From e5e8c5b90a1ae249930fcf7403f3757686cf1a7b Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Thu, 11 Jun 2009 10:03:42 +0200 Subject: dma-debug: check for sg_call_ents in best-fit algorithm too If we don't check for sg_call_ents the hash_bucket_find function might still return the wrong dma_debug_entry for sg mappings. Signed-off-by: Joerg Roedel diff --git a/lib/dma-debug.c b/lib/dma-debug.c index ad65fc0..c71e2dd 100644 --- a/lib/dma-debug.c +++ b/lib/dma-debug.c @@ -262,11 +262,12 @@ static struct dma_debug_entry *hash_bucket_find(struct hash_bucket *bucket, */ matches += 1; match_lvl = 0; - entry->size == ref->size ? ++match_lvl : match_lvl; - entry->type == ref->type ? ++match_lvl : match_lvl; - entry->direction == ref->direction ? ++match_lvl : match_lvl; + entry->size == ref->size ? ++match_lvl : 0; + entry->type == ref->type ? ++match_lvl : 0; + entry->direction == ref->direction ? ++match_lvl : 0; + entry->sg_call_ents == ref->sg_call_ents ? ++match_lvl : 0; - if (match_lvl == 3) { + if (match_lvl == 4) { /* perfect-fit - return the result */ return entry; } else if (match_lvl > last_lvl) { @@ -1076,16 +1077,14 @@ void debug_dma_unmap_sg(struct device *dev, struct scatterlist *sglist, .dev_addr = sg_dma_address(s), .size = sg_dma_len(s), .direction = dir, - .sg_call_ents = 0, + .sg_call_ents = nelems, }; if (mapped_ents && i >= mapped_ents) break; - if (!i) { - ref.sg_call_ents = nelems; + if (!i) mapped_ents = get_nr_mapped_entries(dev, s); - } check_unmap(&ref); } -- cgit v0.10.2 From aa010efb7b6cd0dfbea8ecf37a6ab587dc2a8560 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Jun 2009 15:25:06 +0200 Subject: dma-debug: be more careful when building reference entries The current code is not very careful when it builds reference dma_debug_entries which get passed to hash_bucket_find(). But since this function changed to a best-fit algorithm these entries have to be more acurate. This patch adds this higher level of accuracy. Signed-off-by: Joerg Roedel diff --git a/lib/dma-debug.c b/lib/dma-debug.c index c71e2dd..3b93129 100644 --- a/lib/dma-debug.c +++ b/lib/dma-debug.c @@ -874,72 +874,68 @@ static void check_for_illegal_area(struct device *dev, void *addr, u64 size) "[addr=%p] [size=%llu]\n", addr, size); } -static void check_sync(struct device *dev, dma_addr_t addr, - u64 size, u64 offset, int direction, bool to_cpu) +static void check_sync(struct device *dev, + struct dma_debug_entry *ref, + bool to_cpu) { - struct dma_debug_entry ref = { - .dev = dev, - .dev_addr = addr, - .size = size, - .direction = direction, - }; struct dma_debug_entry *entry; struct hash_bucket *bucket; unsigned long flags; - bucket = get_hash_bucket(&ref, &flags); + bucket = get_hash_bucket(ref, &flags); - entry = hash_bucket_find(bucket, &ref); + entry = hash_bucket_find(bucket, ref); if (!entry) { err_printk(dev, NULL, "DMA-API: device driver tries " "to sync DMA memory it has not allocated " "[device address=0x%016llx] [size=%llu bytes]\n", - (unsigned long long)addr, size); + (unsigned long long)ref->dev_addr, ref->size); goto out; } - if ((offset + size) > entry->size) { + if (ref->size > entry->size) { err_printk(dev, entry, "DMA-API: device driver syncs" " DMA memory outside allocated range " "[device address=0x%016llx] " - "[allocation size=%llu bytes] [sync offset=%llu] " - "[sync size=%llu]\n", entry->dev_addr, entry->size, - offset, size); + "[allocation size=%llu bytes] " + "[sync offset+size=%llu]\n", + entry->dev_addr, entry->size, + ref->size); } - if (direction != entry->direction) { + if (ref->direction != entry->direction) { err_printk(dev, entry, "DMA-API: device driver syncs " "DMA memory with different direction " "[device address=0x%016llx] [size=%llu bytes] " "[mapped with %s] [synced with %s]\n", - (unsigned long long)addr, entry->size, + (unsigned long long)ref->dev_addr, entry->size, dir2name[entry->direction], - dir2name[direction]); + dir2name[ref->direction]); } if (entry->direction == DMA_BIDIRECTIONAL) goto out; if (to_cpu && !(entry->direction == DMA_FROM_DEVICE) && - !(direction == DMA_TO_DEVICE)) + !(ref->direction == DMA_TO_DEVICE)) err_printk(dev, entry, "DMA-API: device driver syncs " "device read-only DMA memory for cpu " "[device address=0x%016llx] [size=%llu bytes] " "[mapped with %s] [synced with %s]\n", - (unsigned long long)addr, entry->size, + (unsigned long long)ref->dev_addr, entry->size, dir2name[entry->direction], - dir2name[direction]); + dir2name[ref->direction]); if (!to_cpu && !(entry->direction == DMA_TO_DEVICE) && - !(direction == DMA_FROM_DEVICE)) + !(ref->direction == DMA_FROM_DEVICE)) err_printk(dev, entry, "DMA-API: device driver syncs " "device write-only DMA memory to device " "[device address=0x%016llx] [size=%llu bytes] " "[mapped with %s] [synced with %s]\n", - (unsigned long long)addr, entry->size, + (unsigned long long)ref->dev_addr, entry->size, dir2name[entry->direction], - dir2name[direction]); + dir2name[ref->direction]); out: put_hash_bucket(bucket, &flags); @@ -1037,19 +1033,16 @@ void debug_dma_map_sg(struct device *dev, struct scatterlist *sg, } EXPORT_SYMBOL(debug_dma_map_sg); -static int get_nr_mapped_entries(struct device *dev, struct scatterlist *s) +static int get_nr_mapped_entries(struct device *dev, + struct dma_debug_entry *ref) { - struct dma_debug_entry *entry, ref; + struct dma_debug_entry *entry; struct hash_bucket *bucket; unsigned long flags; int mapped_ents; - ref.dev = dev; - ref.dev_addr = sg_dma_address(s); - ref.size = sg_dma_len(s), - - bucket = get_hash_bucket(&ref, &flags); - entry = hash_bucket_find(bucket, &ref); + bucket = get_hash_bucket(ref, &flags); + entry = hash_bucket_find(bucket, ref); mapped_ents = 0; if (entry) @@ -1084,7 +1077,7 @@ void debug_dma_unmap_sg(struct device *dev, struct scatterlist *sglist, break; if (!i) - mapped_ents = get_nr_mapped_entries(dev, s); + mapped_ents = get_nr_mapped_entries(dev, &ref); check_unmap(&ref); } @@ -1139,10 +1132,19 @@ EXPORT_SYMBOL(debug_dma_free_coherent); void debug_dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle, size_t size, int direction) { + struct dma_debug_entry ref; + if (unlikely(global_disable)) return; - check_sync(dev, dma_handle, size, 0, direction, true); + ref.type = dma_debug_single; + ref.dev = dev; + ref.dev_addr = dma_handle; + ref.size = size; + ref.direction = direction; + ref.sg_call_ents = 0; + + check_sync(dev, &ref, true); } EXPORT_SYMBOL(debug_dma_sync_single_for_cpu); @@ -1150,10 +1152,19 @@ void debug_dma_sync_single_for_device(struct device *dev, dma_addr_t dma_handle, size_t size, int direction) { + struct dma_debug_entry ref; + if (unlikely(global_disable)) return; - check_sync(dev, dma_handle, size, 0, direction, false); + ref.type = dma_debug_single; + ref.dev = dev; + ref.dev_addr = dma_handle; + ref.size = size; + ref.direction = direction; + ref.sg_call_ents = 0; + + check_sync(dev, &ref, false); } EXPORT_SYMBOL(debug_dma_sync_single_for_device); @@ -1162,10 +1173,19 @@ void debug_dma_sync_single_range_for_cpu(struct device *dev, unsigned long offset, size_t size, int direction) { + struct dma_debug_entry ref; + if (unlikely(global_disable)) return; - check_sync(dev, dma_handle, size, offset, direction, true); + ref.type = dma_debug_single; + ref.dev = dev; + ref.dev_addr = dma_handle; + ref.size = offset + size; + ref.direction = direction; + ref.sg_call_ents = 0; + + check_sync(dev, &ref, true); } EXPORT_SYMBOL(debug_dma_sync_single_range_for_cpu); @@ -1174,10 +1194,19 @@ void debug_dma_sync_single_range_for_device(struct device *dev, unsigned long offset, size_t size, int direction) { + struct dma_debug_entry ref; + if (unlikely(global_disable)) return; - check_sync(dev, dma_handle, size, offset, direction, false); + ref.type = dma_debug_single; + ref.dev = dev; + ref.dev_addr = dma_handle; + ref.size = offset + size; + ref.direction = direction; + ref.sg_call_ents = 0; + + check_sync(dev, &ref, false); } EXPORT_SYMBOL(debug_dma_sync_single_range_for_device); @@ -1191,14 +1220,24 @@ void debug_dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sg, return; for_each_sg(sg, s, nelems, i) { + + struct dma_debug_entry ref = { + .type = dma_debug_sg, + .dev = dev, + .paddr = sg_phys(s), + .dev_addr = sg_dma_address(s), + .size = sg_dma_len(s), + .direction = direction, + .sg_call_ents = nelems, + }; + if (!i) - mapped_ents = get_nr_mapped_entries(dev, s); + mapped_ents = get_nr_mapped_entries(dev, &ref); if (i >= mapped_ents) break; - check_sync(dev, sg_dma_address(s), sg_dma_len(s), 0, - direction, true); + check_sync(dev, &ref, true); } } EXPORT_SYMBOL(debug_dma_sync_sg_for_cpu); @@ -1213,14 +1252,23 @@ void debug_dma_sync_sg_for_device(struct device *dev, struct scatterlist *sg, return; for_each_sg(sg, s, nelems, i) { + + struct dma_debug_entry ref = { + .type = dma_debug_sg, + .dev = dev, + .paddr = sg_phys(s), + .dev_addr = sg_dma_address(s), + .size = sg_dma_len(s), + .direction = direction, + .sg_call_ents = nelems, + }; if (!i) - mapped_ents = get_nr_mapped_entries(dev, s); + mapped_ents = get_nr_mapped_entries(dev, &ref); if (i >= mapped_ents) break; - check_sync(dev, sg_dma_address(s), sg_dma_len(s), 0, - direction, false); + check_sync(dev, &ref, false); } } EXPORT_SYMBOL(debug_dma_sync_sg_for_device); -- cgit v0.10.2 From f75dcc87feab791847605044311a4a8e9335da91 Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Wed, 17 Jun 2009 08:22:32 +0200 Subject: ALSA: hda - Fix memory leak at codec creation The codec->modelname field is allocated twice in snd_hda_codec_new(). Signed-off-by: Takashi Iwai diff --git a/sound/pci/hda/hda_codec.c b/sound/pci/hda/hda_codec.c index 562403a..462e2ce 100644 --- a/sound/pci/hda/hda_codec.c +++ b/sound/pci/hda/hda_codec.c @@ -972,8 +972,6 @@ int /*__devinit*/ snd_hda_codec_new(struct hda_bus *bus, unsigned int codec_addr snd_hda_codec_read(codec, nid, 0, AC_VERB_GET_SUBSYSTEM_ID, 0); } - if (bus->modelname) - codec->modelname = kstrdup(bus->modelname, GFP_KERNEL); /* power-up all before initialization */ hda_set_power_state(codec, -- cgit v0.10.2 From df01b8af5627b7e28c087559cbce4f359ce07c49 Mon Sep 17 00:00:00 2001 From: Sasha Alexandr Date: Tue, 16 Jun 2009 14:46:17 -0400 Subject: ALSA: HDA - Add pci-quirk for MSI MS-7350 motherboard. Add pci-quirk for MSI MS-7350 motherboard with Realtek ALC888. Signed-off-by: Sasha Alexandr Signed-off-by: Takashi Iwai diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index d22b260..a2cebd7 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -9069,6 +9069,7 @@ static struct snd_pci_quirk alc883_cfg_tbl[] = { SND_PCI_QUIRK(0x1462, 0x7267, "MSI", ALC883_3ST_6ch_DIG), SND_PCI_QUIRK(0x1462, 0x7280, "MSI", ALC883_6ST_DIG), SND_PCI_QUIRK(0x1462, 0x7327, "MSI", ALC883_6ST_DIG), + SND_PCI_QUIRK(0x1462, 0x7350, "MSI", ALC883_6ST_DIG), SND_PCI_QUIRK(0x1462, 0xa422, "MSI", ALC883_TARGA_2ch_DIG), SND_PCI_QUIRK(0x147b, 0x1083, "Abit IP35-PRO", ALC883_6ST_DIG), SND_PCI_QUIRK(0x1558, 0x0721, "Clevo laptop M720R", ALC883_CLEVO_M720), -- cgit v0.10.2 From c259249f7d7a6e534f3bafe046929fee236eebfa Mon Sep 17 00:00:00 2001 From: Sasha Alexandr Date: Tue, 16 Jun 2009 14:52:54 -0400 Subject: ALSA: HDA - Name-fixes in code (tagra/targa) Correct some cut+paste typos from 'tagra' to 'targa'. Signed-off-by: Sasha Alexandr Signed-off-by: Takashi Iwai diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index a2cebd7..dc77d75 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -8068,7 +8068,7 @@ static struct snd_kcontrol_new alc883_fivestack_mixer[] = { { } /* end */ }; -static struct snd_kcontrol_new alc883_tagra_mixer[] = { +static struct snd_kcontrol_new alc883_targa_mixer[] = { HDA_CODEC_VOLUME("Front Playback Volume", 0x0c, 0x0, HDA_OUTPUT), HDA_CODEC_MUTE("Headphone Playback Switch", 0x14, 0x0, HDA_OUTPUT), HDA_CODEC_MUTE("Front Playback Switch", 0x1b, 0x0, HDA_OUTPUT), @@ -8088,7 +8088,7 @@ static struct snd_kcontrol_new alc883_tagra_mixer[] = { { } /* end */ }; -static struct snd_kcontrol_new alc883_tagra_2ch_mixer[] = { +static struct snd_kcontrol_new alc883_targa_2ch_mixer[] = { HDA_CODEC_VOLUME("Front Playback Volume", 0x0c, 0x0, HDA_OUTPUT), HDA_CODEC_MUTE("Headphone Playback Switch", 0x14, 0x0, HDA_OUTPUT), HDA_CODEC_MUTE("Front Playback Switch", 0x1b, 0x0, HDA_OUTPUT), @@ -8417,7 +8417,7 @@ static struct hda_verb alc883_2ch_fujitsu_pi2515_verbs[] = { { } /* end */ }; -static struct hda_verb alc883_tagra_verbs[] = { +static struct hda_verb alc883_targa_verbs[] = { {0x0c, AC_VERB_SET_AMP_GAIN_MUTE, AMP_IN_UNMUTE(0)}, {0x0c, AC_VERB_SET_AMP_GAIN_MUTE, AMP_IN_UNMUTE(1)}, @@ -8626,8 +8626,8 @@ static void alc883_medion_md2_init_hook(struct hda_codec *codec) } /* toggle speaker-output according to the hp-jack state */ -#define alc883_tagra_init_hook alc882_targa_init_hook -#define alc883_tagra_unsol_event alc882_targa_unsol_event +#define alc883_targa_init_hook alc882_targa_init_hook +#define alc883_targa_unsol_event alc882_targa_unsol_event static void alc883_clevo_m720_mic_automute(struct hda_codec *codec) { @@ -9166,8 +9166,8 @@ static struct alc_config_preset alc883_presets[] = { .input_mux = &alc883_capture_source, }, [ALC883_TARGA_DIG] = { - .mixers = { alc883_tagra_mixer, alc883_chmode_mixer }, - .init_verbs = { alc883_init_verbs, alc883_tagra_verbs}, + .mixers = { alc883_targa_mixer, alc883_chmode_mixer }, + .init_verbs = { alc883_init_verbs, alc883_targa_verbs}, .num_dacs = ARRAY_SIZE(alc883_dac_nids), .dac_nids = alc883_dac_nids, .dig_out_nid = ALC883_DIGOUT_NID, @@ -9175,12 +9175,12 @@ static struct alc_config_preset alc883_presets[] = { .channel_mode = alc883_3ST_6ch_modes, .need_dac_fix = 1, .input_mux = &alc883_capture_source, - .unsol_event = alc883_tagra_unsol_event, - .init_hook = alc883_tagra_init_hook, + .unsol_event = alc883_targa_unsol_event, + .init_hook = alc883_targa_init_hook, }, [ALC883_TARGA_2ch_DIG] = { - .mixers = { alc883_tagra_2ch_mixer}, - .init_verbs = { alc883_init_verbs, alc883_tagra_verbs}, + .mixers = { alc883_targa_2ch_mixer}, + .init_verbs = { alc883_init_verbs, alc883_targa_verbs}, .num_dacs = ARRAY_SIZE(alc883_dac_nids), .dac_nids = alc883_dac_nids, .adc_nids = alc883_adc_nids_alt, @@ -9189,13 +9189,13 @@ static struct alc_config_preset alc883_presets[] = { .num_channel_mode = ARRAY_SIZE(alc883_3ST_2ch_modes), .channel_mode = alc883_3ST_2ch_modes, .input_mux = &alc883_capture_source, - .unsol_event = alc883_tagra_unsol_event, - .init_hook = alc883_tagra_init_hook, + .unsol_event = alc883_targa_unsol_event, + .init_hook = alc883_targa_init_hook, }, [ALC883_TARGA_8ch_DIG] = { .mixers = { alc883_base_mixer, alc883_chmode_mixer }, .init_verbs = { alc883_init_verbs, alc880_gpio3_init_verbs, - alc883_tagra_verbs }, + alc883_targa_verbs }, .num_dacs = ARRAY_SIZE(alc883_dac_nids), .dac_nids = alc883_dac_nids, .num_adc_nids = ARRAY_SIZE(alc883_adc_nids_rev), @@ -9207,8 +9207,8 @@ static struct alc_config_preset alc883_presets[] = { .channel_mode = alc883_4ST_8ch_modes, .need_dac_fix = 1, .input_mux = &alc883_capture_source, - .unsol_event = alc883_tagra_unsol_event, - .init_hook = alc883_tagra_init_hook, + .unsol_event = alc883_targa_unsol_event, + .init_hook = alc883_targa_init_hook, }, [ALC883_ACER] = { .mixers = { alc883_base_mixer }, @@ -9362,7 +9362,7 @@ static struct alc_config_preset alc883_presets[] = { .init_hook = alc888_lenovo_ms7195_front_automute, }, [ALC883_HAIER_W66] = { - .mixers = { alc883_tagra_2ch_mixer}, + .mixers = { alc883_targa_2ch_mixer}, .init_verbs = { alc883_init_verbs, alc883_haier_w66_verbs}, .num_dacs = ARRAY_SIZE(alc883_dac_nids), .dac_nids = alc883_dac_nids, -- cgit v0.10.2 From def319f9e937f7a6a29718d3e2826c6c32f33245 Mon Sep 17 00:00:00 2001 From: Sasha Alexandr Date: Tue, 16 Jun 2009 16:00:15 -0400 Subject: ALSA: HDA - Correct trivial typos in comments. Correct some trivial typos in comments. Signed-off-by: Sasha Alexandr Signed-off-by: Takashi Iwai diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index dc77d75..8cebe26 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -970,7 +970,7 @@ static void alc_automute_pin(struct hda_codec *codec) } } -#if 0 /* it's broken in some acses -- temporarily disabled */ +#if 0 /* it's broken in some cases -- temporarily disabled */ static void alc_mic_automute(struct hda_codec *codec) { struct alc_spec *spec = codec->spec; @@ -1170,7 +1170,7 @@ static int alc_subsystem_id(struct hda_codec *codec, /* invalid SSID, check the special NID pin defcfg instead */ /* - * 31~30 : port conetcivity + * 31~30 : port connectivity * 29~21 : reserve * 20 : PCBEEP input * 19~16 : Check sum (15:1) @@ -6347,7 +6347,7 @@ static struct hda_channel_mode alc882_sixstack_modes[2] = { }; /* - * macbook pro ALC885 can switch LineIn to LineOut without loosing Mic + * macbook pro ALC885 can switch LineIn to LineOut without losing Mic */ /* @@ -7047,7 +7047,7 @@ static struct hda_verb alc882_auto_init_verbs[] = { #define alc882_loopbacks alc880_loopbacks #endif -/* pcm configuration: identiacal with ALC880 */ +/* pcm configuration: identical with ALC880 */ #define alc882_pcm_analog_playback alc880_pcm_analog_playback #define alc882_pcm_analog_capture alc880_pcm_analog_capture #define alc882_pcm_digital_playback alc880_pcm_digital_playback @@ -8957,7 +8957,7 @@ static void alc889A_mb31_unsol_event(struct hda_codec *codec, unsigned int res) #define alc883_loopbacks alc880_loopbacks #endif -/* pcm configuration: identiacal with ALC880 */ +/* pcm configuration: identical with ALC880 */ #define alc883_pcm_analog_playback alc880_pcm_analog_playback #define alc883_pcm_analog_capture alc880_pcm_analog_capture #define alc883_pcm_analog_alt_capture alc880_pcm_analog_alt_capture @@ -11132,7 +11132,7 @@ static struct hda_verb alc262_toshiba_rx1_unsol_verbs[] = { #define alc262_loopbacks alc880_loopbacks #endif -/* pcm configuration: identiacal with ALC880 */ +/* pcm configuration: identical with ALC880 */ #define alc262_pcm_analog_playback alc880_pcm_analog_playback #define alc262_pcm_analog_capture alc880_pcm_analog_capture #define alc262_pcm_digital_playback alc880_pcm_digital_playback @@ -12287,7 +12287,7 @@ static void alc268_auto_init_mono_speaker_out(struct hda_codec *codec) AC_VERB_SET_AMP_GAIN_MUTE, dac_vol2); } -/* pcm configuration: identiacal with ALC880 */ +/* pcm configuration: identical with ALC880 */ #define alc268_pcm_analog_playback alc880_pcm_analog_playback #define alc268_pcm_analog_capture alc880_pcm_analog_capture #define alc268_pcm_analog_alt_capture alc880_pcm_analog_alt_capture @@ -13198,7 +13198,7 @@ static int alc269_auto_create_analog_input_ctls(struct alc_spec *spec, #define alc269_loopbacks alc880_loopbacks #endif -/* pcm configuration: identiacal with ALC880 */ +/* pcm configuration: identical with ALC880 */ #define alc269_pcm_analog_playback alc880_pcm_analog_playback #define alc269_pcm_analog_capture alc880_pcm_analog_capture #define alc269_pcm_digital_playback alc880_pcm_digital_playback @@ -14060,7 +14060,7 @@ static void alc861_toshiba_unsol_event(struct hda_codec *codec, alc861_toshiba_automute(codec); } -/* pcm configuration: identiacal with ALC880 */ +/* pcm configuration: identical with ALC880 */ #define alc861_pcm_analog_playback alc880_pcm_analog_playback #define alc861_pcm_analog_capture alc880_pcm_analog_capture #define alc861_pcm_digital_playback alc880_pcm_digital_playback @@ -14583,7 +14583,7 @@ static hda_nid_t alc861vd_dac_nids[4] = { /* dac_nids for ALC660vd are in a different order - according to * Realtek's driver. - * This should probably tesult in a different mixer for 6stack models + * This should probably result in a different mixer for 6stack models * of ALC660vd codecs, but for now there is only 3stack mixer * - and it is the same as in 861vd. * adc_nids in ALC660vd are (is) the same as in 861vd @@ -15028,7 +15028,7 @@ static void alc861vd_dallas_init_hook(struct hda_codec *codec) #define alc861vd_loopbacks alc880_loopbacks #endif -/* pcm configuration: identiacal with ALC880 */ +/* pcm configuration: identical with ALC880 */ #define alc861vd_pcm_analog_playback alc880_pcm_analog_playback #define alc861vd_pcm_analog_capture alc880_pcm_analog_capture #define alc861vd_pcm_digital_playback alc880_pcm_digital_playback @@ -15207,7 +15207,7 @@ static void alc861vd_auto_init_hp_out(struct hda_codec *codec) hda_nid_t pin; pin = spec->autocfg.hp_pins[0]; - if (pin) /* connect to front and use dac 0 */ + if (pin) /* connect to front and use dac 0 */ alc861vd_auto_set_output_and_unmute(codec, pin, PIN_HP, 0); pin = spec->autocfg.speaker_pins[0]; if (pin) @@ -16670,7 +16670,7 @@ static struct snd_kcontrol_new alc272_nc10_mixer[] = { #endif -/* pcm configuration: identiacal with ALC880 */ +/* pcm configuration: identical with ALC880 */ #define alc662_pcm_analog_playback alc880_pcm_analog_playback #define alc662_pcm_analog_capture alc880_pcm_analog_capture #define alc662_pcm_digital_playback alc880_pcm_digital_playback -- cgit v0.10.2 From ef39412622b6e8f09c383de9565b07e93553fc27 Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Wed, 17 Jun 2009 11:42:45 +0200 Subject: ASoC: Kill BUS_ID_SIZE Remove the use of BUS_ID_SIZE from txx9aclc.c, as BUS_ID_SIZE will be removed soon later. Also, use snprintf() instead of sprintf() as a safer operation. Signed-off-by: Takashi Iwai diff --git a/sound/soc/txx9/txx9aclc.c b/sound/soc/txx9/txx9aclc.c index fa33661..938a58a 100644 --- a/sound/soc/txx9/txx9aclc.c +++ b/sound/soc/txx9/txx9aclc.c @@ -297,9 +297,9 @@ static int txx9aclc_pcm_new(struct snd_card *card, struct snd_soc_dai *dai, static bool filter(struct dma_chan *chan, void *param) { struct txx9aclc_dmadata *dmadata = param; - char devname[BUS_ID_SIZE + 2]; + char devname[20 + 2]; /* FIXME: old BUS_ID_SIZE + 2 */ - sprintf(devname, "%s.%d", dmadata->dma_res->name, + snprintf(devname, sizeof(devname), "%s.%d", dmadata->dma_res->name, (int)dmadata->dma_res->start); if (strcmp(dev_name(chan->device->dev), devname) == 0) { chan->private = &dmadata->dma_slave; -- cgit v0.10.2 From 0a842c8b60411e200b8a44b65dd78d9665692b91 Mon Sep 17 00:00:00 2001 From: Daniel Mack Date: Wed, 17 Jun 2009 17:45:11 +0200 Subject: ALSA: snd_usb_caiaq: fix legacy input streaming Seems that nobody recently tried the input on the very first supported sound card model, RK2. This patch fixes the byte offset to make it running again. Signed-off-by: Daniel Mack Signed-off-by: Takashi Iwai diff --git a/sound/usb/caiaq/audio.c b/sound/usb/caiaq/audio.c index b144513..8f9b60c 100644 --- a/sound/usb/caiaq/audio.c +++ b/sound/usb/caiaq/audio.c @@ -199,8 +199,9 @@ static int snd_usb_caiaq_pcm_prepare(struct snd_pcm_substream *substream) dev->period_out_count[index] = BYTES_PER_SAMPLE + 1; dev->audio_out_buf_pos[index] = BYTES_PER_SAMPLE + 1; } else { - dev->period_in_count[index] = BYTES_PER_SAMPLE; - dev->audio_in_buf_pos[index] = BYTES_PER_SAMPLE; + int in_pos = (dev->spec.data_alignment == 2) ? 0 : 2; + dev->period_in_count[index] = BYTES_PER_SAMPLE + in_pos; + dev->audio_in_buf_pos[index] = BYTES_PER_SAMPLE + in_pos; } if (dev->streaming) diff --git a/sound/usb/caiaq/device.c b/sound/usb/caiaq/device.c index 2240624..0e5db71 100644 --- a/sound/usb/caiaq/device.c +++ b/sound/usb/caiaq/device.c @@ -35,7 +35,7 @@ #include "input.h" MODULE_AUTHOR("Daniel Mack "); -MODULE_DESCRIPTION("caiaq USB audio, version 1.3.16"); +MODULE_DESCRIPTION("caiaq USB audio, version 1.3.17"); MODULE_LICENSE("GPL"); MODULE_SUPPORTED_DEVICE("{{Native Instruments, RigKontrol2}," "{Native Instruments, RigKontrol3}," -- cgit v0.10.2 From a76c1c23d0c33d98f2d9b36e76e7f71289fc8391 Mon Sep 17 00:00:00 2001 From: Chuck Ebbert Date: Thu, 18 Jun 2009 19:24:10 +0800 Subject: crypto: padlock-aes - work around Nano CPU errata in ECB mode The VIA Nano processor has a bug that makes it prefetch extra data during encryption operations, causing spurious page faults. Extend existing workarounds for ECB mode to copy the data to an temporary buffer to avoid the problem. Signed-off-by: Chuck Ebbert Acked-by: Harald Welte Signed-off-by: Herbert Xu diff --git a/drivers/crypto/padlock-aes.c b/drivers/crypto/padlock-aes.c index 87f92c3..e1d8776 100644 --- a/drivers/crypto/padlock-aes.c +++ b/drivers/crypto/padlock-aes.c @@ -18,9 +18,17 @@ #include #include #include +#include #include #include "padlock.h" +/* number of data blocks actually fetched for each xcrypt insn */ +static unsigned int ecb_fetch_blocks = 2; +static unsigned int cbc_fetch_blocks = 1; + +#define ecb_fetch_bytes (ecb_fetch_blocks * AES_BLOCK_SIZE) +#define cbc_fetch_bytes (cbc_fetch_blocks * AES_BLOCK_SIZE) + /* Control word. */ struct cword { unsigned int __attribute__ ((__packed__)) @@ -173,63 +181,59 @@ static inline void padlock_store_cword(struct cword *cword) */ static inline void padlock_xcrypt(const u8 *input, u8 *output, void *key, - struct cword *control_word) + struct cword *control_word, int count) { asm volatile (".byte 0xf3,0x0f,0xa7,0xc8" /* rep xcryptecb */ : "+S"(input), "+D"(output) - : "d"(control_word), "b"(key), "c"(1)); + : "d"(control_word), "b"(key), "c"(count)); } -static void aes_crypt_copy(const u8 *in, u8 *out, u32 *key, struct cword *cword) +static void aes_crypt_copy(const u8 *in, u8 *out, u32 *key, + struct cword *cword, int count) { - u8 buf[AES_BLOCK_SIZE * 2 + PADLOCK_ALIGNMENT - 1]; + /* + * Padlock prefetches extra data so we must provide mapped input buffers. + * Assume there are at least 16 bytes of stack already in use. + */ + u8 buf[AES_BLOCK_SIZE * 7 + PADLOCK_ALIGNMENT - 1]; u8 *tmp = PTR_ALIGN(&buf[0], PADLOCK_ALIGNMENT); - memcpy(tmp, in, AES_BLOCK_SIZE); - padlock_xcrypt(tmp, out, key, cword); + memcpy(tmp, in, count * AES_BLOCK_SIZE); + padlock_xcrypt(tmp, out, key, cword, count); } static inline void aes_crypt(const u8 *in, u8 *out, u32 *key, - struct cword *cword) + struct cword *cword, int count) { - /* padlock_xcrypt requires at least two blocks of data. */ - if (unlikely(!(((unsigned long)in ^ (PAGE_SIZE - AES_BLOCK_SIZE)) & - (PAGE_SIZE - 1)))) { - aes_crypt_copy(in, out, key, cword); + /* Padlock in ECB mode fetches at least ecb_fetch_bytes of data. + * We could avoid some copying here but it's probably not worth it. + */ + if (unlikely(((unsigned long)in & PAGE_SIZE) + ecb_fetch_bytes > PAGE_SIZE)) { + aes_crypt_copy(in, out, key, cword, count); return; } - padlock_xcrypt(in, out, key, cword); + padlock_xcrypt(in, out, key, cword, count); } static inline void padlock_xcrypt_ecb(const u8 *input, u8 *output, void *key, void *control_word, u32 count) { - if (count == 1) { - aes_crypt(input, output, key, control_word); + u32 initial = count & (ecb_fetch_blocks - 1); + + if (count < ecb_fetch_blocks) { + aes_crypt(input, output, key, control_word, count); return; } - asm volatile ("test $1, %%cl;" - "je 1f;" -#ifndef CONFIG_X86_64 - "lea -1(%%ecx), %%eax;" - "mov $1, %%ecx;" -#else - "lea -1(%%rcx), %%rax;" - "mov $1, %%rcx;" -#endif - ".byte 0xf3,0x0f,0xa7,0xc8;" /* rep xcryptecb */ -#ifndef CONFIG_X86_64 - "mov %%eax, %%ecx;" -#else - "mov %%rax, %%rcx;" -#endif - "1:" - ".byte 0xf3,0x0f,0xa7,0xc8" /* rep xcryptecb */ + if (initial) + asm volatile (".byte 0xf3,0x0f,0xa7,0xc8" /* rep xcryptecb */ + : "+S"(input), "+D"(output) + : "d"(control_word), "b"(key), "c"(initial)); + + asm volatile (".byte 0xf3,0x0f,0xa7,0xc8" /* rep xcryptecb */ : "+S"(input), "+D"(output) - : "d"(control_word), "b"(key), "c"(count) - : "ax"); + : "d"(control_word), "b"(key), "c"(count - initial)); } static inline u8 *padlock_xcrypt_cbc(const u8 *input, u8 *output, void *key, @@ -249,7 +253,7 @@ static void aes_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in) padlock_reset_key(&ctx->cword.encrypt); ts_state = irq_ts_save(); - aes_crypt(in, out, ctx->E, &ctx->cword.encrypt); + aes_crypt(in, out, ctx->E, &ctx->cword.encrypt, 1); irq_ts_restore(ts_state); padlock_store_cword(&ctx->cword.encrypt); } @@ -261,7 +265,7 @@ static void aes_decrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in) padlock_reset_key(&ctx->cword.encrypt); ts_state = irq_ts_save(); - aes_crypt(in, out, ctx->D, &ctx->cword.decrypt); + aes_crypt(in, out, ctx->D, &ctx->cword.decrypt, 1); irq_ts_restore(ts_state); padlock_store_cword(&ctx->cword.encrypt); } @@ -454,6 +458,7 @@ static struct crypto_alg cbc_aes_alg = { static int __init padlock_init(void) { int ret; + struct cpuinfo_x86 *c = &cpu_data(0); if (!cpu_has_xcrypt) { printk(KERN_NOTICE PFX "VIA PadLock not detected.\n"); @@ -476,6 +481,12 @@ static int __init padlock_init(void) printk(KERN_NOTICE PFX "Using VIA PadLock ACE for AES algorithm.\n"); + if (c->x86 == 6 && c->x86_model == 15 && c->x86_mask == 2) { + ecb_fetch_blocks = 8; + cbc_fetch_blocks = 4; /* NOTE: notused */ + printk(KERN_NOTICE PFX "VIA Nano stepping 2 detected: enabling workaround.\n"); + } + out: return ret; -- cgit v0.10.2 From 8d8409f773af2cfd52e23e4b138a7d55a31182cd Mon Sep 17 00:00:00 2001 From: Chuck Ebbert Date: Thu, 18 Jun 2009 19:31:09 +0800 Subject: crypto: padlock-aes - work around Nano CPU errata in CBC mode Extend previous workarounds for the prefetch bug to cover CBC mode, clean up the code a bit. Signed-off-by: Chuck Ebbert Acked-by: Harald Welte Signed-off-by: Herbert Xu diff --git a/drivers/crypto/padlock-aes.c b/drivers/crypto/padlock-aes.c index e1d8776..a9952b1 100644 --- a/drivers/crypto/padlock-aes.c +++ b/drivers/crypto/padlock-aes.c @@ -22,11 +22,16 @@ #include #include "padlock.h" -/* number of data blocks actually fetched for each xcrypt insn */ +/* + * Number of data blocks actually fetched for each xcrypt insn. + * Processors with prefetch errata will fetch extra blocks. + */ static unsigned int ecb_fetch_blocks = 2; -static unsigned int cbc_fetch_blocks = 1; - +#define MAX_ECB_FETCH_BLOCKS (8) #define ecb_fetch_bytes (ecb_fetch_blocks * AES_BLOCK_SIZE) + +static unsigned int cbc_fetch_blocks = 1; +#define MAX_CBC_FETCH_BLOCKS (4) #define cbc_fetch_bytes (cbc_fetch_blocks * AES_BLOCK_SIZE) /* Control word. */ @@ -180,7 +185,7 @@ static inline void padlock_store_cword(struct cword *cword) * should be used only inside the irq_ts_save/restore() context */ -static inline void padlock_xcrypt(const u8 *input, u8 *output, void *key, +static inline void rep_xcrypt_ecb(const u8 *input, u8 *output, void *key, struct cword *control_word, int count) { asm volatile (".byte 0xf3,0x0f,0xa7,0xc8" /* rep xcryptecb */ @@ -188,32 +193,65 @@ static inline void padlock_xcrypt(const u8 *input, u8 *output, void *key, : "d"(control_word), "b"(key), "c"(count)); } -static void aes_crypt_copy(const u8 *in, u8 *out, u32 *key, +static inline u8 *rep_xcrypt_cbc(const u8 *input, u8 *output, void *key, + u8 *iv, struct cword *control_word, int count) +{ + asm volatile (".byte 0xf3,0x0f,0xa7,0xd0" /* rep xcryptcbc */ + : "+S" (input), "+D" (output), "+a" (iv) + : "d" (control_word), "b" (key), "c" (count)); + return iv; +} + +static void ecb_crypt_copy(const u8 *in, u8 *out, u32 *key, struct cword *cword, int count) { /* * Padlock prefetches extra data so we must provide mapped input buffers. * Assume there are at least 16 bytes of stack already in use. */ - u8 buf[AES_BLOCK_SIZE * 7 + PADLOCK_ALIGNMENT - 1]; + u8 buf[AES_BLOCK_SIZE * (MAX_ECB_FETCH_BLOCKS - 1) + PADLOCK_ALIGNMENT - 1]; u8 *tmp = PTR_ALIGN(&buf[0], PADLOCK_ALIGNMENT); memcpy(tmp, in, count * AES_BLOCK_SIZE); - padlock_xcrypt(tmp, out, key, cword, count); + rep_xcrypt_ecb(tmp, out, key, cword, count); } -static inline void aes_crypt(const u8 *in, u8 *out, u32 *key, +static u8 *cbc_crypt_copy(const u8 *in, u8 *out, u32 *key, + u8 *iv, struct cword *cword, int count) +{ + /* + * Padlock prefetches extra data so we must provide mapped input buffers. + * Assume there are at least 16 bytes of stack already in use. + */ + u8 buf[AES_BLOCK_SIZE * (MAX_CBC_FETCH_BLOCKS - 1) + PADLOCK_ALIGNMENT - 1]; + u8 *tmp = PTR_ALIGN(&buf[0], PADLOCK_ALIGNMENT); + + memcpy(tmp, in, count * AES_BLOCK_SIZE); + return rep_xcrypt_cbc(tmp, out, key, iv, cword, count); +} + +static inline void ecb_crypt(const u8 *in, u8 *out, u32 *key, struct cword *cword, int count) { /* Padlock in ECB mode fetches at least ecb_fetch_bytes of data. * We could avoid some copying here but it's probably not worth it. */ if (unlikely(((unsigned long)in & PAGE_SIZE) + ecb_fetch_bytes > PAGE_SIZE)) { - aes_crypt_copy(in, out, key, cword, count); + ecb_crypt_copy(in, out, key, cword, count); return; } - padlock_xcrypt(in, out, key, cword, count); + rep_xcrypt_ecb(in, out, key, cword, count); +} + +static inline u8 *cbc_crypt(const u8 *in, u8 *out, u32 *key, + u8 *iv, struct cword *cword, int count) +{ + /* Padlock in CBC mode fetches at least cbc_fetch_bytes of data. */ + if (unlikely(((unsigned long)in & PAGE_SIZE) + cbc_fetch_bytes > PAGE_SIZE)) + return cbc_crypt_copy(in, out, key, iv, cword, count); + + return rep_xcrypt_cbc(in, out, key, iv, cword, count); } static inline void padlock_xcrypt_ecb(const u8 *input, u8 *output, void *key, @@ -222,7 +260,7 @@ static inline void padlock_xcrypt_ecb(const u8 *input, u8 *output, void *key, u32 initial = count & (ecb_fetch_blocks - 1); if (count < ecb_fetch_blocks) { - aes_crypt(input, output, key, control_word, count); + ecb_crypt(input, output, key, control_word, count); return; } @@ -239,10 +277,19 @@ static inline void padlock_xcrypt_ecb(const u8 *input, u8 *output, void *key, static inline u8 *padlock_xcrypt_cbc(const u8 *input, u8 *output, void *key, u8 *iv, void *control_word, u32 count) { - /* rep xcryptcbc */ - asm volatile (".byte 0xf3,0x0f,0xa7,0xd0" + u32 initial = count & (cbc_fetch_blocks - 1); + + if (count < cbc_fetch_blocks) + return cbc_crypt(input, output, key, iv, control_word, count); + + if (initial) + asm volatile (".byte 0xf3,0x0f,0xa7,0xd0" /* rep xcryptcbc */ + : "+S" (input), "+D" (output), "+a" (iv) + : "d" (control_word), "b" (key), "c" (count)); + + asm volatile (".byte 0xf3,0x0f,0xa7,0xd0" /* rep xcryptcbc */ : "+S" (input), "+D" (output), "+a" (iv) - : "d" (control_word), "b" (key), "c" (count)); + : "d" (control_word), "b" (key), "c" (count-initial)); return iv; } @@ -253,7 +300,7 @@ static void aes_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in) padlock_reset_key(&ctx->cword.encrypt); ts_state = irq_ts_save(); - aes_crypt(in, out, ctx->E, &ctx->cword.encrypt, 1); + ecb_crypt(in, out, ctx->E, &ctx->cword.encrypt, 1); irq_ts_restore(ts_state); padlock_store_cword(&ctx->cword.encrypt); } @@ -265,7 +312,7 @@ static void aes_decrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in) padlock_reset_key(&ctx->cword.encrypt); ts_state = irq_ts_save(); - aes_crypt(in, out, ctx->D, &ctx->cword.decrypt, 1); + ecb_crypt(in, out, ctx->D, &ctx->cword.decrypt, 1); irq_ts_restore(ts_state); padlock_store_cword(&ctx->cword.encrypt); } @@ -482,8 +529,8 @@ static int __init padlock_init(void) printk(KERN_NOTICE PFX "Using VIA PadLock ACE for AES algorithm.\n"); if (c->x86 == 6 && c->x86_model == 15 && c->x86_mask == 2) { - ecb_fetch_blocks = 8; - cbc_fetch_blocks = 4; /* NOTE: notused */ + ecb_fetch_blocks = MAX_ECB_FETCH_BLOCKS; + cbc_fetch_blocks = MAX_CBC_FETCH_BLOCKS; printk(KERN_NOTICE PFX "VIA Nano stepping 2 detected: enabling workaround.\n"); } -- cgit v0.10.2 From e6efaa025384f86a18814a6b9f4e5d54484ab9ff Mon Sep 17 00:00:00 2001 From: Huang Ying Date: Thu, 18 Jun 2009 19:33:57 +0800 Subject: crypto: aes-ni - Fix cbc mode IV saving Original implementation of aesni_cbc_dec do not save IV if input length % 4 == 0. This will make decryption of next block failed. Signed-off-by: Huang Ying Signed-off-by: Herbert Xu diff --git a/arch/x86/crypto/aesni-intel_asm.S b/arch/x86/crypto/aesni-intel_asm.S index caba996..eb0566e 100644 --- a/arch/x86/crypto/aesni-intel_asm.S +++ b/arch/x86/crypto/aesni-intel_asm.S @@ -845,7 +845,7 @@ ENTRY(aesni_cbc_enc) */ ENTRY(aesni_cbc_dec) cmp $16, LEN - jb .Lcbc_dec_ret + jb .Lcbc_dec_just_ret mov 480(KEYP), KLEN add $240, KEYP movups (IVP), IV @@ -891,6 +891,7 @@ ENTRY(aesni_cbc_dec) add $16, OUTP cmp $16, LEN jge .Lcbc_dec_loop1 - movups IV, (IVP) .Lcbc_dec_ret: + movups IV, (IVP) +.Lcbc_dec_just_ret: ret -- cgit v0.10.2 From 9251b64fb2d2326d28f0e0646a9e4fb8bbb51d8e Mon Sep 17 00:00:00 2001 From: Huang Ying Date: Thu, 18 Jun 2009 19:41:27 +0800 Subject: crypto: aes-ni - Do not sleep when using the FPU Because AES-NI instructions will touch XMM state, corresponding code must be enclosed within kernel_fpu_begin/end, which used preempt_disable/enable. So sleep should be prevented between kernel_fpu_begin/end. Signed-off-by: Huang Ying Signed-off-by: Herbert Xu diff --git a/arch/x86/crypto/aesni-intel_glue.c b/arch/x86/crypto/aesni-intel_glue.c index 4e66339..c580c5e 100644 --- a/arch/x86/crypto/aesni-intel_glue.c +++ b/arch/x86/crypto/aesni-intel_glue.c @@ -198,6 +198,7 @@ static int ecb_encrypt(struct blkcipher_desc *desc, blkcipher_walk_init(&walk, dst, src, nbytes); err = blkcipher_walk_virt(desc, &walk); + desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; kernel_fpu_begin(); while ((nbytes = walk.nbytes)) { @@ -221,6 +222,7 @@ static int ecb_decrypt(struct blkcipher_desc *desc, blkcipher_walk_init(&walk, dst, src, nbytes); err = blkcipher_walk_virt(desc, &walk); + desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; kernel_fpu_begin(); while ((nbytes = walk.nbytes)) { @@ -266,6 +268,7 @@ static int cbc_encrypt(struct blkcipher_desc *desc, blkcipher_walk_init(&walk, dst, src, nbytes); err = blkcipher_walk_virt(desc, &walk); + desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; kernel_fpu_begin(); while ((nbytes = walk.nbytes)) { @@ -289,6 +292,7 @@ static int cbc_decrypt(struct blkcipher_desc *desc, blkcipher_walk_init(&walk, dst, src, nbytes); err = blkcipher_walk_virt(desc, &walk); + desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; kernel_fpu_begin(); while ((nbytes = walk.nbytes)) { -- cgit v0.10.2 From b6f34d44cb341ad32f08717d1a2c418e6053a031 Mon Sep 17 00:00:00 2001 From: Huang Ying Date: Thu, 18 Jun 2009 19:44:01 +0800 Subject: crypto: aes-ni - Remove CRYPTO_TFM_REQ_MAY_SLEEP from fpu template kernel_fpu_begin/end used preempt_disable/enable, so sleep should be prevented between kernel_fpu_begin/end. Signed-off-by: Huang Ying Signed-off-by: Herbert Xu diff --git a/arch/x86/crypto/fpu.c b/arch/x86/crypto/fpu.c index 5f9781a..daef6cd 100644 --- a/arch/x86/crypto/fpu.c +++ b/arch/x86/crypto/fpu.c @@ -48,7 +48,7 @@ static int crypto_fpu_encrypt(struct blkcipher_desc *desc_in, struct blkcipher_desc desc = { .tfm = child, .info = desc_in->info, - .flags = desc_in->flags, + .flags = desc_in->flags & ~CRYPTO_TFM_REQ_MAY_SLEEP, }; kernel_fpu_begin(); @@ -67,7 +67,7 @@ static int crypto_fpu_decrypt(struct blkcipher_desc *desc_in, struct blkcipher_desc desc = { .tfm = child, .info = desc_in->info, - .flags = desc_in->flags, + .flags = desc_in->flags & ~CRYPTO_TFM_REQ_MAY_SLEEP, }; kernel_fpu_begin(); -- cgit v0.10.2 From 00540e5d54be972a94a3b2ce6da8621bebe731a2 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 12 Jun 2009 10:04:01 +0200 Subject: lockdep: Select frame pointers on x86 x86 stack traces are a piece of crap without frame pointers, and its not like the 'performance gain' of not having stack pointers matters when you selected lockdep. Reported-by: Andrew Morton LKML-Reference: Cc: Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 6cdcf38..3be4b7c 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -440,7 +440,7 @@ config LOCKDEP bool depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT select STACKTRACE - select FRAME_POINTER if !X86 && !MIPS && !PPC && !ARM_UNWIND && !S390 + select FRAME_POINTER if !MIPS && !PPC && !ARM_UNWIND && !S390 select KALLSYMS select KALLSYMS_ALL -- cgit v0.10.2 From d2fd4b09c07ae0c5ac288c0da6100c26ba9db15b Mon Sep 17 00:00:00 2001 From: Tony Vroon Date: Sun, 21 Jun 2009 00:40:10 +0100 Subject: ALSA: hda - Acer Inspire 6530G model for Realtek ALC888 The selected 4930G model seemed to keep the subwoofer 'tuba' function from operating correctly. Removing the existing PCI ID match made this work again, but it was mapped to 'Side' instead of to LFE as one would expect. This attempts to enable all functionality and keep the amount of available mixer sliders low. Any slider that had no audible effect on the output audio has been removed, and as such EAPD is not currently enabled. Signed-off-by: Tony Vroon Signed-off-by: Takashi Iwai diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index 8cebe26..63cfebb 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -224,6 +224,7 @@ enum { ALC883_ACER, ALC883_ACER_ASPIRE, ALC888_ACER_ASPIRE_4930G, + ALC888_ACER_ASPIRE_6530G, ALC888_ACER_ASPIRE_8930G, ALC883_MEDION, ALC883_MEDION_MD2, @@ -1471,6 +1472,25 @@ static struct hda_verb alc888_acer_aspire_4930g_verbs[] = { }; /* + * ALC888 Acer Aspire 6530G model + */ + +static struct hda_verb alc888_acer_aspire_6530g_verbs[] = { +/* Bias voltage on for external mic port */ + {0x18, AC_VERB_SET_PIN_WIDGET_CONTROL, PIN_IN | PIN_VREF80}, +/* Enable unsolicited event for HP jack */ + {0x15, AC_VERB_SET_UNSOLICITED_ENABLE, ALC880_HP_EVENT | AC_USRSP_EN}, +/* Enable speaker output */ + {0x14, AC_VERB_SET_PIN_WIDGET_CONTROL, PIN_OUT}, + {0x14, AC_VERB_SET_AMP_GAIN_MUTE, AMP_OUT_UNMUTE}, +/* Enable headphone output */ + {0x15, AC_VERB_SET_PIN_WIDGET_CONTROL, PIN_OUT | PIN_HP}, + {0x15, AC_VERB_SET_AMP_GAIN_MUTE, AMP_OUT_UNMUTE}, + {0x15, AC_VERB_SET_CONNECT_SEL, 0x00}, + { } +}; + +/* * ALC889 Acer Aspire 8930G model */ @@ -1544,6 +1564,25 @@ static struct hda_input_mux alc888_2_capture_sources[2] = { } }; +static struct hda_input_mux alc888_acer_aspire_6530_sources[2] = { + /* Interal mic only available on one ADC */ + { + .num_items = 3, + .items = { + { "Ext Mic", 0x0 }, + { "CD", 0x4 }, + { "Int Mic", 0xb }, + }, + }, + { + .num_items = 2, + .items = { + { "Ext Mic", 0x0 }, + { "CD", 0x4 }, + }, + } +}; + static struct hda_input_mux alc889_capture_sources[3] = { /* Digital mic only available on first "ADC" */ { @@ -8153,6 +8192,19 @@ static struct snd_kcontrol_new alc883_acer_aspire_mixer[] = { { } /* end */ }; +static struct snd_kcontrol_new alc888_acer_aspire_6530_mixer[] = { + HDA_CODEC_VOLUME("Front Playback Volume", 0x0c, 0x0, HDA_OUTPUT), + HDA_BIND_MUTE("Front Playback Switch", 0x0c, 2, HDA_INPUT), + HDA_CODEC_VOLUME("LFE Playback Volume", 0x0f, 0x0, HDA_OUTPUT), + HDA_BIND_MUTE("LFE Playback Switch", 0x0f, 2, HDA_INPUT), + HDA_CODEC_VOLUME("CD Playback Volume", 0x0b, 0x04, HDA_INPUT), + HDA_CODEC_MUTE("CD Playback Switch", 0x0b, 0x04, HDA_INPUT), + HDA_CODEC_VOLUME("Mic Playback Volume", 0x0b, 0x0, HDA_INPUT), + HDA_CODEC_VOLUME("Mic Boost", 0x18, 0, HDA_INPUT), + HDA_CODEC_MUTE("Mic Playback Switch", 0x0b, 0x0, HDA_INPUT), + { } /* end */ +}; + static struct snd_kcontrol_new alc888_lenovo_sky_mixer[] = { HDA_CODEC_VOLUME("Front Playback Volume", 0x0c, 0x0, HDA_OUTPUT), HDA_BIND_MUTE("Front Playback Switch", 0x0c, 2, HDA_INPUT), @@ -9021,7 +9073,7 @@ static struct snd_pci_quirk alc883_cfg_tbl[] = { SND_PCI_QUIRK(0x1025, 0x015e, "Acer Aspire 6930G", ALC888_ACER_ASPIRE_4930G), SND_PCI_QUIRK(0x1025, 0x0166, "Acer Aspire 6530G", - ALC888_ACER_ASPIRE_4930G), + ALC888_ACER_ASPIRE_6530G), /* default Acer -- disabled as it causes more problems. * model=auto should work fine now */ @@ -9256,6 +9308,24 @@ static struct alc_config_preset alc883_presets[] = { .unsol_event = alc_automute_amp_unsol_event, .init_hook = alc888_acer_aspire_4930g_init_hook, }, + [ALC888_ACER_ASPIRE_6530G] = { + .mixers = { alc888_acer_aspire_6530_mixer }, + .init_verbs = { alc883_init_verbs, alc880_gpio1_init_verbs, + alc888_acer_aspire_6530g_verbs }, + .num_dacs = ARRAY_SIZE(alc883_dac_nids), + .dac_nids = alc883_dac_nids, + .num_adc_nids = ARRAY_SIZE(alc883_adc_nids_rev), + .adc_nids = alc883_adc_nids_rev, + .capsrc_nids = alc883_capsrc_nids_rev, + .dig_out_nid = ALC883_DIGOUT_NID, + .num_channel_mode = ARRAY_SIZE(alc883_3ST_2ch_modes), + .channel_mode = alc883_3ST_2ch_modes, + .num_mux_defs = + ARRAY_SIZE(alc888_2_capture_sources), + .input_mux = alc888_acer_aspire_6530_sources, + .unsol_event = alc_automute_amp_unsol_event, + .init_hook = alc888_acer_aspire_4930g_init_hook, + }, [ALC888_ACER_ASPIRE_8930G] = { .mixers = { alc888_base_mixer, alc883_chmode_mixer }, -- cgit v0.10.2 From b1a914690c581f8f88b897d83a79b1c6eaf494c9 Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Sun, 21 Jun 2009 10:56:44 +0200 Subject: ALSA: hda - Add model=6530g option Add the new model string corresponding to the previous Acer Aspire 6530G support. Signed-off-by: Takashi Iwai diff --git a/Documentation/sound/alsa/HD-Audio-Models.txt b/Documentation/sound/alsa/HD-Audio-Models.txt index de8e10a..0d8d235 100644 --- a/Documentation/sound/alsa/HD-Audio-Models.txt +++ b/Documentation/sound/alsa/HD-Audio-Models.txt @@ -139,6 +139,7 @@ ALC883/888 acer Acer laptops (Travelmate 3012WTMi, Aspire 5600, etc) acer-aspire Acer Aspire 9810 acer-aspire-4930g Acer Aspire 4930G + acer-aspire-6530g Acer Aspire 6530G acer-aspire-8930g Acer Aspire 8930G medion Medion Laptops medion-md2 Medion MD2 diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index 63cfebb..bf4b78a 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -9030,6 +9030,7 @@ static const char *alc883_models[ALC883_MODEL_LAST] = { [ALC883_ACER] = "acer", [ALC883_ACER_ASPIRE] = "acer-aspire", [ALC888_ACER_ASPIRE_4930G] = "acer-aspire-4930g", + [ALC888_ACER_ASPIRE_6530G] = "acer-aspire-6530g", [ALC888_ACER_ASPIRE_8930G] = "acer-aspire-8930g", [ALC883_MEDION] = "medion", [ALC883_MEDION_MD2] = "medion-md2", -- cgit v0.10.2 From 0d1bb41ad4ebca92fafbab6d6c60438d7efef386 Mon Sep 17 00:00:00 2001 From: Ben Dooks Date: Sun, 14 Jun 2009 13:52:37 +0100 Subject: sdhci-s3c: Samsung S3C based SDHCI controller glue Add support for the 'HSMMC' block(s) in the Samsung SoC line. These are compatible with the SDHCI driver so add the necessary setup and driver binding for the platform devices. Signed-off-by: Ben Dooks Signed-off-by: Pierre Ossman diff --git a/MAINTAINERS b/MAINTAINERS index dc226e7..ba7ee2b 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -5094,6 +5094,13 @@ L: sdhci-devel@lists.ossman.eu S: Maintained F: drivers/mmc/host/sdhci.