[PATCH] IB/ipath: allow diags on any unit
authorBryan O'Sullivan <bos@pathscale.com>
Sat, 1 Jul 2006 11:35:52 +0000 (04:35 -0700)
committerLinus Torvalds <torvalds@g5.osdl.org>
Sat, 1 Jul 2006 16:55:58 +0000 (09:55 -0700)
There is no longer a /dev/ipath_diag file; instead, there's
/dev/ipath_diag0, 1, etc.

It's still not possible to have diags run on more than one unit at a time,
but that's easy to fix at some point.

Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
drivers/infiniband/hw/ipath/ipath_diag.c
drivers/infiniband/hw/ipath/ipath_driver.c
drivers/infiniband/hw/ipath/ipath_file_ops.c
drivers/infiniband/hw/ipath/ipath_kernel.h
drivers/infiniband/hw/ipath/ipath_layer.c

index 0fcd507..d7f17f2 100644 (file)
@@ -67,18 +67,20 @@ static struct file_operations diag_file_ops = {
        .release = ipath_diag_release
 };
 
-static struct cdev *diag_cdev;
-static struct class_device *diag_class_dev;
-
-int ipath_diag_init(void)
+int ipath_diag_add(struct ipath_devdata *dd)
 {
-       return ipath_cdev_init(IPATH_DIAG_MINOR, "ipath_diag",
-                              &diag_file_ops, &diag_cdev, &diag_class_dev);
+       char name[16];
+
+       snprintf(name, sizeof(name), "ipath_diag%d", dd->ipath_unit);
+
+       return ipath_cdev_init(IPATH_DIAG_MINOR_BASE + dd->ipath_unit, name,
+                              &diag_file_ops, &dd->diag_cdev,
+                              &dd->diag_class_dev);
 }
 
-void ipath_diag_cleanup(void)
+void ipath_diag_remove(struct ipath_devdata *dd)
 {
-       ipath_cdev_cleanup(&diag_cdev, &diag_class_dev);
+       ipath_cdev_cleanup(&dd->diag_cdev, &dd->diag_class_dev);
 }
 
 /**
@@ -102,8 +104,7 @@ static int ipath_read_umem64(struct ipath_devdata *dd, void __user *uaddr,
        int ret;
 
        /* not very efficient, but it works for now */
-       if (reg_addr < dd->ipath_kregbase ||
-           reg_end > dd->ipath_kregend) {
+       if (reg_addr < dd->ipath_kregbase || reg_end > dd->ipath_kregend) {
                ret = -EINVAL;
                goto bail;
        }
@@ -140,8 +141,7 @@ static int ipath_write_umem64(struct ipath_devdata *dd, void __iomem *caddr,
        int ret;
 
        /* not very efficient, but it works for now */
-       if (reg_addr < dd->ipath_kregbase ||
-           reg_end > dd->ipath_kregend) {
+       if (reg_addr < dd->ipath_kregbase || reg_end > dd->ipath_kregend) {
                ret = -EINVAL;
                goto bail;
        }
@@ -240,59 +240,45 @@ bail:
 
 static int ipath_diag_open(struct inode *in, struct file *fp)
 {
+       int unit = iminor(in) - IPATH_DIAG_MINOR_BASE;
        struct ipath_devdata *dd;
-       int unit = 0; /* XXX this is bogus */
-       unsigned long flags;
        int ret;
 
-       dd = ipath_lookup(unit);
-
        mutex_lock(&ipath_mutex);
-       spin_lock_irqsave(&ipath_devs_lock, flags);
 
        if (ipath_diag_inuse) {
                ret = -EBUSY;
                goto bail;
        }
 
-       list_for_each_entry(dd, &ipath_dev_list, ipath_list) {
-               /*
-                * we need at least one infinipath device to be present
-                * (don't use INITTED, because we want to be able to open
-                * even if device is in freeze mode, which cleared INITTED).
-                * There is a small amount of risk to this, which is why we
-                * also verify kregbase is set.
-                */
-
-               if (!(dd->ipath_flags & IPATH_PRESENT) ||
-                   !dd->ipath_kregbase)
-                       continue;
-
-               ipath_diag_inuse = 1;
-               diag_set_link = 0;
-               ret = 0;
+       dd = ipath_lookup(unit);
+
+       if (dd == NULL || !(dd->ipath_flags & IPATH_PRESENT) ||
+           !dd->ipath_kregbase) {
+               ret = -ENODEV;
                goto bail;
        }
 
