linux-wireless.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Samuel Ortiz <sameo@linux.intel.com>
To: Johannes Berg <johannes@sipsolutions.net>
Cc: John Linville <linville@tuxdriver.com>,
	linux-wireless@vger.kernel.org, Zhu Yi <yi.zhu@intel.com>,
	Samuel Ortiz <samuel.ortiz@intel.com>
Subject: Re: [PATCH 01/12][2.6.31 and w-t] iwmc3200wifi: fix sdio initialisation race
Date: Mon, 15 Jun 2009 19:52:26 +0200	[thread overview]
Message-ID: <20090615175225.GF4094@sortiz.org> (raw)
In-Reply-To: <1245080647.23912.3.camel@johannes.local>

On Mon, Jun 15, 2009 at 05:44:07PM +0200, Johannes Berg wrote:
> On Mon, 2009-06-15 at 17:39 +0200, Samuel Ortiz wrote:
> > From: Samuel Ortiz <samuel.ortiz@intel.com>
> > 
> > When modprobing this driver, the netdev interface is ready before the SDIO
> > function is correctly set.
> > This can lead to userspace calling ndo_open() before the SDIO bus is ready.
> > We fix that by returning an error from ndo_open() when the
> > IWM_STATUS_BUS_READY bit is not set yet.
> 
> Can't you wait for it to be set instead?
I thought about it, but I didnt want to add one more completion/wait_queue
pointer to our already crowded iwm structures.
But if you or John insist on it, I can definitely do it.

Cheers,
Samuel.
 
> johannes
> 
> > Signed-off-by: Samuel Ortiz <samuel.ortiz@intel.com>
> > ---
> >  drivers/net/wireless/iwmc3200wifi/iwm.h    |   11 ++++++-----
> >  drivers/net/wireless/iwmc3200wifi/main.c   |    6 +++++-
> >  drivers/net/wireless/iwmc3200wifi/netdev.c |   10 ++++++----
> >  drivers/net/wireless/iwmc3200wifi/sdio.c   |    4 ++++
> >  4 files changed, 21 insertions(+), 10 deletions(-)
> > 
> > diff --git a/drivers/net/wireless/iwmc3200wifi/iwm.h b/drivers/net/wireless/iwmc3200wifi/iwm.h
> > index 635c16e..3cc4e91 100644
> > --- a/drivers/net/wireless/iwmc3200wifi/iwm.h
> > +++ b/drivers/net/wireless/iwmc3200wifi/iwm.h
> > @@ -180,11 +180,12 @@ struct iwm_key {
> >  
> >  #define IWM_SCAN_ID_MAX 0xff
> >  
> > -#define IWM_STATUS_READY		0
> > -#define IWM_STATUS_SCANNING		1
> > -#define IWM_STATUS_SCAN_ABORTING	2
> > -#define IWM_STATUS_ASSOCIATING		3
> > -#define IWM_STATUS_ASSOCIATED		4
> > +#define IWM_STATUS_BUS_READY		0
> > +#define IWM_STATUS_READY		1
> > +#define IWM_STATUS_SCANNING		2
> > +#define IWM_STATUS_SCAN_ABORTING	3
> > +#define IWM_STATUS_ASSOCIATING		4
> > +#define IWM_STATUS_ASSOCIATED		5
> >  
> >  #define IWM_RADIO_RFKILL_OFF		0
> >  #define IWM_RADIO_RFKILL_HW		1
> > diff --git a/drivers/net/wireless/iwmc3200wifi/main.c b/drivers/net/wireless/iwmc3200wifi/main.c
> > index 6a2640f..09ed247 100644
> > --- a/drivers/net/wireless/iwmc3200wifi/main.c
> > +++ b/drivers/net/wireless/iwmc3200wifi/main.c
> > @@ -227,11 +227,15 @@ int iwm_priv_init(struct iwm_priv *iwm)
> >  void iwm_reset(struct iwm_priv *iwm)
> >  {
> >  	struct iwm_notif *notif, *next;
> > +	unsigned long status = 0;
> >  
> >  	if (test_bit(IWM_STATUS_READY, &iwm->status))
> >  		iwm_target_reset(iwm);
> >  
> > -	iwm->status = 0;
> > +	if (test_bit(IWM_STATUS_BUS_READY, &iwm->status))
> > +		set_bit(IWM_STATUS_BUS_READY, &status);
> > +
> > +	iwm->status = status;
> >  	iwm->scan_id = 1;
> >  
> >  	list_for_each_entry_safe(notif, next, &iwm->pending_notif, pending) {
> > diff --git a/drivers/net/wireless/iwmc3200wifi/netdev.c b/drivers/net/wireless/iwmc3200wifi/netdev.c
> > index 68e2c3b..5b7aa34 100644
> > --- a/drivers/net/wireless/iwmc3200wifi/netdev.c
> > +++ b/drivers/net/wireless/iwmc3200wifi/netdev.c
> > @@ -54,9 +54,10 @@
> >  static int iwm_open(struct net_device *ndev)
> >  {
> >  	struct iwm_priv *iwm = ndev_to_iwm(ndev);
> > -	int ret = 0;
> > +	int ret = -ENODEV;
> >  
> > -	if (!test_bit(IWM_RADIO_RFKILL_SW, &iwm->radio))
> > +	if (!test_bit(IWM_RADIO_RFKILL_SW, &iwm->radio)
> > +	    && test_bit(IWM_STATUS_BUS_READY, &iwm->status))
> >  		ret = iwm_up(iwm);
> >  
> >  	return ret;
> > @@ -65,9 +66,10 @@ static int iwm_open(struct net_device *ndev)
> >  static int iwm_stop(struct net_device *ndev)
> >  {
> >  	struct iwm_priv *iwm = ndev_to_iwm(ndev);
> > -	int ret = 0;
> > +	int ret = -ENODEV;
> >  
> > -	if (!test_bit(IWM_RADIO_RFKILL_SW, &iwm->radio))
> > +	if (!test_bit(IWM_RADIO_RFKILL_SW, &iwm->radio)
> > +	    && test_bit(IWM_STATUS_BUS_READY, &iwm->status))
> >  		ret = iwm_down(iwm);
> >  
> >  	return ret;
> > diff --git a/drivers/net/wireless/iwmc3200wifi/sdio.c b/drivers/net/wireless/iwmc3200wifi/sdio.c
> > index b54da67..97e7da3 100644
> > --- a/drivers/net/wireless/iwmc3200wifi/sdio.c
> > +++ b/drivers/net/wireless/iwmc3200wifi/sdio.c
> > @@ -454,6 +454,9 @@ static int iwm_sdio_probe(struct sdio_func *func,
> >  
> >  	INIT_WORK(&hw->isr_worker, iwm_sdio_isr_worker);
> >  
> > +	/* The SDIO bus is set and ready to be enabled */
> > +	set_bit(IWM_STATUS_BUS_READY, &iwm->status);
> > +
> >  	dev_info(dev, "IWM SDIO probe\n");
> >  
> >  	return 0;
> > @@ -476,6 +479,7 @@ static void iwm_sdio_remove(struct sdio_func *func)
> >  	destroy_workqueue(hw->isr_wq);
> >  
> >  	sdio_set_drvdata(func, NULL);
> > +	clear_bit(IWM_STATUS_BUS_READY, &iwm->status);
> >  
> >  	dev_info(dev, "IWM SDIO remove\n");
> >  



