Commit 42a87d41 authored by Harald Freudenberger's avatar Harald Freudenberger Committed by Martin Schwidefsky

s390/zcrypt: make sysfs reset attribute trigger queue reset

Until now there is no way to reset a AP queue or card. Driving a card
or queue offline and online again does only toggle the 'software'
online state. The only way to trigger a (hardware) reset is by running
hot-unplug/hot-plug for example on the HMC.

This patch makes the queue reset attribute in sysfs writable.
Writing into this attribute triggers a reset on the AP queue's state
machine. So the AP queue is flushed and state machine runs through the
initial states which cause a reset (PQAP(RAPQ)) and a re-registration
to interrupts (PQAP(AQIC)) if available.

The reset sysfs attribute is writable by root only. So only an
administrator is allowed to initiate a reset of AP queues. Please note
that the queue's counter values are left untouched by the reset.
Signed-off-by: default avatarHarald Freudenberger <freude@linux.ibm.com>
Signed-off-by: default avatarMartin Schwidefsky <schwidefsky@de.ibm.com>
parent 98dfd326
......@@ -14,6 +14,9 @@
#include <asm/facility.h>
#include "ap_bus.h"
#include "ap_debug.h"
static void __ap_flush_queue(struct ap_queue *aq);
/**
* ap_queue_enable_interruption(): Enable interruption on an AP queue.
......@@ -541,7 +544,25 @@ static ssize_t reset_show(struct device *dev,
return rc;
}
static DEVICE_ATTR_RO(reset);
static ssize_t reset_store(struct device *dev,
struct device_attribute *attr,
const char *buf, size_t count)
{
struct ap_queue *aq = to_ap_queue(dev);
spin_lock_bh(&aq->lock);
__ap_flush_queue(aq);
aq->state = AP_STATE_RESET_START;
ap_wait(ap_sm_event(aq, AP_EVENT_POLL));
spin_unlock_bh(&aq->lock);
AP_DBF(DBF_INFO, "reset queue=%02x.%04x triggered by user\n",
AP_QID_CARD(aq->qid), AP_QID_QUEUE(aq->qid));
return count;
}
static DEVICE_ATTR_RW(reset);
static ssize_t interrupt_show(struct device *dev,
struct device_attribute *attr, char *buf)
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment