From 6f6ef82cc9de24153ba7d5cedab5970e276aefa1 Mon Sep 17 00:00:00 2001 From: Carlos Corbacho Date: Sat, 26 Dec 2009 19:24:31 +0000 Subject: acer-wmi: Respect current backlight level when loading Set the backlight to use the current brightness when loaded, rather than always resetting the backlight to maximum brightness. Fixes kernel bugzilla #14207 Signed-off-by: Carlos Corbacho Reported-by: Denis Mukhin Signed-off-by: Len Brown --- drivers/platform/x86/acer-wmi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/platform/x86/acer-wmi.c b/drivers/platform/x86/acer-wmi.c index 07d14dfdf0b4..226b3e93498c 100644 --- a/drivers/platform/x86/acer-wmi.c +++ b/drivers/platform/x86/acer-wmi.c @@ -934,7 +934,7 @@ static int __devinit acer_backlight_init(struct device *dev) acer_backlight_device = bd; bd->props.power = FB_BLANK_UNBLANK; - bd->props.brightness = max_brightness; + bd->props.brightness = read_brightness(bd); bd->props.max_brightness = max_brightness; backlight_update_status(bd); return 0; -- cgit v1.2.3 From 455c0d71d46e86b0b7ff2c9dcfc19bc162302ee9 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Thu, 18 Feb 2010 10:28:20 -0800 Subject: ACPI: Fix regression where _PPC is not read at boot even when ignore_ppc=0 Earlier, Ingo Molnar posted a patch to make it so that the kernel would avoid reading _PPC on his broken T60. Unfortunately, it seems that with Thomas Renninger's patch last July to eliminate _PPC evaluations when the processor driver loads, the kernel never actually reads _PPC at all! This is problematic if you happen to boot your non-T60 computer in a state where the BIOS _wants_ _PPC to be something other than zero. So, put the _PPC evaluation back into acpi_processor_get_performance_info if ignore_ppc isn't 1. Signed-off-by: Darrick J. Wong Signed-off-by: Len Brown --- drivers/acpi/processor_perflib.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/acpi/processor_perflib.c b/drivers/acpi/processor_perflib.c index 2cabadcc4d8c..a959f6a07508 100644 --- a/drivers/acpi/processor_perflib.c +++ b/drivers/acpi/processor_perflib.c @@ -413,7 +413,11 @@ static int acpi_processor_get_performance_info(struct acpi_processor *pr) if (result) goto update_bios; - return 0; + /* We need to call _PPC once when cpufreq starts */ + if (ignore_ppc != 1) + result = acpi_processor_get_platform_limit(pr); + + return result; /* * Having _PPC but missing frequencies (_PSS, _PCT) is a very good hint that -- cgit v1.2.3 From 5e2f75b8993a0d83d469388b50716dd5551f2eb4 Mon Sep 17 00:00:00 2001 From: Dan Halperin Date: Thu, 18 Feb 2010 22:01:39 -0800 Subject: iwlwifi: set HT flags after channel in rxon The HT extension channel settings require priv->staging_rxon.channel to be accurate. However, iwl_set_rxon_ht was being called before iwl_set_rxon_channel and thus HT40 could be broken unless another call to iwl_mac_config came in. This problem was recently introduced by "iwlwifi: Fix to set correct ht configuration" The particular setting in which I noticed this was monitor mode: iwconfig wlan0 mode monitor ifconfig wlan0 up ./iw wlan0 set channel 64 HT40- #./iw wlan0 set channel 64 HT40- tcpdump -i wlan0 -y IEEE802_11_RADIO would only catch HT40 packets if I issued the IW command twice. From visual inspection, iwl_set_rxon_channel does not depend on iwl_set_rxon_ht, so simply swapping them should be safe and fixes this problem. Signed-off-by: Daniel Halperin Acked-by: Wey-Yi Guy Signed-off-by: Reinette Chatre CC: stable@kernel.org Signed-off-by: John W. Linville --- drivers/net/wireless/iwlwifi/iwl-core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/net/wireless/iwlwifi/iwl-core.c b/drivers/net/wireless/iwlwifi/iwl-core.c index d10bea64fce3..f36f804804fc 100644 --- a/drivers/net/wireless/iwlwifi/iwl-core.c +++ b/drivers/net/wireless/iwlwifi/iwl-core.c @@ -2744,8 +2744,8 @@ int iwl_mac_config(struct ieee80211_hw *hw, u32 changed) if ((le16_to_cpu(priv->staging_rxon.channel) != ch)) priv->staging_rxon.flags = 0; - iwl_set_rxon_ht(priv, ht_conf); iwl_set_rxon_channel(priv, conf->channel); + iwl_set_rxon_ht(priv, ht_conf); iwl_set_flags_for_band(priv, conf->channel->band); spin_unlock_irqrestore(&priv->lock, flags); -- cgit v1.2.3 From d9c4f846997c6d37e4f56907d93f1be022c17c6b Mon Sep 17 00:00:00 2001 From: Samu Onkalo Date: Fri, 19 Feb 2010 23:17:58 -0800 Subject: Input: polldev can cause crash in case when polling disabled When polled input device is opened and closed and there are no other users of polled device, the workqueue is created and destroyed in every open / close operation. It is probable that at some point dynamic allocation of internal parts of the workqueue cause changes to the workqueue. When a work is queued to the workqueue the work struct contains pointers to the workqueue data. If the workqueue has been changed and the work has never been queued to the new workqueue, work-struct contains pointers to the non-existing workqueue. This will cause crash at the work cancellation during device close since cancellation of a work assumes that the workqueue exists. To prevent that, work struct is cleaned up at device close. This keeps work struct clean for the next use. Signed-off-by: Samu Onkalo Signed-off-by: Dmitry Torokhov --- drivers/input/input-polldev.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'drivers') diff --git a/drivers/input/input-polldev.c b/drivers/input/input-polldev.c index aa6713b4a988..291d9393d359 100644 --- a/drivers/input/input-polldev.c +++ b/drivers/input/input-polldev.c @@ -100,6 +100,12 @@ static void input_close_polled_device(struct input_dev *input) struct input_polled_dev *dev = input_get_drvdata(input); cancel_delayed_work_sync(&dev->work); + /* + * Clean up work struct to remove references to the workqueue. + * It may be destroyed by the next call. This causes problems + * at next device open-close in case of poll_interval == 0. + */ + INIT_DELAYED_WORK(&dev->work, dev->work.work.func); input_polldev_stop_workqueue(); if (dev->close) -- cgit v1.2.3 From 3dae93ec3ee1fceec69f40ef9b97892ce62ba7a5 Mon Sep 17 00:00:00 2001 From: Mike Frysinger Date: Mon, 15 Feb 2010 19:32:25 -0500 Subject: [WATCHDOG] bfin: fix max timeout calculation Relying on overflow/wrap around isn't exact because if you wrap far enough, you get back to "valid" values. Reported-by: Thorsten Pohlmann Signed-off-by: Mike Frysinger Signed-off-by: Wim Van Sebroeck Cc: stable --- drivers/watchdog/bfin_wdt.