-       ret = -ENODEV;
-
-bail:
-       spin_unlock_irqrestore(&ipath_devs_lock, flags);
+       fp->private_data = dd;
+       ipath_diag_inuse = 1;
+       diag_set_link = 0;
+       ret = 0;
 
        /* Only expose a way to reset the device if we
           make it into diag mode. */
-       if (ret == 0)
-               ipath_expose_reset(&dd->pcidev->dev);
+       ipath_expose_reset(&dd->pcidev->dev);
 
+bail:
        mutex_unlock(&ipath_mutex);
 
        return ret;
 }
 
-static int ipath_diag_release(struct inode *i, struct file *f)
+static int ipath_diag_release(struct inode *in, struct file *fp)
 {
        mutex_lock(&ipath_mutex);
        ipath_diag_inuse = 0;
+       fp->private_data = NULL;
        mutex_unlock(&ipath_mutex);
        return 0;
 }
@@ -300,17 +286,10 @@ static int ipath_diag_release(struct inode *i, struct file *f)
 static ssize_t ipath_diag_read(struct file *fp, char __user *data,
                               size_t count, loff_t *off)
 {
-       int unit = 0; /* XXX provide for reads on other units some day */
-       struct ipath_devdata *dd;
+       struct ipath_devdata *dd = fp->private_data;
        void __iomem *kreg_base;
        ssize_t ret;
 
-       dd = ipath_lookup(unit);
-       if (!dd) {
-               ret = -ENODEV;
-               goto bail;
-       }
-
        kreg_base = dd->ipath_kregbase;
 
        if (count == 0)
@@ -329,23 +308,16 @@ static ssize_t ipath_diag_read(struct file *fp, char __user *data,
                ret = count;
        }
 
-bail:
        return ret;
 }
 
 static ssize_t ipath_diag_write(struct file *fp, const char __user *data,
                                size_t count, loff_t *off)
 {
-       int unit = 0; /* XXX this is bogus */
-       struct ipath_devdata *dd;
+       struct ipath_devdata *dd = fp->private_data;
        void __iomem *kreg_base;
        ssize_t ret;
 
-       dd = ipath_lookup(unit);
-       if (!dd) {
-               ret = -ENODEV;
-               goto bail;
-       }
        kreg_base = dd->ipath_kregbase;
 
        if (count == 0)
@@ -364,6 +336,5 @@ static ssize_t ipath_diag_write(struct file *fp, const char __user *data,
                ret = count;
        }
 
-bail:
        return ret;
 }
index 05e00ba..0a93e22 100644 (file)
@@ -546,6 +546,7 @@ static int __devinit ipath_init_one(struct pci_dev *pdev,
        ipath_device_create_group(&pdev->dev, dd);
        ipathfs_add_device(dd);
        ipath_user_add(dd);
+       ipath_diag_add(dd);
        ipath_layer_add(dd);
 
        goto bail;
@@ -578,8 +579,9 @@ static void __devexit ipath_remove_one(struct pci_dev *pdev)
                return;
 
        dd = pci_get_drvdata(pdev);
-       ipath_layer_del(dd);
-       ipath_user_del(dd);
+       ipath_layer_remove(dd);
+       ipath_diag_remove(dd);
+       ipath_user_remove(dd);
        ipathfs_remove_device(dd);
        ipath_device_remove_group(&pdev->dev, dd);
        ipath_cdbg(VERBOSE, "Releasing pci memory regions, dd %p, "
index 8ff97e4..cd2d5cd 100644 (file)
@@ -1443,16 +1443,16 @@ done:
 
 static int ipath_open(struct inode *in, struct file *fp)
 {
-       int ret, minor;
+       int ret, user_minor;
 
        mutex_lock(&ipath_mutex);
 
-       minor = iminor(in);
+       user_minor = iminor(in) - IPATH_USER_MINOR_BASE;
        ipath_cdbg(VERBOSE, "open on dev %lx (minor %d)\n",
-                  (long)in->i_rdev, minor);
+                  (long)in->i_rdev, user_minor);
 
-       if (minor)
-               ret = find_free_port(minor - 1, fp);
+       if (user_minor)
+               ret = find_free_port(user_minor - 1, fp);
        else
                ret = find_best_unit(fp);
 
@@ -1860,19 +1860,12 @@ int ipath_user_add(struct ipath_devdata *dd)
                                      "error %d\n", -ret);
                        goto bail;
                }
-               ret = ipath_diag_init();
-               if (ret < 0) {
-                       ipath_dev_err(dd, "Unable to set up diag support: "
-                                     "error %d\n", -ret);
-                       goto bail_sma;
-               }
-
                ret = init_cdev(0, "ipath", &ipath_file_ops, &wildcard_cdev,
                                &wildcard_class_dev);
                if (ret < 0) {
                        ipath_dev_err(dd, "Could not create wildcard "
                                      "minor: error %d\n", -ret);
-                       goto bail_diag;
+                       goto bail_sma;
                }
 
                atomic_set(&user_setup, 1);
@@ -1881,31 +1874,28 @@ int ipath_user_add(struct ipath_devdata *dd)
        snprintf(name, sizeof(name), "ipath%d", dd->ipath_unit);
 
        ret = init_cdev(dd->ipath_unit + 1, name, &ipath_file_ops,
-                       &dd->cdev, &dd->class_dev);
+                       &dd->user_cdev, &dd->user_class_dev);
        if (ret < 0)
                ipath_dev_err(dd, "Could not create user minor %d, %s\n",
                              dd->ipath_unit + 1, name);
 
        goto bail;
 
-bail_diag:
-       ipath_diag_cleanup();
 bail_sma:
        user_cleanup();
 bail:
        return ret;
 }
 
-void ipath_user_del(struct ipath_devdata *dd)
+void ipath_user_remove(struct ipath_devdata *dd)
 {
-       cleanup_cdev(&dd->cdev, &dd->class_dev);
+       cleanup_cdev(&dd->user_cdev, &dd->user_class_dev);
 
        if (atomic_dec_return(&user_count) == 0) {
                if (atomic_read(&user_setup) == 0)
                        goto bail;
 
                cleanup_cdev(&wildcard_cdev, &wildcard_class_dev);
-               ipath_diag_cleanup();
                user_cleanup();
 
                atomic_set(&user_setup, 0);
index 57dea8c..412480d 100644 (file)
@@ -355,8 +355,10 @@ struct ipath_devdata {
        char *ipath_freezemsg;
        /* pci access data structure */
        struct pci_dev *pcidev;
-       struct cdev *cdev;
-       struct class_device *class_dev;
+       struct cdev *user_cdev;
+       struct cdev *diag_cdev;
+       struct class_device *user_class_dev;
+       struct class_device *diag_class_dev;
        /* timer used to prevent stats overflow, error throttling, etc. */
        struct timer_list ipath_stats_timer;
        /* check for stale messages in rcv queue */
@@ -538,7 +540,7 @@ extern int __ipath_verbs_piobufavail(struct ipath_devdata *);
 extern int __ipath_verbs_rcv(struct ipath_devdata *, void *, void *, u32);
 
 void ipath_layer_add(struct ipath_devdata *);
-void ipath_layer_del(struct ipath_devdata *);
+void ipath_layer_remove(struct ipath_devdata *);
 
 int ipath_init_chip(struct ipath_devdata *, int);
 int ipath_enable_wc(struct ipath_devdata *dd);
@@ -552,14 +554,14 @@ int ipath_cdev_init(int minor, char *name, struct file_operations *fops,
 void ipath_cdev_cleanup(struct cdev **cdevp,
                        struct class_device **class_devp);
 
-int ipath_diag_init(void);
-void ipath_diag_cleanup(void);
+int ipath_diag_add(struct ipath_devdata *);
+void ipath_diag_remove(struct ipath_devdata *);
 void ipath_diag_bringup_link(struct ipath_devdata *);
 
 extern wait_queue_head_t ipath_sma_state_wait;
 
 int ipath_user_add(struct ipath_devdata *dd);
-void ipath_user_del(struct ipath_devdata *dd);
+void ipath_user_remove(struct ipath_devdata *dd);
 
 struct sk_buff *ipath_alloc_skb(struct ipath_devdata *dd, gfp_t);
 
@@ -843,9 +845,10 @@ extern struct mutex ipath_mutex;
 
 #define IPATH_DRV_NAME         "ipath_core"
 #define IPATH_MAJOR            233
+#define IPATH_USER_MINOR_BASE  0
 #define IPATH_SMA_MINOR                128
-#define IPATH_DIAG_MINOR       129
-#define IPATH_NMINORS          130
+#define IPATH_DIAG_MINOR_BASE  129
+#define IPATH_NMINORS          255
 
 #define ipath_dev_err(dd,fmt,...) \
        do { \
index fd95b55..bb15b17 100644 (file)
@@ -404,7 +404,7 @@ void ipath_layer_add(struct ipath_devdata *dd)
        mutex_unlock(&ipath_layer_mutex);
 }
 
-void ipath_layer_del(struct ipath_devdata *dd)
+void ipath_layer_remove(struct ipath_devdata *dd)
 {
        mutex_lock(&ipath_layer_mutex);