mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@suse.de>
To: linux-kernel@vger.kernel.org
Cc: Pavan Savoy <pavan_savoy@ti.com>, Greg Kroah-Hartman <gregkh@suse.de>
Subject: [PATCH 4/8] drivers:misc: ti-st: protect registrations
Date: Mon,  9 Jan 2012 11:25:01 -0800	[thread overview]
Message-ID: <1326137105-9960-4-git-send-email-gregkh@suse.de> (raw)
In-Reply-To: <1326137105-9960-1-git-send-email-gregkh@suse.de>

From: Pavan Savoy <pavan_savoy@ti.com>

Concurrent access to UART2/combo-interface by multiple protocol drivers such
as BT, FM and GPS caused issues during firmware download failure cases or
cases when the firmware download took longer than usual.

This was because of un-safe access to protos_registered & st_states.
Protecting this will also make the registration complete callback un-safe for
sleep.

Signed-off-by: Pavan Savoy <pavan_savoy@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
---
 drivers/misc/ti-st/st_core.c |   18 ++++++++++++++----
 1 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/drivers/misc/ti-st/st_core.c b/drivers/misc/ti-st/st_core.c
index ba168a7..2b62232 100644
--- a/drivers/misc/ti-st/st_core.c
+++ b/drivers/misc/ti-st/st_core.c
@@ -137,6 +137,8 @@ void st_send_frame(unsigned char chnl_id, struct st_data_s *st_gdata)
  * st_reg_complete -
  * to call registration complete callbacks
  * of all protocol stack drivers
+ * This function is being called with spin lock held, protocol drivers are
+ * only expected to complete their waits and do nothing more than that.
  */
 void st_reg_complete(struct st_data_s *st_gdata, char err)
 {
@@ -538,11 +540,12 @@ long st_register(struct st_proto_s *new_proto)
 		set_bit(ST_REG_IN_PROGRESS, &st_gdata->st_state);
 		st_recv = st_kim_recv;
 
+		/* enable the ST LL - to set default chip state */
+		st_ll_enable(st_gdata);
+
 		/* release lock previously held - re-locked below */
 		spin_unlock_irqrestore(&st_gdata->lock, flags);
 
-		/* enable the ST LL - to set default chip state */
-		st_ll_enable(st_gdata);
 		/* this may take a while to complete
 		 * since it involves BT fw download
 		 */
@@ -553,10 +556,13 @@ long st_register(struct st_proto_s *new_proto)
 			    (test_bit(ST_REG_PENDING, &st_gdata->st_state))) {
 				pr_err(" KIM failure complete callback ");
 				st_reg_complete(st_gdata, err);
+				clear_bit(ST_REG_PENDING, &st_gdata->st_state);
 			}
 			return -EINVAL;
 		}
 
+		spin_lock_irqsave(&st_gdata->lock, flags);
+
 		clear_bit(ST_REG_IN_PROGRESS, &st_gdata->st_state);
 		st_recv = st_int_recv;
 
@@ -576,10 +582,10 @@ long st_register(struct st_proto_s *new_proto)
 		if (st_gdata->is_registered[new_proto->chnl_id] == true) {
 			pr_err(" proto %d already registered ",
 				   new_proto->chnl_id);
+			spin_unlock_irqrestore(&st_gdata->lock, flags);
 			return -EALREADY;
 		}
 
-		spin_lock_irqsave(&st_gdata->lock, flags);
 		add_channel_to_table(st_gdata, new_proto);
 		st_gdata->protos_registered++;
 		new_proto->write = st_write;
@@ -619,7 +625,7 @@ long st_unregister(struct st_proto_s *proto)
 
 	spin_lock_irqsave(&st_gdata->lock, flags);
 
-	if (st_gdata->list[proto->chnl_id] == NULL) {
+	if (st_gdata->is_registered[proto->chnl_id] == false) {
 		pr_err(" chnl_id %d not registered", proto->chnl_id);
 		spin_unlock_irqrestore(&st_gdata->lock, flags);
 		return -EPROTONOSUPPORT;
@@ -629,6 +635,10 @@ long st_unregister(struct st_proto_s *proto)
 	remove_channel_from_table(st_gdata, proto);
 	spin_unlock_irqrestore(&st_gdata->lock, flags);
 
+	/* paranoid check */
+	if (st_gdata->protos_registered < ST_EMPTY)
+		st_gdata->protos_registered = ST_EMPTY;
+
 	if ((st_gdata->protos_registered == ST_EMPTY) &&
 	    (!test_bit(ST_REG_PENDING, &st_gdata->st_state))) {
 		pr_info(" all chnl_ids unregistered ");
-- 
1.7.8


  parent reply	other threads:[~2012-01-09 19:25 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-01-09 19:07 [GIT PATCH] char and misc patches for 3.3 Greg KH
2012-01-09 19:24 ` [PATCH 1/8] misc: ad525x_dpot: Add support for SPI module device table matching Greg Kroah-Hartman
2012-01-09 19:24   ` [PATCH 2/8] s390: tape_class.h: remove kobj_map.h inclusion Greg Kroah-Hartman
2012-01-09 19:25   ` [PATCH 3/8] char_dev.c: fix up some whitespace errors Greg Kroah-Hartman
2012-01-09 19:25   ` Greg Kroah-Hartman [this message]
2012-01-09 19:25   ` [PATCH 5/8] drivers:misc: ti-st: flush UART upon fw failure Greg Kroah-Hartman
2012-01-09 19:25   ` [PATCH 6/8] drivers:misc: ti-st: DEBUG uart, baud rate mods Greg Kroah-Hartman
2012-01-09 19:25   ` [PATCH 7/8] BMP085: Remove redundant semi-colon from return statement Greg Kroah-Hartman
2012-01-09 19:25   ` [PATCH 8/8] isl29020: Remove a " Greg Kroah-Hartman

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=1326137105-9960-4-git-send-email-gregkh@suse.de \
    --to=gregkh@suse.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pavan_savoy@ti.com \
    /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

Powered by JetHome