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) (limited to 'drivers') diff --git a/drivers/watchdog/bfin_wdt.c b/drivers/watchdog/bfin_wdt.c index c7b3f9df2317..2159e668751c 100644 --- a/drivers/watchdog/bfin_wdt.c +++ b/drivers/watchdog/bfin_wdt.c @@ -1,9 +1,8 @@ /* * Blackfin On-Chip Watchdog Driver - * Supports BF53[123]/BF53[467]/BF54[2489]/BF561 * * Originally based on softdog.c - * Copyright 2006-2007 Analog Devices Inc. + * Copyright 2006-2010 Analog Devices Inc. * Copyright 2006-2007 Michele d'Amico * Copyright 1996 Alan Cox * @@ -137,13 +136,15 @@ static int bfin_wdt_running(void) */ static int bfin_wdt_set_timeout(unsigned long t) { - u32 cnt; + u32 cnt, max_t, sclk; unsigned long flags; - stampit(); + sclk = get_sclk(); + max_t = -1 / sclk; + cnt = t * sclk; + stamp("maxtimeout=%us newtimeout=%lus (cnt=%#x)", max_t, t, cnt); - cnt = t * get_sclk(); - if (cnt < get_sclk()) { + if (t > max_t) { printk(KERN_WARNING PFX "timeout value is too large\n"); return -EINVAL; } -- cgit v1.2.3 From eb083ba260f21ad79e83e1ad05a0d27e93b58c83 Mon Sep 17 00:00:00 2001 From: Roy Yin Date: Sun, 21 Feb 2010 22:52:49 -0800 Subject: Input: usbtouchscreen - extend coordinate range for Generaltouch devices Generaltouch protocol allows for coordinates in [0, 0xffff] range and there are devices reporting coordinates as high as 0x7fff so let's update the driver to reflect that. Signed-off-by: Roy Yin Signed-off-by: Dmitry Torokhov --- drivers/input/touchscreen/usbtouchscreen.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'drivers') diff --git a/drivers/input/touchscreen/usbtouchscreen.c b/drivers/input/touchscreen/usbtouchscreen.c index 09a5e7341bd5..5256123a5228 100644 --- a/drivers/input/touchscreen/usbtouchscreen.c +++ b/drivers/input/touchscreen/usbtouchscreen.c @@ -618,8 +618,8 @@ static int idealtek_read_data(struct usbtouch_usb *dev, unsigned char *pkt) #ifdef CONFIG_TOUCHSCREEN_USB_GENERAL_TOUCH static int general_touch_read_data(struct usbtouch_usb *dev, unsigned char *pkt) { - dev->x = ((pkt[2] & 0x0F) << 8) | pkt[1] ; - dev->y = ((pkt[4] & 0x0F) << 8) | pkt[3] ; + dev->x = (pkt[2] << 8) | pkt[1]; + dev->y = (pkt[4] << 8) | pkt[3]; dev->press = pkt[5] & 0xff; dev->touch = pkt[0] & 0x01; @@ -809,9 +809,9 @@ static struct usbtouch_device_info usbtouch_dev_info[] = { #ifdef CONFIG_TOUCHSCREEN_USB_GENERAL_TOUCH [DEVTYPE_GENERAL_TOUCH] = { .min_xc = 0x0, - .max_xc = 0x0500, + .max_xc = 0x7fff, .min_yc = 0x0, - .max_yc = 0x0500, + .max_yc = 0x7fff, .rept_size = 7, .read_data = general_touch_read_data, }, -- cgit v1.2.3 From a239a8b47cc0e5e6d7416a89f340beac06d5edaa Mon Sep 17 00:00:00 2001 From: Wey-Yi Guy Date: Fri, 19 Feb 2010 15:47:32 -0800 Subject: iwlwifi: error checking for number of tfds in queue When receive reply_tx and ready to decrement the count for number of tfds in queue, do error checking to prevent error condition and tfds_in_queue become negative number. Signed-off-by: Wey-Yi Guy Signed-off-by: Reinette Chatre CC: stable@kernel.org Signed-off-by: John W. Linville --- drivers/net/wireless/iwlwifi/iwl-4965.c | 2 +- drivers/net/wireless/iwlwifi/iwl-5000.c | 4 ++-- drivers/net/wireless/iwlwifi/iwl-core.h | 2 ++ drivers/net/wireless/iwlwifi/iwl-tx.c | 16 +++++++++++++++- 4 files changed, 20 insertions(+), 4 deletions(-) (limited to 'drivers') diff --git a/drivers/net/wireless/iwlwifi/iwl-4965.c b/drivers/net/wireless/iwlwifi/iwl-4965.c index 9b4b8b5c7574..31462813bac0 100644 --- a/drivers/net/wireless/iwlwifi/iwl-4965.c +++ b/drivers/net/wireless/iwlwifi/iwl-4965.c @@ -2008,7 +2008,7 @@ static void iwl4965_rx_reply_tx(struct iwl_priv *priv, IWL_DEBUG_TX_REPLY(priv, "Retry scheduler reclaim scd_ssn " "%d index %d\n", scd_ssn , index); freed = iwl_tx_queue_reclaim(priv, txq_id, index); - priv->stations[sta_id].tid[tid].tfds_in_queue -= freed; + iwl_free_tfds_in_queue(priv, sta_id, tid, freed); if (priv->mac80211_registered && (iwl_queue_space(&txq->q) > txq->q.low_mark) && diff --git a/drivers/net/wireless/iwlwifi/iwl-5000.c b/drivers/net/wireless/iwlwifi/iwl-5000.c index de45f308b744..f27c514886a5 100644 --- a/drivers/net/wireless/iwlwifi/iwl-5000.c +++ b/drivers/net/wireless/iwlwifi/iwl-5000.c @@ -1125,7 +1125,7 @@ static void iwl5000_rx_reply_tx(struct iwl_priv *priv, scd_ssn , index, txq_id, txq->swq_id); freed = iwl_tx_queue_reclaim(priv, txq_id, index); - priv->stations[sta_id].tid[tid].tfds_in_queue -= freed; + iwl_free_tfds_in_queue(priv, sta_id, tid, freed); if (priv->mac80211_registered && (iwl_queue_space(&txq->q) > txq->q.low_mark) && @@ -1154,7 +1154,7 @@ static void iwl5000_rx_reply_tx(struct iwl_priv *priv, freed = iwl_tx_queue_reclaim(priv, txq_id, index); if (ieee80211_is_data_qos(tx_resp->frame_ctrl)) - priv->stations[sta_id].tid[tid].tfds_in_queue -= freed; + iwl_free_tfds_in_queue(priv, sta_id, tid, freed); if (priv->mac80211_registered && (iwl_queue_space(&txq->q) > txq->q.low_mark)) diff --git a/drivers/net/wireless/iwlwifi/iwl-core.h b/drivers/net/wireless/iwlwifi/iwl-core.h index 675b7df632fc..fe37d6a6bf97 100644 --- a/drivers/net/wireless/iwlwifi/iwl-core.h +++ b/drivers/net/wireless/iwlwifi/iwl-core.h @@ -446,6 +446,8 @@ void iwl_hw_txq_ctx_free(struct iwl_priv *priv); int iwl_hw_tx_queue_init(struct iwl_priv *priv, struct iwl_tx_queue *txq); int iwl_txq_update_write_ptr(struct iwl_priv *priv, struct iwl_tx_queue *txq); +void iwl_free_tfds_in_queue(struct iwl_priv *priv, + int sta_id, int tid, int freed); int iwl_tx_queue_init(struct iwl_priv *priv, struct iwl_tx_queue *txq, int slots_num, u32 txq_id); void iwl_tx_queue_free(struct iwl_priv *priv, int txq_id); diff --git a/drivers/net/wireless/iwlwifi/iwl-tx.c b/drivers/net/wireless/iwlwifi/iwl-tx.c index 87ce2bd292c7..72136c8f51da 100644 --- a/drivers/net/wireless/iwlwifi/iwl-tx.c +++ b/drivers/net/wireless/iwlwifi/iwl-tx.c @@ -120,6 +120,20 @@ int iwl_txq_update_write_ptr(struct iwl_priv *priv, struct iwl_tx_queue *txq) EXPORT_SYMBOL(iwl_txq_update_write_ptr); +void