* +SECURE DIGITAL HOST CONTROLLER INTERFACE (SDHCI) SAMSUNG DRIVER +P: Ben Dooks +M: ben-linux@fluff.org +L: sdhci-devel@lists.ossman.eu +S: Maintained +F: drivers/mmc/host/sdhci-s3c.c + SECURITY SUBSYSTEM P: James Morris M: jmorris@namei.org diff --git a/drivers/mmc/host/Kconfig b/drivers/mmc/host/Kconfig index 40111a6..7609922 100644 --- a/drivers/mmc/host/Kconfig +++ b/drivers/mmc/host/Kconfig @@ -94,6 +94,31 @@ config MMC_SDHCI_PLTFM If unsure, say N. +config MMC_SDHCI_S3C + tristate "SDHCI support on Samsung S3C SoC" + depends on MMC_SDHCI && (PLAT_S3C24XX || PLAT_S3C64XX) + help + This selects the Secure Digital Host Controller Interface (SDHCI) + often referrered to as the HSMMC block in some of the Samsung S3C + range of SoC. + + Note, due to the problems with DMA, the DMA support is only + available with CONFIG_EXPERIMENTAL is selected. + + If you have a controller with this interface, say Y or M here. + + If unsure, say N. + +config MMC_SDHCI_S3C_DMA + bool "DMA support on S3C SDHCI" + depends on MMC_SDHCI_S3C && EXPERIMENTAL + help + Enable DMA support on the Samsung S3C SDHCI glue. The DMA + has proved to be problematic if the controller encounters + certain errors, and thus should be treated with care. + + YMMV. + config MMC_OMAP tristate "TI OMAP Multimedia Card Interface support" depends on ARCH_OMAP diff --git a/drivers/mmc/host/Makefile b/drivers/mmc/host/Makefile index 79da397..e637f2f 100644 --- a/drivers/mmc/host/Makefile +++ b/drivers/mmc/host/Makefile @@ -15,6 +15,7 @@ obj-$(CONFIG_MMC_SDHCI_PCI) += sdhci-pci.o obj-$(CONFIG_MMC_RICOH_MMC) += ricoh_mmc.o obj-$(CONFIG_MMC_SDHCI_OF) += sdhci-of.o obj-$(CONFIG_MMC_SDHCI_PLTFM) += sdhci-pltfm.o +obj-$(CONFIG_MMC_SDHCI_S3C) += sdhci-s3c.o obj-$(CONFIG_MMC_WBSD) += wbsd.o obj-$(CONFIG_MMC_AU1X) += au1xmmc.o obj-$(CONFIG_MMC_OMAP) += omap.o diff --git a/drivers/mmc/host/sdhci-s3c.c b/drivers/mmc/host/sdhci-s3c.c new file mode 100644 index 0000000..19246fe --- /dev/null +++ b/drivers/mmc/host/sdhci-s3c.c @@ -0,0 +1,425 @@ +/* linux/drivers/mmc/host/sdhci-s3c.c + * + * Copyright 2008 Openmoko Inc. + * Copyright 2008 Simtec Electronics + * Ben Dooks + * http://armlinux.simtec.co.uk/ + * + * SDHCI (HSMMC) support for Samsung SoC + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include +#include +#include +#include +#include + +#include + +#include +#include + +#include "sdhci.h" + +#define MAX_BUS_CLK (4) + +/** + * struct sdhci_s3c - S3C SDHCI instance + * @host: The SDHCI host created + * @pdev: The platform device we where created from. + * @ioarea: The resource created when we claimed the IO area. + * @pdata: The platform data for this controller. + * @cur_clk: The index of the current bus clock. + * @clk_io: The clock for the internal bus interface. + * @clk_bus: The clocks that are available for the SD/MMC bus clock. + */ +struct sdhci_s3c { + struct sdhci_host *host; + struct platform_device *pdev; + struct resource *ioarea; + struct s3c_sdhci_platdata *pdata; + unsigned int cur_clk; + + struct clk *clk_io; + struct clk *clk_bus[MAX_BUS_CLK]; +}; + +static inline struct sdhci_s3c *to_s3c(struct sdhci_host *host) +{ + return sdhci_priv(host); +} + +/** + * get_curclk - convert ctrl2 register to clock source number + * @ctrl2: Control2 register value. + */ +static u32 get_curclk(u32 ctrl2) +{ + ctrl2 &= S3C_SDHCI_CTRL2_SELBASECLK_MASK; + ctrl2 >>= S3C_SDHCI_CTRL2_SELBASECLK_SHIFT; + + return ctrl2; +} + +static void sdhci_s3c_check_sclk(struct sdhci_host *host) +{ + struct sdhci_s3c *ourhost = to_s3c(host); + u32 tmp = readl(host->ioaddr + S3C_SDHCI_CONTROL2); + + if (get_curclk(tmp) != ourhost->cur_clk) { + dev_dbg(&ourhost->pdev->dev, "restored ctrl2 clock setting\n"); + + tmp &= ~S3C_SDHCI_CTRL2_SELBASECLK_MASK; + tmp |= ourhost->cur_clk << S3C_SDHCI_CTRL2_SELBASECLK_SHIFT; + writel(tmp, host->ioaddr + 0x80); + } +} + +/** + * sdhci_s3c_get_max_clk - callback to get maximum clock frequency. + * @host: The SDHCI host instance. + * + * Callback to return the maximum clock rate acheivable by the controller. +*/ +static unsigned int sdhci_s3c_get_max_clk(struct sdhci_host *host) +{ + struct sdhci_s3c *ourhost = to_s3c(host); + struct clk *busclk; + unsigned int rate, max; + int clk; + + /* note, a reset will reset the clock source */ + + sdhci_s3c_check_sclk(host); + + for (max = 0, clk = 0; clk < MAX_BUS_CLK; clk++) { + busclk = ourhost->clk_bus[clk]; + if (!busclk) + continue; + + rate = clk_get_rate(busclk); + if (rate > max) + max = rate; + } + + return max; +} + +static unsigned int sdhci_s3c_get_timeout_clk(struct sdhci_host *host) +{ + return sdhci_s3c_get_max_clk(host) / 1000000; +} + +/** + * sdhci_s3c_consider_clock - consider one the bus clocks for current setting + * @ourhost: Our SDHCI instance. + * @src: The source clock index. + * @wanted: The clock frequency wanted. + */ +static unsigned int sdhci_s3c_consider_clock(struct sdhci_s3c *ourhost, + unsigned int src, + unsigned int wanted) +{ + unsigned long rate; + struct clk *clksrc = ourhost->clk_bus[src]; + int div; + + if (!clksrc) + return UINT_MAX; + + rate = clk_get_rate(clksrc); + + for (div = 1; div < 256; div *= 2) { + if ((rate / div) <= wanted) + break; + } + + dev_dbg(&ourhost->pdev->dev, "clk %d: rate %ld, want %d, got %ld\n", + src, rate, wanted, rate / div); + + return (wanted - (rate / div)); +} + +/** + * sdhci_s3c_set_clock - callback on clock change + * @host: The SDHCI host being changed + * @clock: The clock rate being requested. + * + * When the card's clock is going to be changed, look at the new frequency + * and find the best clock source to go with it. +*/ +static void sdhci_s3c_set_clock(struct sdhci_host *host, unsigned int clock) +{ + struct sdhci_s3c *ourhost = to_s3c(host); + unsigned int best = UINT_MAX; + unsigned int delta; + int best_src = 0; + int src; + u32 ctrl; + + /* don't bother if the clock is going off. */ + if (clock == 0) + return; + + for (src = 0; src < MAX_BUS_CLK; src++) { + delta = sdhci_s3c_consider_clock(ourhost, src, clock); + if (delta < best) { + best = delta; + best_src = src; + } + } + + dev_dbg(&ourhost->pdev->dev, + "selected source %d, clock %d, delta %d\n", + best_src, clock, best); + + /* select the new clock source */ + + if (ourhost->cur_clk != best_src) { + struct clk *clk = ourhost->clk_bus[best_src]; + + /* turn clock off to card before changing clock source */ + writew(0, host->ioaddr + SDHCI_CLOCK_CONTROL); + + ourhost->cur_clk = best_src; + host->max_clk = clk_get_rate(clk); + host->timeout_clk = sdhci_s3c_get_timeout_clk(host); + + ctrl = readl(host->ioaddr + S3C_SDHCI_CONTROL2); + ctrl &= ~S3C_SDHCI_CTRL2_SELBASECLK_MASK; + ctrl |= best_src << S3C_SDHCI_CTRL2_SELBASECLK_SHIFT; + writel(ctrl, host->ioaddr + S3C_SDHCI_CONTROL2); + } + + /* reconfigure the hardware for new clock rate */ + + { + struct mmc_ios ios; + + ios.clock = clock; + + if (ourhost->pdata->cfg_card) + (ourhost->pdata->cfg_card)(ourhost->pdev, host->ioaddr, + &ios, NULL); + } +} + +static struct sdhci_ops sdhci_s3c_ops = { + .get_max_clock = sdhci_s3c_get_max_clk, + .get_timeout_clock = sdhci_s3c_get_timeout_clk, + .set_clock = sdhci_s3c_set_clock, +}; + +static int __devinit sdhci_s3c_probe(struct platform_device *pdev) +{ + struct s3c_sdhci_platdata *pdata = pdev->dev.platform_data; + struct device *dev = &pdev->dev; + struct sdhci_host *host; + struct sdhci_s3c *sc; + struct resource *res; + int ret, irq, ptr, clks; + + if (!pdata) { + dev_err(dev, "no device data specified\n"); + return -ENOENT; + } + + irq = platform_get_irq(pdev, 0); + if (irq < 0) { + dev_err(dev, "no irq specified\n"); + return irq; + } + + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (!res) { + dev_err(dev, "no memory specified\n"); + return -ENOENT; + } + + host = sdhci_alloc_host(dev, sizeof(struct