Linux-ARM-Kernel Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: y@google.com (y at google.com)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH] omap: resource: Add per-resource mutex for OMAP resource framework
Date: Tue,  9 Mar 2010 16:09:12 -0800	[thread overview]
Message-ID: <1268179753-4063-1-git-send-email-y> (raw)

From: Chunqiu Wang <cqwang@motorola.com>

Current OMAP resource fwk uses a global res_mutex
for resource_request and resource_release calls
for all the available resources.It may cause dead
lock if resource_request/resource_release is called
recursively.

For current OMAP3 VDD1/VDD2 resource, the change_level
implementation is mach-omap2/resource34xx.c/set_opp(),
when using resource_release to remove vdd1 constraint,
this function may call resource_release again to release
Vdd2 constrait if target vdd1 level is less than OPP3.
in this scenario, the global res_mutex down operation
will be called again, this will cause the second
down operation hang there.

To fix the problem, per-resource mutex is added
to avoid hangup when resource_request/resource_release
is called recursively.

Signed-off-by: Chunqiu Wang <cqwang@motorola.com>
Signed-off-by: Mike Chan <mike@android.com>
---
 arch/arm/plat-omap/include/plat/resource.h |    2 ++
 arch/arm/plat-omap/resource.c              |   21 ++++++++++-----------
 2 files changed, 12 insertions(+), 11 deletions(-)

diff --git a/arch/arm/plat-omap/include/plat/resource.h b/arch/arm/plat-omap/include/plat/resource.h
index 9acebcc..b5aff1f 100644
--- a/arch/arm/plat-omap/include/plat/resource.h
+++ b/arch/arm/plat-omap/include/plat/resource.h
@@ -54,6 +54,8 @@ struct shared_resource {
 	/* Shared resource operations */
 	struct shared_resource_ops *ops;
 	struct list_head node;
+	/* Protect each resource */
+	struct mutex resource_mutex;
 };
 
 struct shared_resource_ops {
diff --git a/arch/arm/plat-omap/resource.c b/arch/arm/plat-omap/resource.c
index f1cdecf..0a7b79b 100644
--- a/arch/arm/plat-omap/resource.c
+++ b/arch/arm/plat-omap/resource.c
@@ -277,6 +277,7 @@ int resource_register(struct shared_resource *resp)
 	}
 
 	INIT_LIST_HEAD(&resp->users_list);
+	mutex_init(&resp->resource_mutex);
 
 	/* Add the resource to the resource list */
 	list_add(&resp->node, &res_list);
@@ -339,14 +340,13 @@ int resource_request(const char *name, struct device *dev,
 	struct  users_list *user;
 	int 	found = 0, ret = 0;
 
-	down(&res_mutex);
-	resp = _resource_lookup(name);
+	resp = resource_lookup(name);
 	if (!resp) {
 		printk(KERN_ERR "resource_request: Invalid resource name\n");
-		ret = -EINVAL;
-		goto res_unlock;
+		return -EINVAL;
 	}
 
+	mutex_lock(&resp->resource_mutex);
 	/* Call the resource specific validate function */
 	if (resp->ops->validate_level) {
 		ret = resp->ops->validate_level(resp, level);
@@ -375,7 +375,7 @@ int resource_request(const char *name, struct device *dev,
 	user->level = level;
 
 res_unlock:
-	up(&res_mutex);
+	mutex_unlock(&resp->resource_mutex);
 	/*
 	 * Recompute and set the current level for the resource.
 	 * NOTE: update_resource level moved out of spin_lock, as it may call
@@ -406,14 +406,13 @@ int resource_release(const char *name, struct device *dev)
 	struct users_list *user;
 	int found = 0, ret = 0;
 
-	down(&res_mutex);
-	resp = _resource_lookup(name);
+	resp = resource_lookup(name);
 	if (!resp) {
 		printk(KERN_ERR "resource_release: Invalid resource name\n");
-		ret = -EINVAL;
-		goto res_unlock;
+		return -EINVAL;
 	}
 
+	mutex_lock(&resp->resource_mutex);
 	list_for_each_entry(user, &resp->users_list, node) {
 		if (user->dev == dev) {
 			found = 1;
@@ -434,7 +433,7 @@ int resource_release(const char *name, struct device *dev)
 	/* Recompute and set the current level for the resource */
 	ret = update_resource_level(resp);
 res_unlock:
-	up(&res_mutex);
+	mutex_unlock(&resp->resource_mutex);
 	return ret;
 }
 EXPORT_SYMBOL(resource_release);
@@ -458,7 +457,7 @@ int resource_get_level(const char *name)
 		up(&res_mutex);
 		return -EINVAL;
 	}
-	ret = resp->curr_level;
+	ret =  resp->curr_level;
 	up(&res_mutex);
 	return ret;
 }
-- 
1.6.6.2

             reply	other threads:[~2010-03-10  0:09 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-03-10  0:09 y at google.com [this message]
     [not found] <-1645286213079228429@unknownmsgid>
2010-03-10  0:11 ` [PATCH] omap: resource: Add per-resource mutex for OMAP resource framework Mike Chan
2010-03-10 17:27   ` Kevin Hilman

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1268179753-4063-1-git-send-email-y \
    --to=y@google.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox