_info, &m_sem, "kmcheck");
ctl_set_bit(14, 28); /* enable channel report MCH */
return 0;
}