iwl_free_tfds_in_queue(struct iwl_priv *priv, + int sta_id, int tid, int freed) +{ + if (priv->stations[sta_id].tid[tid].tfds_in_queue >= freed) + priv->stations[sta_id].tid[tid].tfds_in_queue -= freed; + else { + IWL_ERR(priv, "free more than tfds_in_queue (%u:%d)\n", + priv->stations[sta_id].tid[tid].tfds_in_queue, + freed); + priv->stations[sta_id].tid[tid].tfds_in_queue = 0; + } +} +EXPORT_SYMBOL(iwl_free_tfds_in_queue); + /** * iwl_tx_queue_free - Deallocate DMA queue. * @txq: Transmit queue to deallocate. @@ -1559,7 +1573,7 @@ void iwl_rx_reply_compressed_ba(struct iwl_priv *priv, if (txq->q.read_ptr != (ba_resp_scd_ssn & 0xff)) { /* calculate mac80211 ampdu sw queue to wake */ int freed = iwl_tx_queue_reclaim(priv, scd_flow, index); - priv->stations[sta_id].tid[tid].tfds_in_queue -= freed; + iwl_free_tfds_in_queue(priv, sta_id, tid, freed); if ((iwl_queue_space(&txq->q) > txq->q.low_mark) && priv->mac80211_registered && -- cgit v1.2.3 From a120e912eb51e347f36c71b60a1d13af74d30e83 Mon Sep 17 00:00:00 2001 From: Stanislaw Gruszka Date: Fri, 19 Feb 2010 15:47:33 -0800 Subject: iwlwifi: sanity check before counting number of tfds can be free Check the frame control for ieee80211_is_data_qos() is true before counting the number of tfds can be free, the tfds_in_queue only increment when ieee80211_is_data_qos() is true before transmit; so it should only decrement if the type match. Remove ieee80211_is_data_qos check for frame_ctrl in tx_resp to avoid invalid information pass from uCode. Signed-off-by: Stanislaw Gruszka Signed-off-by: Wey-Yi Guy Signed-off-by: Reinette Chatre CC: stable@kernel.org Signed-off-by: John W. Linville --- drivers/net/wireless/iwlwifi/iwl-5000.c | 6 ++---- drivers/net/wireless/iwlwifi/iwl-tx.c | 6 +++++- 2 files changed, 7 insertions(+), 5 deletions(-) (limited to 'drivers') diff --git a/drivers/net/wireless/iwlwifi/iwl-5000.c b/drivers/net/wireless/iwlwifi/iwl-5000.c index f27c514886a5..cffaae772d51 100644 --- a/drivers/net/wireless/iwlwifi/iwl-5000.c +++ b/drivers/net/wireless/iwlwifi/iwl-5000.c @@ -1153,16 +1153,14 @@ static void iwl5000_rx_reply_tx(struct iwl_priv *priv, tx_resp->failure_frame); freed = iwl_tx_queue_reclaim(priv, txq_id, index); - if (ieee80211_is_data_qos(tx_resp->frame_ctrl)) - iwl_free_tfds_in_queue(priv, sta_id, tid, freed); + iwl_free_tfds_in_queue(priv, sta_id, tid, freed); if (priv->mac80211_registered && (iwl_queue_space(&txq->q) > txq->q.low_mark)) iwl_wake_queue(priv, txq_id); } - if (ieee80211_is_data_qos(tx_resp->frame_ctrl)) - iwl_txq_check_empty(priv, sta_id, tid, txq_id); + iwl_txq_check_empty(priv, sta_id, tid, txq_id); if (iwl_check_bits(status, TX_ABORT_REQUIRED_MSK)) IWL_ERR(priv, "TODO: Implement Tx ABORT REQUIRED!!!\n"); diff --git a/drivers/net/wireless/iwlwifi/iwl-tx.c b/drivers/net/wireless/iwlwifi/iwl-tx.c index 72136c8f51da..8f4071562857 100644 --- a/drivers/net/wireless/iwlwifi/iwl-tx.c +++ b/drivers/net/wireless/iwlwifi/iwl-tx.c @@ -1145,6 +1145,7 @@ int iwl_tx_queue_reclaim(struct iwl_priv *priv, int txq_id, int index) struct iwl_queue *q = &txq->q; struct iwl_tx_info *tx_info; int nfreed = 0; + struct ieee80211_hdr *hdr; if ((index >= q->n_bd) || (iwl_queue_used(q, index) == 0)) { IWL_ERR(priv, "Read index for DMA queue txq id (%d), index %d, " @@ -1159,13 +1160,16 @@ int iwl_tx_queue_reclaim(struct iwl_priv *priv, int txq_id, int index) tx_info = &txq->txb[txq->q.read_ptr]; iwl_tx_status(priv, tx_info->skb[0]); + + hdr = (struct ieee80211_hdr *)tx_info->skb[0]->data; + if (hdr && ieee80211_is_data_qos(hdr->frame_control)) + nfreed++; tx_info->skb[0] = NULL; if (priv->cfg->ops->lib->txq_inval_byte_cnt_tbl) priv->cfg->ops->lib->txq_inval_byte_cnt_tbl(priv, txq); priv->cfg->ops->lib->txq_free_tfd(priv, txq); - nfreed++; } return nfreed; } -- cgit v1.2.3 From 61caf87cb5c2a198966018343a6ce4c5ab6cf8df Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Thu, 18 Feb 2010 23:06:27 +0100 Subject: i915 / PM: Fix hibernate regression caused by suspend/resume splitting Commit 84b79f8d2882b0a84330c04839ed4d3cefd2ff77 (drm/i915: Fix crash while aborting hibernation) attempted to fix a regression introduced by commit cbda12d77ea590082edb6d30bd342a67ebc459e0 (drm/i915: implement new pm ops for i915), but it went too far trying to split the freeze/suspend and resume/thaw parts of the code. As a result, it introduced another regression, which only is visible on some systems. Fix the problem by merging i915_drm_suspend() with i915_drm_freeze() and moving some code from i915_resume() into i915_drm_thaw(), so that intel_opregion_free() and intel_opregion_init() are also executed in the freeze and thaw code paths, respectively. Signed-off-by: Rafael J. Wysocki Reported-and-tested-by: Pedro Ribeiro Tested-by: Tino Keitel Acked-by: Eric Anholt Signed-off-by: Linus Torvalds --- drivers/gpu/drm/i915/i915_drv.c | 30 +++++++++--------------------- 1 file changed, 9 insertions(+), 21 deletions(-) (limited to 'drivers') diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 79beffcf5936..cf4cb3e9a0c2 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -176,6 +176,8 @@ MODULE_DEVICE_TABLE(pci, pciidlist); static int i915_drm_freeze(struct drm_device *dev) { + struct drm_i915_private *dev_priv = dev->dev_private; + pci_save_state(dev->pdev); /* If KMS is active, we do the leavevt stuff here */ @@ -191,17 +193,12 @@ static int i915_drm_freeze(struct drm_device *dev) i915_save_state(dev); - return 0; -} - -static void i915_drm_suspend(struct drm_device *dev) -{ - struct drm_i915_private *dev_priv = dev->dev_private; - intel_opregion_free(dev, 1); /* Modeset on resume, not lid events */ dev_priv->modeset_on_lid = 0; + + return 0; } static int i915_suspend(struct drm_device *dev, pm_message_t state) @@ -221,8 +218,6 @@ static int i915_suspend(struct drm_device *dev, pm_message_t state) if (error) return error; - i915_drm_suspend(dev); - if (state.event == PM_EVENT_SUSPEND) { /* Shut down the device */ pci_disable_device(dev->pdev); @@ -237,6 +232,10 @@ static int i915_drm_thaw(struct drm_device *dev) struct drm_i915_private *dev_priv = dev->dev_private; int error = 0; + i915_restore_state(dev); + + intel_opregion_init(dev, 