[SCSI] ibmvscsi: Fix softlockup on resume

This fixes a softlockup seen on resume. During resume, the CRQ
must be reenabled. However, the H_ENABLE_CRQ hcall used to do
this may return H_BUSY or H_LONG_BUSY. When this happens, the
caller is expected to retry later. This patch changes a simple
loop, which was causing the softlockup, to a loop at task level
which sleeps between retries rather than simply spinning.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
This commit is contained in:
Brian King 2010-06-17 13:56:00 -05:00 committed by James Bottomley
parent 06395193b2
commit 0f33ece5bc
3 changed files with 106 additions and 31 deletions

View File

@ -73,6 +73,7 @@
#include <linux/slab.h>
#include <linux/of.h>
#include <linux/pm.h>
#include <linux/kthread.h>
#include <asm/firmware.h>
#include <asm/vio.h>
#include <scsi/scsi.h>
@ -504,14 +505,8 @@ static void ibmvscsi_reset_host(struct ibmvscsi_host_data *hostdata)
atomic_set(&hostdata->request_limit, 0);
purge_requests(hostdata, DID_ERROR);
if ((ibmvscsi_ops->reset_crq_queue(&hostdata->queue, hostdata)) ||
(ibmvscsi_ops->send_crq(hostdata, 0xC001000000000000LL, 0)) ||
(vio_enable_interrupts(to_vio_dev(hostdata->dev)))) {
atomic_set(&hostdata->request_limit, -1);
dev_err(hostdata->dev, "error after reset\n");
}
scsi_unblock_requests(hostdata->host);
hostdata->reset_crq = 1;
wake_up(&hostdata->work_wait_q);
}
/**
@ -1462,30 +1457,14 @@ void ibmvscsi_handle_crq(struct viosrp_crq *crq,
/* We need to re-setup the interpartition connection */
dev_info(hostdata->dev, "Re-enabling adapter!\n");
hostdata->client_migrated = 1;
hostdata->reenable_crq = 1;
purge_requests(hostdata, DID_REQUEUE);
if ((ibmvscsi_ops->reenable_crq_queue(&hostdata->queue,
hostdata)) ||
(ibmvscsi_ops->send_crq(hostdata,
0xC001000000000000LL, 0))) {
atomic_set(&hostdata->request_limit,
-1);
dev_err(hostdata->dev, "error after enable\n");
}
wake_up(&hostdata->work_wait_q);
} else {
dev_err(hostdata->dev, "Virtual adapter failed rc %d!\n",
crq->format);
purge_requests(hostdata, DID_ERROR);
if ((ibmvscsi_ops->reset_crq_queue(&hostdata->queue,
hostdata)) ||
(ibmvscsi_ops->send_crq(hostdata,
0xC001000000000000LL, 0))) {
atomic_set(&hostdata->request_limit,
-1);
dev_err(hostdata->dev, "error after reset\n");
}
ibmvscsi_reset_host(hostdata);
}
scsi_unblock_requests(hostdata->host);
return;
case 0x80: /* real payload */
break;
@ -1850,6 +1829,75 @@ static unsigned long ibmvscsi_get_desired_dma(struct vio_dev *vdev)
return desired_io;
}
static void ibmvscsi_do_work(struct ibmvscsi_host_data *hostdata)
{
int rc;
char *action = "reset";
if (hostdata->reset_crq) {
smp_rmb();
hostdata->reset_crq = 0;
rc = ibmvscsi_ops->reset_crq_queue(&hostdata->queue, hostdata);
if (!rc)
rc = ibmvscsi_ops->send_crq(hostdata, 0xC001000000000000LL, 0);
if (!rc)
rc = vio_enable_interrupts(to_vio_dev(hostdata->dev));
} else if (hostdata->reenable_crq) {
smp_rmb();
action = "enable";
rc = ibmvscsi_ops->reenable_crq_queue(&hostdata->queue, hostdata);
hostdata->reenable_crq = 0;
if (!rc)
rc = ibmvscsi_ops->send_crq(hostdata, 0xC001000000000000LL, 0);
} else
return;
if (rc) {
atomic_set(&hostdata->request_limit, -1);
dev_err(hostdata->dev, "error after %s\n", action);
}
scsi_unblock_requests(hostdata->host);
}
static int ibmvscsi_work_to_do(struct ibmvscsi_host_data *hostdata)
{
if (kthread_should_stop())
return 1;
else if (hostdata->reset_crq) {
smp_rmb();
return 1;
} else if (hostdata->reenable_crq) {
smp_rmb();
return 1;
}
return 0;
}
static int ibmvscsi_work(void *data)
{
struct ibmvscsi_host_data *hostdata = data;
int rc;
set_user_nice(current, -20);
while (1) {
rc = wait_event_interruptible(hostdata->work_wait_q,
ibmvscsi_work_to_do(hostdata));
BUG_ON(rc);
if (kthread_should_stop())
break;
ibmvscsi_do_work(hostdata);
}
return 0;
}
/**
* Called by bus code for each adapter
*/
@ -1875,6 +1923,7 @@ static int ibmvscsi_probe(struct vio_dev *vdev, const struct vio_device_id *id)
hostdata = shost_priv(host);
memset(hostdata, 0x00, sizeof(*hostdata));
INIT_LIST_HEAD(&hostdata->sent);
init_waitqueue_head(&hostdata->work_wait_q);
hostdata->host = host;
hostdata->dev = dev;
atomic_set(&hostdata->request_limit, -1);
@ -1885,10 +1934,19 @@ static int ibmvscsi_probe(struct vio_dev *vdev, const struct vio_device_id *id)
goto persist_bufs_failed;
}
hostdata->work_thread = kthread_run(ibmvscsi_work, hostdata, "%s_%d",
"ibmvscsi", host->host_no);
if (IS_ERR(hostdata->work_thread)) {
dev_err(&vdev->dev, "couldn't initialize kthread. rc=%ld\n",
PTR_ERR(hostdata->work_thread));
goto init_crq_failed;
}
rc = ibmvscsi_ops->init_crq_queue(&hostdata->queue, hostdata, max_events);
if (rc != 0 && rc != H_RESOURCE) {
dev_err(&vdev->dev, "couldn't initialize crq. rc=%d\n", rc);
goto init_crq_failed;
goto kill_kthread;
}
if (initialize_event_pool(&hostdata->pool, max_events, hostdata) != 0) {
dev_err(&vdev->dev, "couldn't initialize event pool\n");
@ -1944,6 +2002,8 @@ static int ibmvscsi_probe(struct vio_dev *vdev, const struct vio_device_id *id)
release_event_pool(&hostdata->pool, hostdata);
init_pool_failed:
ibmvscsi_ops->release_crq_queue(&hostdata->queue, hostdata, max_events);
kill_kthread:
kthread_stop(hostdata->work_thread);
init_crq_failed:
unmap_persist_bufs(hostdata);
persist_bufs_failed:
@ -1960,6 +2020,7 @@ static int ibmvscsi_remove(struct vio_dev *vdev)
ibmvscsi_ops->release_crq_queue(&hostdata->queue, hostdata,
max_events);
kthread_stop(hostdata->work_thread);
srp_remove_host(hostdata->host);
scsi_remove_host(hostdata->host);
scsi_host_put(hostdata->host);

