linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: Benjamin Herrenschmidt <benh@ozlabs.org>
To: Paul Mackerras <paulus@samba.org>
Cc: linuxppc-dev@ozlabs.org
Subject: [PATCH 6/6] [POWERPC] Fix device-tree locking vs. interrupts
Date: Wed, 09 Apr 2008 17:21:36 +1000	[thread overview]
Message-ID: <20080409072233.45667DDF19@ozlabs.org> (raw)

Lockdep found out that we can occasionally take the device-tree
lock for reading from softirq time (from rtas_token called
by the rtas real time clock code called by the NTP code),
while we take it occasionally for writing without masking
interrupts. The combination of those two can thus deadlock.

While some of those cases of interrupt read lock could be fixed
(such as caching the RTAS tokens) I figured that taking the
lock for writing is so rare (device-tree modification) that we
may as well penalize that case and allow reading from interrupts.

Thus, this turns all the writers to take the lock with irqs
masked to avoid the situation.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/kernel/prom.c |   28 +++++++++++++++++-----------
 1 file changed, 17 insertions(+), 11 deletions(-)

--- linux-work.orig/arch/powerpc/kernel/prom.c	2008-04-09 14:23:18.000000000 +1000
+++ linux-work/arch/powerpc/kernel/prom.c	2008-04-09 14:25:48.000000000 +1000
@@ -1244,12 +1244,14 @@ EXPORT_SYMBOL(of_node_put);
  */
 void of_attach_node(struct device_node *np)
 {
-	write_lock(&devtree_lock);
+	unsigned long flags;
+
+	write_lock_irqsave(&devtree_lock, flags);
 	np->sibling = np->parent->child;
 	np->allnext = allnodes;
 	np->parent->child = np;
 	allnodes = np;
-	write_unlock(&devtree_lock);
+	write_unlock_irqrestore(&devtree_lock, flags);
 }
 
 /*
@@ -1260,8 +1262,9 @@ void of_attach_node(struct device_node *
 void of_detach_node(struct device_node *np)
 {
 	struct device_node *parent;
+	unsigned long flags;
 
-	write_lock(&devtree_lock);
+	write_lock_irqsave(&devtree_lock, flags);
 
 	parent = np->parent;
 	if (!parent)
@@ -1292,7 +1295,7 @@ void of_detach_node(struct device_node *
 	of_node_set_flag(np, OF_DETACHED);
 
 out_unlock:
-	write_unlock(&devtree_lock);
+	write_unlock_irqrestore(&devtree_lock, flags);
 }
 
 #ifdef CONFIG_PPC_PSERIES
@@ -1373,20 +1376,21 @@ __initcall(prom_reconfig_setup);
 int prom_add_property(struct device_node* np, struct property* prop)
 {
 	struct property **next;
+	unsigned long flags;
 
 	prop->next = NULL;	
-	write_lock(&devtree_lock);
+	write_lock_irqsave(&devtree_lock, flags);
 	next = &np->properties;
 	while (*next) {
 		if (strcmp(prop->name, (*next)->name) == 0) {
 			/* duplicate ! don't insert it */
-			write_unlock(&devtree_lock);
+			write_unlock_irqrestore(&devtree_lock, flags);
 			return -1;
 		}
 		next = &(*next)->next;
 	}
 	*next = prop;
-	write_unlock(&devtree_lock);
+	write_unlock_irqrestore(&devtree_lock, flags);
 
 #ifdef CONFIG_PROC_DEVICETREE
 	/* try to add to proc as well if it was initialized */
@@ -1406,9 +1410,10 @@ int prom_add_property(struct device_node
 int prom_remove_property(struct device_node *np, struct property *prop)
 {
 	struct property **next;
+	unsigned long flags;
 	int found = 0;
 
-	write_lock(&devtree_lock);
+	write_lock_irqsave(&devtree_lock, flags);
 	next = &np->properties;
 	while (*next) {
 		if (*next == prop) {
@@ -1421,7 +1426,7 @@ int prom_remove_property(struct device_n
 		}
 		next = &(*next)->next;
 	}
-	write_unlock(&devtree_lock);
+	write_unlock_irqrestore(&devtree_lock, flags);
 
 	if (!found)
 		return -ENODEV;
@@ -1447,9 +1452,10 @@ int prom_update_property(struct device_n
 			 struct property *oldprop)
 {
 	struct property **next;
+	unsigned long flags;
 	int found = 0;
 
-	write_lock(&devtree_lock);
+	write_lock_irqsave(&devtree_lock, flags);
 	next = &np->properties;
 	while (*next) {
 		if (*next == oldprop) {
@@ -1463,7 +1469,7 @@ int prom_update_property(struct device_n
 		}
 		next = &(*next)->next;
 	}
-	write_unlock(&devtree_lock);
+	write_unlock_irqrestore(&devtree_lock, flags);
 
 	if (!found)
 		return -ENODEV;

                 reply	other threads:[~2008-04-09  7:21 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=20080409072233.45667DDF19@ozlabs.org \
    --to=benh@ozlabs.org \
    --cc=linuxppc-dev@ozlabs.org \
    --cc=paulus@samba.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;
as well as URLs for NNTP newsgroup(s).