1); + /* KMS EnterVT equivalent */ if (drm_core_check_feature(dev, DRIVER_MODESET)) { mutex_lock(&dev->struct_mutex); @@ -263,10 +262,6 @@ static int i915_resume(struct drm_device *dev) pci_set_master(dev->pdev); - i915_restore_state(dev); - - intel_opregion_init(dev, 1); - return i915_drm_thaw(dev); } @@ -423,8 +418,6 @@ static int i915_pm_suspend(struct device *dev) if (error) return error; - i915_drm_suspend(drm_dev); - pci_disable_device(pdev); pci_set_power_state(pdev, PCI_D3hot); @@ -464,13 +457,8 @@ static int i915_pm_poweroff(struct device *dev) { struct pci_dev *pdev = to_pci_dev(dev); struct drm_device *drm_dev = pci_get_drvdata(pdev); - int error; - - error = i915_drm_freeze(drm_dev); - if (!error) - i915_drm_suspend(drm_dev); - return error; + return i915_drm_freeze(drm_dev); } const struct dev_pm_ops i915_pm_ops = { -- cgit v1.2.3 From d306ebc28649b89877a22158fe0076f06cc46f60 Mon Sep 17 00:00:00 2001 From: "Pallipadi, Venkatesh" Date: Wed, 10 Feb 2010 10:35:31 -0800 Subject: ACPI: Be in TS_POLLING state during mwait based C-state entry ACPI deep C-state entry had a long standing bug/missing feature, wherein we were sending resched IPIs when an idle CPU is in mwait based deep C-state. Only mwait based C1 was using the write to the monitored address to wake up mwait'ing CPU. This patch changes the code to retain TS_POLLING bit if we are entering an mwait based deep C-state. The patch has been verified to reduce the number of resched IPIs in general and also improves the performance/power on workloads with low system utilization (i.e., when mwait based deep C-states are being used). Fixes "netperf ~50% regression with 2.6.33-rc1, bisect to 1b9508f" http://marc.info/?l=linux-kernel&m=126441481427331&w=4 Reported-by: Lin Ming Tested-by: Alex Shi Signed-off-by: Venkatesh Pallipadi Signed-off-by: Len Brown --- drivers/acpi/processor_idle.c | 28 ++++++++++++++++------------ 1 file changed, 16 insertions(+), 12 deletions(-) (limited to 'drivers') diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c index e88e8ae04fdb..cc978a8c00b7 100644 --- a/drivers/acpi/processor_idle.c +++ b/drivers/acpi/processor_idle.c @@ -880,12 +880,14 @@ static int acpi_idle_enter_simple(struct cpuidle_device *dev, return(acpi_idle_enter_c1(dev, state)); local_irq_disable(); - current_thread_info()->status &= ~TS_POLLING; - /* - * TS_POLLING-cleared state must be visible before we test - * NEED_RESCHED: - */ - smp_mb(); + if (cx->entry_method != ACPI_CSTATE_FFH) { + current_thread_info()->status &= ~TS_POLLING; + /* + * TS_POLLING-cleared state must be visible before we test + * NEED_RESCHED: + */ + smp_mb(); + } if (unlikely(need_resched())) { current_thread_info()->status |= TS_POLLING; @@ -965,12 +967,14 @@ static int acpi_idle_enter_bm(struct cpuidle_device *dev, } local_irq_disable(); - current_thread_info()->status &= ~TS_POLLING; - /* - * TS_POLLING-cleared state must be visible before we test - * NEED_RESCHED: - */ - smp_mb(); + if (cx->entry_method != ACPI_CSTATE_FFH) { + current_thread_info()->status &= ~TS_POLLING; + /* + * TS_POLLING-cleared state must be visible before we test + * NEED_RESCHED: + */ + smp_mb(); + } if (unlikely(need_resched())) { current_thread_info()->status |= TS_POLLING; -- cgit v1.2.3 From e7e65caefd57913260c82fb751f07655671ec47e Mon Sep 17 00:00:00 2001 From: Francisco Jerez Date: Wed, 10 Feb 2010 19:25:58 +0100 Subject: drm/nouveau: Fix up pre-nv17 analog load detection. Signed-off-by: Francisco Jerez Signed-off-by: Ben Skeggs --- drivers/gpu/drm/nouveau/nv04_dac.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/gpu/drm/nouveau/nv04_dac.c b/drivers/gpu/drm/nouveau/nv04_dac.c index d0e038d28948..1d73b15d70da 100644 --- a/drivers/gpu/drm/nouveau/nv04_dac.c +++ b/drivers/gpu/drm/nouveau/nv04_dac.c @@ -119,7 +119,7 @@ static enum drm_connector_status nv04_dac_detect(struct drm_encoder *encoder, struct drm_connector *connector) { struct drm_device *dev = encoder->dev; - uint8_t saved_seq1, saved_pi, saved_rpc1; + uint8_t saved_seq1, saved_pi, saved_rpc1, saved_cr_mode; uint8_t saved_palette0[3], saved_palette_mask; uint32_t saved_rtest_ctrl, saved_rgen_ctrl; int i; @@ -135,6 +135,9 @@ static enum drm_connector_status nv04_dac_detect(struct drm_encoder *encoder, /* only implemented for head A for now */ NVSetOwner(dev, 0); + saved_cr_mode = NVReadVgaCrtc(dev, 0, NV_CIO_CR_MODE_INDEX); + NVWriteVgaCrtc(dev, 0, NV_CIO_CR_MODE_INDEX, saved_cr_mode | 0x80); + saved_seq1 = NVReadVgaSeq(dev, 0, NV_VIO_SR_CLOCK_INDEX); NVWriteVgaSeq(dev, 0, NV_VIO_SR_CLOCK_INDEX, saved_seq1 & ~0x20); @@ -203,6 +206,7 @@ out: NVWriteVgaCrtc(dev, 0, NV_CIO_CRE_PIXEL_INDEX, saved_pi); NVWriteVgaCrtc(dev, 0, NV_CIO_CRE_RPC1_INDEX, saved_rpc1); NVWriteVgaSeq(dev, 0, NV_VIO_SR_CLOCK_INDEX, saved_seq1); + NVWriteVgaCrtc(dev, 0, NV_CIO_CR_MODE_INDEX, saved_cr_mode); if (blue == 0x18) { NV_INFO(dev, "Load detected on head A\n"); -- cgit v1.2.3 From 66b6ebaccb176a2068bbe328f162614dce524621 Mon Sep 17 00:00:00 2001 From: Ben Skeggs Date: Thu, 11 Feb 2010 10:23:30 +1000 Subject: drm/nv50: make nv50_mem_vm_{bind,unbind} operate only on vram GART is handled elsewhere, no reason to have the code for it here too. Signed-off-by: Ben Skeggs --- drivers/gpu/drm/nouveau/nouveau_mem.c | 43 +++++++++++------------------------ 1 file changed, 13 insertions(+), 30 deletions(-) (limited to 'drivers') diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c index 8f3a12f614ed..04885d2fb15f 100644 --- a/drivers/gpu/drm/nouveau/nouveau_mem.c +++ b/drivers/gpu/drm/nouveau/nouveau_mem.c @@ -285,53 +285,36 @@ nv50_mem_vm_bind_linear(struct drm_device *dev, uint64_t virt, uint32_t size, uint32_t flags, uint64_t phys) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpuobj **pgt; - unsigned psz, pfl, pages; - - if (virt >= dev_priv->vm_gart_base && - (virt + size) < (dev_priv->vm_gart_base + dev_priv->vm_gart_size)) { - psz = 12; - pgt = &dev_priv->gart_info.sg_ctxdma; - pfl = 0x21; - virt -= dev_priv->vm_gart_base; - } else - if (virt >= dev_priv->vm_vram_base && - (virt + size) < (dev_priv->vm_vram_base + dev_priv->vm_vram_size)) { - psz = 16; - pgt = dev_priv->vm_vram_pt; - pfl = 0x01; - virt -= dev_priv->vm_vram_base; - } else { - NV_ERROR(dev, "Invalid address: 0x%16llx-0x%16llx\n", - virt, virt + size - 1); - return -EINVAL; - } + unsigned pages; - pages = size >> psz; + virt -= dev_priv->vm_vram_base; + pages = size >> 16; dev_priv->engine.instmem.prepare_access(dev, true); if (flags & 0x80000000) { while (pages--) { - struct nouveau_gpuobj *pt = pgt[virt >> 29]; - unsigned pte = ((virt & 0x1fffffffULL) >> psz) << 1; + struct nouveau_gpuobj *pt = + dev_priv->vm_vram_pt[virt >> 29]; + unsigned pte = ((virt & 0x1fffffffULL) >> 16) << 1; nv_wo32(dev, pt, pte++, 0x00000000); nv_wo32(dev, pt, pte++, 0x00000000); - virt += (1 << psz); + virt += (1 << 16); } } else { while (pages--) { - struct nouveau_gpuobj *pt = pgt[virt >> 29]; - unsigned pte = ((virt & 0x1fffffffULL) >> psz) << 1; + struct nouveau_gpuobj *pt = + dev_priv->vm_vram_pt[virt >> 29]; + unsigned pte = ((virt & 0x1fffffffULL) >> 16) << 1; unsigned offset_h = upper_32_bits(phys) & 0xff; unsigned offset_l = lower_32_bits(phys); - nv_wo32(dev, pt, pte++, offset_l | pfl); + nv_wo32(dev, pt, pte++, offset_l | 1); nv_wo32(dev, pt, pte++, offset_h | flags); - phys += (1 << psz); - virt += (1 << psz); + phys += (1 << 16); + virt += (1 << 16); } } dev_priv->engine.instmem.finish_access(dev); -- cgit v1.2.3 From 4c27bd339d226175ac0e4dc3ab8289ba696db8be Mon Sep 17 00:00:00 2001 From: Ben Skeggs Date: Thu, 11 Feb 2010 10:25:53 +1000 Subject: drm/nv50: more efficient clearing of gpu page table entries Signed-off-by: Ben Skeggs --- drivers/gpu/drm/nouveau/nouveau_mem.c | 68 ++++++++++++++++++++++------------- 1 file changed, 44 insertions(+), 24 deletions(-) (limited to 'drivers') diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c index 04885d2fb15f..6832c4c969a3 100644 --- a/drivers/gpu/drm/nouveau/nouveau_mem.c +++ b/drivers/gpu/drm/nouveau/nouveau_mem.c @@ -291,31 +291,17 @@ nv50_mem_vm_bind_linear(struct drm_device *dev, uint64_t virt, uint32_t size, pages = size >> 16; dev_priv->engine.instmem.prepare_access(dev, true); - if (flags & 0x80000000) { - while (pages--) { - struct nouveau_gpuobj *pt = - dev_priv->vm_vram_pt[virt >> 29]; - unsigned pte = ((virt & 0x1fffffffULL) >> 16) << 1; + while (pages--) { + struct nouveau_gpuobj *pt = dev_priv->vm_vram_pt[virt >> 29]; + unsigned pte = ((virt & 0x1fffffffULL) >> 16) << 1; + unsigned offset_h = upper_32_bits(phys) & 0xff; + unsigned offset_l = lower_32_bits(phys); - nv_wo32(dev, pt, pte++, 0x00000000); - nv_wo32(dev, pt, pte++, 0x00000000); + nv_wo32(dev, pt, pte++, offset_l | 1); + nv_wo32(dev, pt, pte++, offset_h | flags); - virt += (1 << 16); - } - } else { - while (pages--) { - struct nouveau_gpuobj *pt = - dev_priv->vm_vram_pt[virt >> 29]; - unsigned pte = ((virt & 0x1fffffffULL) >> 16) << 1; - unsigned offset_h = upper_32_bits(phys) & 0xff; - unsigned offset_l = lower_32_bits(phys); - - nv_wo32(dev, pt, pte++, offset_l | 1); - nv_wo32(dev, pt, pte++, offset_h | flags); - - phys += (1 << 16); - virt += (1 << 16); - } + phys += (1 << 16); + virt += (1 << 16); } dev_priv->engine.instmem.finish_access(dev); @@ -339,7 +325,41 @@ nv50_mem_vm_bind_linear(struct drm_device *dev, uint64_t virt, uint32_t size, void nv50_mem_vm_unbind(struct drm_device *dev, uint64_t virt, uint32_t size) { - nv50_mem_vm_bind_linear(dev, virt, size, 0x80000000, 0); + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpuobj *pgt; + unsigned pages, pte, end; + + virt -= dev_priv->vm_vram_base; + pages = (size >> 16) << 1; + + dev_priv->engine.instmem.prepare_access(dev, true); + while (pages) { + pgt = dev_priv->vm_vram_pt[virt >> 29]; + pte = (virt & 0x1ffe0000ULL) >> 15; + + end = pte + pages; + if (end > 16384) + end = 16384; + pages -= (end - pte); + virt += (end - pte) << 15; + + while (pte < end) + nv_wo32(dev, pgt, pte++, 0); + } + dev_priv->engine.instmem.finish_access(dev); + + nv_wr32(dev, 0x100c80, 0x00050001); + if (!nv_wait(0x100c80, 0x00000001, 0x00000000)) { + NV_ERROR(dev, "timeout: (0x100c80 & 1) == 0 (2)\n"); + NV_ERROR(dev, "0x100c80 = 0x%08x\n", nv_rd32(dev, 0x100c80)); + return; + } + + nv_wr32(dev, 0x100c80, 0x00000001); + if (!nv_wait(0x100c80, 0x00000001, 0x00000000)) { + NV_ERROR(dev, "timeout: (0x100c80 & 1) == 0 (2)\n"); + NV_ERROR(dev, "0x100c80 = 0x%08x\n", nv_rd32(dev, 0x100c80)); + } } /* -- cgit v1.2.3 From 531e77139f26e8da32ee694b9ee5e6f4c764f1db Mon Sep 17 00:00:00 2001 From: Ben Skeggs Date: Thu, 11 Feb 2010 11:31:44 +1000 Subject: drm/nv50: improve vram page table construction This commit changes nouveau to construct PTEs which look very much like the ones the binary driver creates. I presume that filling multiple PTEs identically with length flags and the physical address of the start of a block of VRAM is a hint to the memory controller that it need not perform additional page table lookups for that range of addresses. Signed-off-by: Ben Skeggs --- drivers/gpu/drm/nouveau/nouveau_mem.c | 44 ++++++++++++++++++++++++++--------- 1 file changed, 33 insertions(+), 11 deletions(-) (limited to 'drivers') diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c index 6832c4c969a3..134fedbb7669 100644 --- a/drivers/gpu/drm/nouveau/nouveau_mem.c +++ b/drivers/gpu/drm/nouveau/nouveau_mem.c @@ -285,23 +285,45 @@ nv50_mem_vm_bind_linear(struct drm_device *dev, uint64_t virt, uint32_t size, uint32_t flags, uint64_t phys) { struct drm_nouveau_private *dev_priv = dev->dev_private; - unsigned pages; + struct nouveau_gpuobj *pgt; + unsigned block; + int i; - virt -= dev_priv->vm_vram_base; - pages = size >> 16; + virt = ((virt - dev_priv->vm_vram_base) >> 16) << 1; + size = (size >> 16) << 1; + phys |= ((uint64_t)flags << 32) | 1; dev_priv->engine.instmem.prepare_access(dev, true); - while (pages--) { - struct nouveau_gpuobj *pt = dev_priv->vm_vram_pt[virt >> 29]; - unsigned pte = ((virt & 0x1fffffffULL) >> 16) << 1; - unsigned offset_h = upper_32_bits(phys) & 0xff; + while (size) { + unsigned offset_h = upper_32_bits(phys); unsigned offset_l = lower_32_bits(phys); + unsigned pte, end; + + for (i = 7; i >= 0; i--) { + block = 1 << (i + 1); + if (size >= block && !