View File

@ -91,12 +91,16 @@ struct event_pool {
struct ibmvscsi_host_data {
atomic_t request_limit;
int client_migrated;
int reset_crq;
int reenable_crq;
struct device *dev;
struct event_pool pool;
struct crq_queue queue;
struct tasklet_struct srp_task;
struct list_head sent;
struct Scsi_Host *host;
struct task_struct *work_thread;
wait_queue_head_t work_wait_q;
struct mad_adapter_info_data madapter_info;
struct capabilities caps;
dma_addr_t caps_addr;

View File

@ -31,6 +31,7 @@
#include <asm/prom.h>
#include <asm/iommu.h>
#include <asm/hvcall.h>
#include <linux/delay.h>
#include <linux/dma-mapping.h>
#include <linux/gfp.h>
#include <linux/interrupt.h>
@ -71,11 +72,13 @@ static void rpavscsi_release_crq_queue(struct crq_queue *queue,
struct ibmvscsi_host_data *hostdata,
int max_requests)
{
long rc;
long rc = 0;
struct vio_dev *vdev = to_vio_dev(hostdata->dev);
free_irq(vdev->irq, (void *)hostdata);
tasklet_kill(&hostdata->srp_task);
do {
if (rc)
msleep(100);
rc = plpar_hcall_norets(H_FREE_CRQ, vdev->unit_address);
} while ((rc == H_BUSY) || (H_IS_LONG_BUSY(rc)));
dma_unmap_single(hostdata->dev,
@ -200,11 +203,13 @@ static void set_adapter_info(struct ibmvscsi_host_data *hostdata)
static int rpavscsi_reset_crq_queue(struct crq_queue *queue,
struct ibmvscsi_host_data *hostdata)
{
int rc;
int rc = 0;
struct vio_dev *vdev = to_vio_dev(hostdata->dev);
/* Close the CRQ */
do {
if (rc)
msleep(100);
rc = plpar_hcall_norets(H_FREE_CRQ, vdev->unit_address);
} while ((rc == H_BUSY) || (H_IS_LONG_BUSY(rc)));
@ -301,7 +306,10 @@ static int rpavscsi_init_crq_queue(struct crq_queue *queue,
req_irq_failed:
tasklet_kill(&hostdata->srp_task);
rc = 0;
do {
if (rc)
msleep(100);
rc = plpar_hcall_norets(H_FREE_CRQ, vdev->unit_address);
} while ((rc == H_BUSY) || (H_IS_LONG_BUSY(rc)));
reg_crq_failed:
@ -323,11 +331,13 @@ static int rpavscsi_init_crq_queue(struct crq_queue *queue,
static int rpavscsi_reenable_crq_queue(struct crq_queue *queue,
struct ibmvscsi_host_data *hostdata)
{
int rc;
int rc = 0;
struct vio_dev *vdev = to_vio_dev(hostdata->dev);
/* Re-enable the CRQ */
do {
if (rc)
msleep(100);
rc = plpar_hcall_norets(H_ENABLE_CRQ, vdev->unit_address);
} while ((rc == H_IN_PROGRESS) || (rc == H_BUSY) || (H_IS_LONG_BUSY(rc)));