-- 
Intel Open Source Technology Centre
http://oss.intel.com/

  reply	other threads:[~2009-06-15 17:50 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-06-15 15:39 [PATCH 00/12] iwmc3200wifi updates Samuel Ortiz
2009-06-15 15:39 ` [PATCH 01/12][2.6.31 and w-t] iwmc3200wifi: fix sdio initialisation race Samuel Ortiz
2009-06-15 15:44   ` Johannes Berg
2009-06-15 17:52     ` Samuel Ortiz [this message]
2009-06-15 15:39 ` [PATCH 02/12][2.6.31 and w-t] iwmc3200wifi: check for iwm_priv_init error Samuel Ortiz
2009-06-15 15:39 ` [PATCH 03/12][2.6.31 and w-t] iwmc3200wifi: add iwm_if_add and iwm_if_remove Samuel Ortiz
2009-06-15 15:39 ` [PATCH 04/12][2.6.31 and w-t] iwmc3200wifi: fix potential kernel oops on module removal Samuel Ortiz
2009-06-15 15:39 ` [PATCH 05/12][2.6.31 and w-t] iwmc3200wifi: add a mutex to protect iwm_reset_worker Samuel Ortiz
2009-06-15 15:39 ` [PATCH 06/12][w-t] iwmc3200wifi: invalidate keys when changing the BSSID Samuel Ortiz
2009-06-15 15:39 ` [PATCH 07/12][w-t] iwmc3200wifi: handling wifi_if_ntfy responses Samuel Ortiz
2009-06-15 15:39 ` [PATCH 08/12][w-t] iwmc3200wifi: cfg80211 key hooks implemetation Samuel Ortiz
2009-06-15 15:39 ` [PATCH 09/12][w-t] iwmc3200wifi: change coexist periodic calibration flag Samuel Ortiz
2009-06-15 15:39 ` [PATCH 10/12][w-t] iwmc3200wifi: cache keys when interface is down Samuel Ortiz
2009-06-15 15:39 ` [PATCH 11/12][w-t] iwmc3200wifi: remove select LIB80211 from Kconfig Samuel Ortiz
2009-06-15 15:39 ` [PATCH 12/12][w-t] iwmc3200wifi: rfkill cleanup Samuel Ortiz

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=20090615175225.GF4094@sortiz.org \
    --to=sameo@linux.intel.com \
    --cc=johannes@sipsolutions.net \
    --cc=linux-wireless@vger.kernel.org \
    --cc=linville@tuxdriver.com \
    --cc=samuel.ortiz@intel.com \
    --cc=yi.zhu@intel.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;
as well as URLs for NNTP newsgroup(s).