(virt & (block - 1))) + break; + } + offset_l |= (i << 7); + + phys += block << 15; + size -= block; - nv_wo32(dev, pt, pte++, offset_l | 1); - nv_wo32(dev, pt, pte++, offset_h | flags); + while (block) { + pgt = dev_priv->vm_vram_pt[virt >> 14]; + pte = virt & 0x3ffe; - phys += (1 << 16); - virt += (1 << 16); + end = pte + block; + if (end > 16384) + end = 16384; + block -= (end - pte); + virt += (end - pte); + + while (pte < end) { + nv_wo32(dev, pgt, pte++, offset_l); + nv_wo32(dev, pgt, pte++, offset_h); + } + } } dev_priv->engine.instmem.finish_access(dev); -- cgit v1.2.3 From 76befb8c30cebe2af83fa346bdaf75b430893511 Mon Sep 17 00:00:00 2001 From: Ben Skeggs Date: Sat, 20 Feb 2010 08:06:36 +1000 Subject: drm/nv50: fix instmem binding on IGPs to point at stolen system memory This also modifies the unused PRAMIN PT entries to be all zeroes, can't really recall why I used 9/0 initially, just that it didn't work for some reason. It was likely masking a bug elsewhere that's since been fixed. Signed-off-by: Ben Skeggs --- drivers/gpu/drm/nouveau/nouveau_drv.h | 1 + drivers/gpu/drm/nouveau/nv50_instmem.c | 58 +++++++++++++++++++++++----------- 2 files changed, 41 insertions(+), 18 deletions(-) (limited to 'drivers') diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h index 5445cefdd03e..1c15ef37b71c 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drv.h +++ b/drivers/gpu/drm/nouveau/nouveau_drv.h @@ -583,6 +583,7 @@ struct drm_nouveau_private { uint64_t vm_end; struct nouveau_gpuobj *vm_vram_pt[NV50_VM_VRAM_NR]; int vm_vram_pt_nr; + uint64_t vram_sys_base; /* the mtrr covering the FB */ int fb_mtrr; diff --git a/drivers/gpu/drm/nouveau/nv50_instmem.c b/drivers/gpu/drm/nouveau/nv50_instmem.c index 94400f777e7f..f0dc4e36ef05 100644 --- a/drivers/gpu/drm/nouveau/nv50_instmem.c +++ b/drivers/gpu/drm/nouveau/nv50_instmem.c @@ -76,6 +76,11 @@ nv50_instmem_init(struct drm_device *dev) for (i = 0x1700; i <= 0x1710; i += 4) priv->save1700[(i-0x1700)/4] = nv_rd32(dev, i); + if (dev_priv->chipset == 0xaa || dev_priv->chipset == 0xac) + dev_priv->vram_sys_base = nv_rd32(dev, 0x100e10) << 12; + else + dev_priv->vram_sys_base = 0; + /* Reserve the last MiB of VRAM, we should probably try to avoid * setting up the below tables over the top of the VBIOS image at * some point. @@ -172,16 +177,28 @@ nv50_instmem_init(struct drm_device *dev) * We map the entire fake channel into the start of the PRAMIN BAR */ ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, pt_size, 0x1000, - 0, &priv->pramin_pt); + 0, &priv->pramin_pt); if (ret) return ret; - for (i = 0, v = c_offset; i < pt_size; i += 8, v += 0x1000) { - if (v < (c_offset + c_size)) - BAR0_WI32(priv->pramin_pt->gpuobj, i + 0, v | 1); - else - BAR0_WI32(priv->pramin_pt->gpuobj, i + 0, 0x00000009); + v = c_offset | 1; + if (dev_priv->vram_sys_base) { + v += dev_priv->vram_sys_base; + v |= 0x30; + } + + i = 0; + while (v < dev_priv->vram_sys_base + c_offset + c_size) { + BAR0_WI32(priv->pramin_pt->gpuobj, i + 0, v); + BAR0_WI32(priv->pramin_pt->gpuobj, i + 4, 0x00000000); + v += 0x1000; + i += 8; + } + + while (i < pt_size) { + BAR0_WI32(priv->pramin_pt->gpuobj, i + 0, 0x00000000); BAR0_WI32(priv->pramin_pt->gpuobj, i + 4, 0x00000000); + i += 8; } BAR0_WI32(chan->vm_pd, 0x00, priv->pramin_pt->instance | 0x63); @@ -416,7 +433,9 @@ nv50_instmem_bind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nv50_instmem_priv *priv = dev_priv->engine.instmem.priv; - uint32_t pte, pte_end, vram; + struct nouveau_gpuobj *pramin_pt = priv->pramin_pt->gpuobj; + uint32_t pte, pte_end; + uint64_t vram; if (!gpuobj->im_backing || !gpuobj->im_pramin || gpuobj->im_bound) return -EINVAL; @@ -424,20 +443,24 @@ nv50_instmem_bind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj) NV_DEBUG(dev, "st=0x%0llx sz=0x%0llx\n", gpuobj->im_pramin->start, gpuobj->im_pramin->size); - pte = (gpuobj->im_pramin->start >> 12) << 3; - pte_end = ((gpuobj->im_pramin->size >> 12) << 3) + pte; + pte = (gpuobj->im_pramin->start >> 12) << 1; + pte_end = ((gpuobj->im_pramin->size >> 12) << 1) + pte; vram = gpuobj->im_backing_start; NV_DEBUG(dev, "pramin=0x%llx, pte=%d, pte_end=%d\n", gpuobj->im_pramin->start, pte, pte_end); NV_DEBUG(dev, "first vram page: 0x%08x\n", gpuobj->im_backing_start); + vram |= 1; + if (dev_priv->vram_sys_base) { + vram += dev_priv->vram_sys_base; + vram |= 0x30; + } + dev_priv->engine.instmem.prepare_access(dev, true); while (pte < pte_end) { - nv_wo32(dev, priv->pramin_pt->gpuobj, (pte + 0)/4, vram | 1); - nv_wo32(dev, priv->pramin_pt->gpuobj, (pte + 4)/4, 0x00000000); - - pte += 8; + nv_wo32(dev, pramin_pt, pte++, lower_32_bits(vram)); + nv_wo32(dev, pramin_pt, pte++, upper_32_bits(vram)); vram += NV50_INSTMEM_PAGE_SIZE; } dev_priv->engine.instmem.finish_access(dev); @@ -470,14 +493,13 @@ nv50_instmem_unbind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj) if (gpuobj->im_bound == 0) return -EINVAL; - pte = (gpuobj->im_pramin->start >> 12) << 3; - pte_end = ((gpuobj->im_pramin->size >> 12) << 3) + pte; + pte = (gpuobj->im_pramin->start >> 12) << 1; + pte_end = ((gpuobj->im_pramin->size >> 12) << 1) + pte; dev_priv->engine.instmem.prepare_access(dev, true); while (pte < pte_end) { - nv_wo32(dev, priv->pramin_pt->gpuobj, (pte + 0)/4, 0x00000009); - nv_wo32(dev, priv->pramin_pt->gpuobj, (pte + 4)/4, 0x00000000); - pte += 8; + nv_wo32(dev, priv->pramin_pt->gpuobj, pte++, 0x00000000); + nv_wo32(dev, priv->pramin_pt->gpuobj, pte++, 0x00000000); } dev_priv->engine.instmem.finish_access(dev); -- cgit v1.2.3 From 6c42966768b0254f465a8f451333795283f53d22 Mon Sep 17 00:00:00 2001 From: Ben Skeggs Date: Sat, 20 Feb 2010 08:10:11 +1000 Subject: drm/nv50: fix vram ptes on IGPs to point at stolen system memory Signed-off-by: Ben Skeggs --- drivers/gpu/drm/nouveau/nouveau_mem.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c index 134fedbb7669..2dc09dbd817d 100644 --- a/drivers/gpu/drm/nouveau/nouveau_mem.c +++ b/drivers/gpu/drm/nouveau/nouveau_mem.c @@ -291,7 +291,13 @@ nv50_mem_vm_bind_linear(struct drm_device *dev, uint64_t virt, uint32_t size, virt = ((virt - dev_priv->vm_vram_base) >> 16) << 1; size = (size >> 16) << 1; - phys |= ((uint64_t)flags << 32) | 1; + + phys |= ((uint64_t)flags << 32); + phys |= 1; + if (dev_priv->vram_sys_base) { + phys += dev_priv->vram_sys_base; + phys |= 0x30; + } dev_priv->engine.instmem.prepare_access(dev, true); while (size) { -- cgit v1.2.3 From 115079aad949cba31755eb4e2576edba7fddfdbc Mon Sep 17 00:00:00 2001 From: Jens Rottmann Date: Mon, 22 Feb 2010 12:44:20 -0800 Subject: geode-mfgpt: restore previous behavior for selecting IRQ geode-mfgpt: restore previous behavior for selecting IRQ The MFGPT IRQ used to be, in order of decreasing priority, * IRQ supplied by the user as a boot-time parameter, * IRQ previously set by the BIOS or another driver, * default IRQ given at compile time. Return to this behavior, which got broken when splitting the MFGPT/clocksource driver for 2.6.33-rc1. Signed-off-by: Jens Rottmann Acked-by: Andres Salomon Cc: Jordan Crouse Cc: Ingo Molnar Cc: john stultz Acked-by: Thomas Gleixner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/clocksource/cs5535-clockevt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/clocksource/cs5535-clockevt.c b/drivers/clocksource/cs5535-clockevt.c index 27d20fac19d1..b314a999aabe 100644 --- a/drivers/clocksource/cs5535-clockevt.c +++ b/drivers/clocksource/cs5535-clockevt.c @@ -21,7 +21,7 @@ #define DRV_NAME "cs5535-clockevt" -static int timer_irq = CONFIG_CS5535_MFGPT_DEFAULT_IRQ; +static int timer_irq; module_param_named(irq, timer_irq, int, 0644); MODULE_PARM_DESC(irq, "Which IRQ to use for the clock source MFGPT ticks."); -- cgit v1.2.3 From 89f3f2199084a160a3a45fa6d9af235696321758 Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Mon, 22 Feb 2010 12:44:22 -0800 Subject: efifb: fix framebuffer handoff Commit 4410f3910947dcea8672280b3adecd53cec4e85e ("fbdev: add support for handoff from firmware to hw framebuffers") didn't add fb_destroy operation to efifb. Fix it and change aperture_size to match size passed to request_mem_region. Addresses http://bugzilla.kernel.org/show_bug.cgi?id=15151 Signed-off-by: Marcin Slusarz Reported-by: Alex Zhavnerchik Tested-by: Alex Zhavnerchik Acked-by: Peter Jones Cc: Huang Ying Cc: Dave Airlie Cc: "Rafael J. Wysocki" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/video/efifb.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/video/efifb.c b/drivers/video/efifb.c index eb12182b2059..d25df51bb0d2 100644 --- a/drivers/video/efifb.c +++ b/drivers/video/efifb.c @@ -161,8 +161,17 @@ static int efifb_setcolreg(unsigned regno, unsigned red, unsigned green, return 0; } +static void efifb_destroy(struct fb_info *info) +{ + if (info->screen_base) + iounmap(info->screen_base); + release_mem_region(info->aperture_base, info->aperture_size); + framebuffer_release(info); +} + static struct fb_ops efifb_ops = { .owner = THIS_MODULE, + .fb_destroy = efifb_destroy, .fb_setcolreg = efifb_setcolreg, .fb_fillrect = cfb_fillrect, .fb_copyarea = cfb_copyarea, @@ -281,7 +290,7 @@ static int __init efifb_probe(struct platform_device *dev) info->par = NULL; info->aperture_base = efifb_fix.smem_start; - info->aperture_size = size_total; + info->aperture_size = size_remap; info->screen_base = ioremap(efifb_fix.smem_start, efifb_fix.smem_len); if (!info->screen_base) { -- cgit v1.2.3 From 4e4ddd47774313accc86b233d6ca2c6a9037a671 Mon Sep 17 00:00:00 2001 From: Thomas Hellstrom Date: Sun, 21 Feb 2010 14:54:55 +0000 Subject: drm/vmwgfx: Fix queries if no dma buffer thrashing is occuring. Intercept query commands and apply relocations to their guest pointers. Signed-off-by: Thomas Hellstrom Signed-off-by: Dave Airlie --- drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c | 108 +++++++++++++++++++++++++++----- 1 file changed, 92 insertions(+), 16 deletions(-) (limited to 'drivers') diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c index d69caf92ffe7..0897359b3e4e 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c @@ -182,25 +182,19 @@ static int vmw_cmd_present_check(struct vmw_private *dev_priv, return vmw_cmd_sid_check(dev_priv, sw_context, &cmd->body.sid); } -static int vmw_cmd_dma(struct vmw_private *dev_priv, - struct vmw_sw_context *sw_context, - SVGA3dCmdHeader *header) +static int vmw_translate_guest_ptr(struct vmw_private *dev_priv, + struct vmw_sw_context *sw_context, + SVGAGuestPtr *ptr, + struct vmw_dma_buffer **vmw_bo_p) { - uint32_t handle; struct vmw_dma_buffer *vmw_bo = NULL; struct ttm_buffer_object *bo; - struct vmw_surface *srf = NULL; - struct vmw_dma_cmd { - SVGA3dCmdHeader header; - SVGA3dCmdSurfaceDMA dma; - } *cmd; + uint32_t handle = ptr->gmrId; struct vmw_relocation *reloc; - int ret; uint32_t cur_validate_node; struct ttm_validate_buffer *val_buf; + int ret; - cmd = container_of(header, struct vmw_dma_cmd, header); - handle = cmd->dma.guest.ptr.gmrId; ret = vmw_user_dmabuf_lookup(sw_context->tfile, handle, &vmw_bo); if (unlikely(ret != 0)) { DRM_ERROR("Could not find or use GMR region.\n"); @@ -209,14 +203,14 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv, bo = &vmw_bo->base; if (unlikely(sw_context->cur_reloc >= VMWGFX_MAX_RELOCATIONS)) { - DRM_ERROR("Max number of DMA commands per submission" + DRM_ERROR("Max number relocations per submission" " exceeded\n"); ret = -EINVAL; goto out_no_reloc; } reloc = &sw_context->relocs[sw_context->cur_reloc++]; - reloc->location = &cmd->dma.guest.ptr; + reloc->location = ptr; cur_validate_node = vmw_dmabuf_validate_node(bo, sw_context->cur_val_buf); if (unlikely(cur_validate_node >= VMWGFX_MAX_GMRS)) { @@ -234,7 +228,89 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv, list_add_tail(&val_buf->head, &sw_context->validate_nodes); ++sw_context->cur_val_buf; } + *vmw_bo_p = vmw_bo; + return 0; + +out_no_reloc: + vmw_dmabuf_unreference(&vmw_bo); + vmw_bo_p = NULL; + return ret; +} + +static int vmw_cmd_end_query(struct vmw_private *dev_priv, + struct vmw_sw_context *sw_context, + SVGA3dCmdHeader *header) +{ + struct vmw_dma_buffer *vmw_bo; + struct vmw_query_cmd { + SVGA3dCmdHeader header; + SVGA3dCmdEndQuery q; + } *cmd; + int ret; + + cmd = container_of(header, struct vmw_query_cmd, header); + ret = vmw_cmd_cid_check(dev_priv, sw_context, header); + if (unlikely(ret != 0)) + return ret; + + ret = vmw_translate_guest_ptr(dev_priv, sw_context, + &cmd->q.guestResult, + &vmw_bo); + if (unlikely(ret != 0)) + return ret; + + vmw_dmabuf_unreference(&vmw_bo); + return 