From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tony Houghton Date: Sun, 5 Jun 2011 14:11:30 +0100 Subject: [ath9k-devel] ath9k causes lockups since kernel 2.6.35 In-Reply-To: References: <20110310131950.61580121@toddler> <20110314162721.1c3e1429@toddler> <20110509164541.364bac2c@realh.co.uk> <20110604180847.45088f0d@realh.co.uk> Message-ID: <20110605141130.11995b77@toddler> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ath9k-devel@lists.ath9k.org On Sun, 5 Jun 2011 07:30:34 -0400 Camilo Mesias wrote: > I have a suspicion the lockup happens in the rf off code, because I > can cause the lockup reliably simply with iwconfig wlan0 tx off > Can other people confirm this? Is it worth putting more defensive > code in that function? This is what I found (I forgot to use Reply To All so I only sent this to Mohammed Shafi before): : The freeze is in ath9k_hw_set_power_awake during the for() loop when it : calls : REG_SET_BIT(ah, AR_RTC_FORCE_WAKE, AR_RTC_FORCE_WAKE_EN); That could be wrong because I used printk to track it down and there may be a race condition between the crash and the messages appearing on the console. I also discovered that val = REG_READ(ah, AR_RTC_STATUS) & AR_RTC_STATUS_M; had a result of 15, but the status should only have one bit set at a time I think? I also wondered why it was trying to go into a wake state when it was supposed to be shutting down.