From: Imre Deak Date: Mon, 5 Apr 2010 23:16:06 +0000 (+0300) Subject: gpu: pvr: remove unrelevant comments about caller ID X-Git-Url: http://git.openpandora.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=df979f2d74e49fc0df1a2075b80973c88f9b8782;p=sgx.git gpu: pvr: remove unrelevant comments about caller ID Signed-off-by: Imre Deak CC: Mark Underwood CC: Mark Riding Reviewed-by: Topi Pohjolainen Reviewed-by: Ville Syrjälä --- diff --git a/pvr/pvr_debug.c b/pvr/pvr_debug.c index 0a3707e..7cd60c1 100644 --- a/pvr/pvr_debug.c +++ b/pvr/pvr_debug.c @@ -295,14 +295,6 @@ static int pvr_dbg_reset(void *data, u64 val) goto exit; } - /* - * Yes, this is kinda braindead. KERNEL_ID, IMG_TRUE above means - * take the power lock - not just try lock - and keep it. TIMER_ID - * here means that we have already the power lock, so don't take it. - * Also - regardless of the ID - the following will release the lock. - * Finally we pass KERNEL_ID again to take and release the lock. - * Yay! - */ HWRecoveryResetSGX(node); SGXTestActivePowerEvent(node); diff --git a/pvr/sgxinit.c b/pvr/sgxinit.c index 301b94f..7b3982a 100644 --- a/pvr/sgxinit.c +++ b/pvr/sgxinit.c @@ -895,16 +895,6 @@ static void SGX_MISRHandler(void *pvData) int dev_idx; enum PVRSRV_ERROR err; - /* - * Yet again we have to cope with the caller ID argument, which - * determines whether the power lock is properly locked or just - * trylock'd. If trylock fails we just get back here with an error - * so we need a proper lock. Actually trylock should never fail - * either because of the next assumption. - * The code afterwards depends on noone turning off the power and - * also the power lock to be released, so ask for the lock to be - * released when returning from the function. - */ dev_idx = psDeviceNode->sDevId.ui32DeviceIndex; err = PVRSRVSetDevicePowerStateKM(dev_idx, PVRSRV_POWER_STATE_D0); BUG_ON(err != PVRSRV_OK);