0; +} +static int vmw_cmd_wait_query(struct vmw_private *dev_priv, + struct vmw_sw_context *sw_context, + SVGA3dCmdHeader *header) +{ + struct vmw_dma_buffer *vmw_bo; + struct vmw_query_cmd { + SVGA3dCmdHeader header; + SVGA3dCmdWaitForQuery q; + } *cmd; + int ret; + + cmd = container_of(header, struct vmw_query_cmd, header); + ret = vmw_cmd_cid_check(dev_priv, sw_context, header); + if (unlikely(ret != 0)) + return ret; + + ret = vmw_translate_guest_ptr(dev_priv, sw_context, + &cmd->q.guestResult, + &vmw_bo); + if (unlikely(ret != 0)) + return ret; + + vmw_dmabuf_unreference(&vmw_bo); + return 0; +} + + +static int vmw_cmd_dma(struct vmw_private *dev_priv, + struct vmw_sw_context *sw_context, + SVGA3dCmdHeader *header) +{ + struct vmw_dma_buffer *vmw_bo = NULL; + struct ttm_buffer_object *bo; + struct vmw_surface *srf = NULL; + struct vmw_dma_cmd { + SVGA3dCmdHeader header; + SVGA3dCmdSurfaceDMA dma; + } *cmd; + int ret; + + cmd = container_of(header, struct vmw_dma_cmd, header); + ret = vmw_translate_guest_ptr(dev_priv, sw_context, + &cmd->dma.guest.ptr, + &vmw_bo); + if (unlikely(ret != 0)) + return ret; + + bo = &vmw_bo->base; ret = vmw_user_surface_lookup_handle(dev_priv, sw_context->tfile, cmd->dma.host.sid, &srf); if (ret) { @@ -379,8 +455,8 @@ static vmw_cmd_func vmw_cmd_funcs[SVGA_3D_CMD_MAX] = { VMW_CMD_DEF(SVGA_3D_CMD_DRAW_PRIMITIVES, &vmw_cmd_draw), VMW_CMD_DEF(SVGA_3D_CMD_SETSCISSORRECT, &vmw_cmd_cid_check), VMW_CMD_DEF(SVGA_3D_CMD_BEGIN_QUERY, &vmw_cmd_cid_check), - VMW_CMD_DEF(SVGA_3D_CMD_END_QUERY, &vmw_cmd_cid_check), - VMW_CMD_DEF(SVGA_3D_CMD_WAIT_FOR_QUERY, &vmw_cmd_cid_check), + VMW_CMD_DEF(SVGA_3D_CMD_END_QUERY, &vmw_cmd_end_query), + VMW_CMD_DEF(SVGA_3D_CMD_WAIT_FOR_QUERY, &vmw_cmd_wait_query), VMW_CMD_DEF(SVGA_3D_CMD_PRESENT_READBACK, &vmw_cmd_ok), VMW_CMD_DEF(SVGA_3D_CMD_BLIT_SURFACE_TO_SCREEN, &vmw_cmd_blt_surf_screen_check) -- cgit v1.2.3 From b5abb028e214cca68f4231d4f3bc0847ddbc986e Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Fri, 19 Feb 2010 17:54:53 +0000 Subject: e1000: Fix DMA mapping error handling on RX Check for error return from pci_map_single/pci_map_page and clean up. With this and the previous patch the driver was able to handle a significant percentage of errors (I set the fault injection rate to 10% and could still download large files at a reasonable speed). Signed-off-by: Anton Blanchard Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller --- drivers/net/e1000/e1000_main.c | 19 ++++++++++++++++++- 1 file changed, 18 insertions(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/net/e1000/e1000_main.c b/drivers/net/e1000/e1000_main.c index d29bb532eccf..765543663a4f 100644 --- a/drivers/net/e1000/e1000_main.c +++ b/drivers/net/e1000/e1000_main.c @@ -4006,11 +4006,21 @@ check_page: } } - if (!buffer_info->dma) + if (!buffer_info->dma) { buffer_info->dma = pci_map_page(pdev, buffer_info->page, 0, buffer_info->length, PCI_DMA_FROMDEVICE); + if (pci_dma_mapping_error(pdev, buffer_info->dma)) { + put_page(buffer_info->page); + dev_kfree_skb(skb); + buffer_info->page = NULL; + buffer_info->skb = NULL; + buffer_info->dma = 0; + adapter->alloc_rx_buff_failed++; + break; /* while !buffer_info->skb */ + } + } rx_desc = E1000_RX_DESC(*rx_ring, i); rx_desc->buffer_addr = cpu_to_le64(buffer_info->dma); @@ -4101,6 +4111,13 @@ map_skb: skb->data, buffer_info->length, PCI_DMA_FROMDEVICE); + if (pci_dma_mapping_error(pdev, buffer_info->dma)) { + dev_kfree_skb(skb); + buffer_info->skb = NULL; + buffer_info->dma = 0; + adapter->alloc_rx_buff_failed++; + break; /* while !buffer_info->skb */ + } /* * XXX if it was allocated cleanly it will never map to a -- cgit v1.2.3 From cac43a1b7b091b17113502e4128dcb0ff7e3503d Mon Sep 17 00:00:00 2001 From: Torgny Johansson Date: Fri, 19 Feb 2010 01:59:15 +0000 Subject: cdc_ether: new PID for Ericsson C3607w to the whitelist (resubmit) This patch adds a new vid/pid to the cdc_ether whitelist. Device added: - Ericsson Mobile Broadband variant C3607w Signed-off-by: Torgny Johansson -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Signed-off-by: David S. Miller --- drivers/net/usb/cdc_ether.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'drivers') diff --git a/drivers/net/usb/cdc_ether.c b/drivers/net/usb/cdc_ether.c index 4f27f022fbf7..5f3b9eaeb04f 100644 --- a/drivers/net/usb/cdc_ether.c +++ b/drivers/net/usb/cdc_ether.c @@ -583,6 +583,11 @@ static const struct usb_device_id products [] = { USB_DEVICE_AND_INTERFACE_INFO(0x0bdb, 0x1049, USB_CLASS_COMM, USB_CDC_SUBCLASS_MDLM, USB_CDC_PROTO_NONE), .driver_info = (unsigned long) &mbm_info, +}, { + /* Ericsson C3607w ver 2 */ + USB_DEVICE_AND_INTERFACE_INFO(0x0bdb, 0x190b, USB_CLASS_COMM, + USB_CDC_SUBCLASS_MDLM, USB_CDC_PROTO_NONE), + .driver_info = (unsigned long) &mbm_info, }, { /* Toshiba F3507g */ USB_DEVICE_AND_INTERFACE_INFO(0x0930, 0x130b, USB_CLASS_COMM, -- cgit v1.2.3 From 662a96bd6f020782dfbdc0d0bd177c7dbb556687 Mon Sep 17 00:00:00 2001 From: Atsushi Nemoto Date: Fri, 19 Feb 2010 05:13:58 +0000 Subject: tc35815: Remove a wrong netif_wake_queue() call which triggers BUG_ON The netif_wake_queue() is called correctly (i.e. only on !txfull condition) from txdone routine. So Unconditional call to the netif_wake_queue() here is wrong. This might cause calling of start_xmit routine on txfull state and trigger BUG_ON. This bug does not happen when NAPI disabled. After txdone there must be at least one free tx slot. But with NAPI, this is not true anymore and the BUG_ON can hits on heavy load. In this driver NAPI was enabled on 2.6.33-rc1 so this is regression from 2.6.32 kernel. Reported-by: Ralf Roesch Signed-off-by: Atsushi Nemoto Signed-off-by: David S. Miller --- drivers/net/tc35815.c | 1 - 1 file changed, 1 deletion(-) (limited to 'drivers') diff --git a/drivers/net/tc35815.c b/drivers/net/tc35815.c index 75a669d48e5e..d71c1976072e 100644 --- a/drivers/net/tc35815.c +++ b/drivers/net/tc35815.c @@ -1437,7 +1437,6 @@ static int tc35815_do_interrupt(struct net_device *dev, u32 status, int limit) /* Transmit complete. */ lp->lstats.tx_ints++; tc35815_txdone(dev); - netif_wake_queue(dev); if (ret < 0) ret = 0; } -- cgit v1.2.3