sdhci_s3c)); + if (IS_ERR(host)) { + dev_err(dev, "sdhci_alloc_host() failed\n"); + return PTR_ERR(host); + } + + sc = sdhci_priv(host); + + sc->host = host; + sc->pdev = pdev; + sc->pdata = pdata; + + platform_set_drvdata(pdev, host); + + sc->clk_io = clk_get(dev, "hsmmc"); + if (IS_ERR(sc->clk_io)) { + dev_err(dev, "failed to get io clock\n"); + ret = PTR_ERR(sc->clk_io); + goto err_io_clk; + } + + /* enable the local io clock and keep it running for the moment. */ + clk_enable(sc->clk_io); + + for (clks = 0, ptr = 0; ptr < MAX_BUS_CLK; ptr++) { + struct clk *clk; + char *name = pdata->clocks[ptr]; + + if (name == NULL) + continue; + + clk = clk_get(dev, name); + if (IS_ERR(clk)) { + dev_err(dev, "failed to get clock %s\n", name); + continue; + } + + clks++; + sc->clk_bus[ptr] = clk; + clk_enable(clk); + + dev_info(dev, "clock source %d: %s (%ld Hz)\n", + ptr, name, clk_get_rate(clk)); + } + + if (clks == 0) { + dev_err(dev, "failed to find any bus clocks\n"); + ret = -ENOENT; + goto err_no_busclks; + } + + sc->ioarea = request_mem_region(res->start, resource_size(res), + mmc_hostname(host->mmc)); + if (!sc->ioarea) { + dev_err(dev, "failed to reserve register area\n"); + ret = -ENXIO; + goto err_req_regs; + } + + host->ioaddr = ioremap_nocache(res->start, resource_size(res)); + if (!host->ioaddr) { + dev_err(dev, "failed to map registers\n"); + ret = -ENXIO; + goto err_req_regs; + } + + /* Ensure we have minimal gpio selected CMD/CLK/Detect */ + if (pdata->cfg_gpio) + pdata->cfg_gpio(pdev, pdata->max_width); + + host->hw_name = "samsung-hsmmc"; + host->ops = &sdhci_s3c_ops; + host->quirks = 0; + host->irq = irq; + + /* Setup quirks for the controller */ + + /* Currently with ADMA enabled we are getting some length + * interrupts that are not being dealt with, do disable + * ADMA until this is sorted out. */ + host->quirks |= SDHCI_QUIRK_BROKEN_ADMA; + host->quirks |= SDHCI_QUIRK_32BIT_ADMA_SIZE; + +#ifndef CONFIG_MMC_SDHCI_S3C_DMA + + /* we currently see overruns on errors, so disable the SDMA + * support as well. */ + host->quirks |= SDHCI_QUIRK_BROKEN_DMA; + +#endif /* CONFIG_MMC_SDHCI_S3C_DMA */ + + /* It seems we do not get an DATA transfer complete on non-busy + * transfers, not sure if this is a problem with this specific + * SDHCI block, or a missing configuration that needs to be set. */ + host->quirks |= SDHCI_QUIRK_NO_BUSY_IRQ; + + host->quirks |= (SDHCI_QUIRK_32BIT_DMA_ADDR | + SDHCI_QUIRK_32BIT_DMA_SIZE); + + ret = sdhci_add_host(host); + if (ret) { + dev_err(dev, "sdhci_add_host() failed\n"); + goto err_add_host; + } + + return 0; + + err_add_host: + release_resource(sc->ioarea); + kfree(sc->ioarea); + + err_req_regs: + for (ptr = 0; ptr < MAX_BUS_CLK; ptr++) { + clk_disable(sc->clk_bus[ptr]); + clk_put(sc->clk_bus[ptr]); + } + + err_no_busclks: + clk_disable(sc->clk_io); + clk_put(sc->clk_io); + + err_io_clk: + sdhci_free_host(host); + + return ret; +} + +static int __devexit sdhci_s3c_remove(struct platform_device *pdev) +{ + return 0; +} + +#ifdef CONFIG_PM + +static int sdhci_s3c_suspend(struct platform_device *dev, pm_message_t pm) +{ + struct sdhci_host *host = platform_get_drvdata(dev); + + sdhci_suspend_host(host, pm); + return 0; +} + +static int sdhci_s3c_resume(struct platform_device *dev) +{ + struct sdhci_host *host = platform_get_drvdata(dev); + + sdhci_resume_host(host); + return 0; +} + +#else +#define sdhci_s3c_suspend NULL +#define sdhci_s3c_resume NULL +#endif + +static struct platform_driver sdhci_s3c_driver = { + .probe = sdhci_s3c_probe, + .remove = __devexit_p(sdhci_s3c_remove), + .suspend = sdhci_s3c_suspend, + .resume = sdhci_s3c_resume, + .driver = { + .owner = THIS_MODULE, + .name = "s3c-sdhci", + }, +}; + +static int __init sdhci_s3c_init(void) +{ + return platform_driver_register(&sdhci_s3c_driver); +} + +static void __exit sdhci_s3c_exit(void) +{ + platform_driver_unregister(&sdhci_s3c_driver); +} + +module_init(sdhci_s3c_init); +module_exit(sdhci_s3c_exit); + +MODULE_DESCRIPTION("Samsung SDHCI (HSMMC) glue"); +MODULE_AUTHOR("Ben Dooks, "); +MODULE_LICENSE("GPL v2"); +MODULE_ALIAS("platform:s3c-sdhci"); -- cgit v0.10.2 From 6882a8c071d609f4c088bee41e79572c7cfc1790 Mon Sep 17 00:00:00 2001 From: Ben Dooks Date: Sun, 14 Jun 2009 13:52:38 +0100 Subject: sdhci: Add better ADMA error reporting Update the ADMA error reporting to not only show the overall controller state but also to print the ADMA descriptor list. Signed-off-by: Ben Dooks Signed-off-by: Pierre Ossman diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index 35789c6..e54d6fb 100644 --- a/drivers/mmc/host/sdhci.c +++ b/drivers/mmc/host/sdhci.c @@ -1382,6 +1382,35 @@ static void sdhci_cmd_irq(struct sdhci_host *host, u32 intmask) sdhci_finish_command(host); } +#ifdef DEBUG +static void sdhci_show_adma_error(struct sdhci_host *host) +{ + const char *name = mmc_hostname(host->mmc); + u8 *desc = host->adma_desc; + __le32 *dma; + __le16 *len; + u8 attr; + + sdhci_dumpregs(host); + + while (true) { + dma = (__le32 *)(desc + 4); + len = (__le16 *)(desc + 2); + attr = *desc; + + DBG("%s: %p: DMA 0x%08x, LEN 0x%04x, Attr=0x%02x\n", + name, desc, le32_to_cpu(*dma), le16_to_cpu(*len), attr); + + desc += 8; + + if (attr & 2) + break; + } +} +#else +static void sdhci_show_adma_error(struct sdhci_host *host) { } +#endif + static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) { BUG_ON(intmask == 0); @@ -1411,8 +1440,11 @@ static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) host->data->error = -ETIMEDOUT; else if (intmask & (SDHCI_INT_DATA_CRC | SDHCI_INT_DATA_END_BIT)) host->data->error = -EILSEQ; - else if (intmask & SDHCI_INT_ADMA_ERROR) + else if (intmask & SDHCI_INT_ADMA_ERROR) { + printk(KERN_ERR "%s: ADMA error\n", mmc_hostname(host->mmc)); + sdhci_show_adma_error(host); host->data->error = -EIO; + } if (host->data->error) sdhci_finish_data(host); -- cgit v0.10.2 From 1388eefd5a5e6aaa3cb04070bfc2b944c1d24b82 Mon Sep 17 00:00:00 2001 From: Ben Dooks Date: Sun, 14 Jun 2009 12:40:53 +0100 Subject: sdhci: Add SDHCI_QUIRK_NO_MULTIBLOCK quirk Add quirk to show the controller cannot do multi-block IO. This is mainly for the Samsung SDHCI controller that currently cannot manage to do multi-block PIO without timing out. Signed-off-by: Ben Dooks Signed-off-by: Pierre Ossman diff --git a/drivers/mmc/host/sdhci-s3c.c b/drivers/mmc/host/sdhci-s3c.c index 19246fe..50997d2 100644 --- a/drivers/mmc/host/sdhci-s3c.c +++ b/drivers/mmc/host/sdhci-s3c.c @@ -329,6 +329,9 @@ static int __devinit sdhci_s3c_probe(struct platform_device *pdev) * support as well. */ host->quirks |= SDHCI_QUIRK_BROKEN_DMA; + /* PIO currently has problems with multi-block IO */ + host->quirks |= SDHCI_QUIRK_NO_MULTIBLOCK; + #endif /* CONFIG_MMC_SDHCI_S3C_DMA */ /* It seems we do not get an DATA transfer complete on non-busy diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index e54d6fb..3856669 100644 --- a/drivers/mmc/host/sdhci.c +++ b/drivers/mmc/host/sdhci.c @@ -1834,7 +1834,7 @@ int sdhci_add_host(struct sdhci_host *host) /* * Maximum block count. */ - mmc->max_blk_count = 65535; + mmc->max_blk_count = (host->quirks & SDHCI_QUIRK_NO_MULTIBLOCK) ? 1 : 65535; /* * Init tasklets. diff --git a/drivers/mmc/host/sdhci.h b/drivers/mmc/host/sdhci.h index 2de0834..84d1490 100644 --- a/drivers/mmc/host/sdhci.h +++ b/drivers/mmc/host/sdhci.h @@ -226,6 +226,8 @@ struct sdhci_host { #define SDHCI_QUIRK_RESTORE_IRQS_AFTER_RESET (1<<19) /* Controller has to be forced to use block size of 2048 bytes */ #define SDHCI_QUIRK_FORCE_BLK_SZ_2048 (1<<20) +/* Controller cannot do multi-block transfers */ +#define SDHCI_QUIRK_NO_MULTIBLOCK (1<<21) int irq; /* Device IRQ */ void __iomem * ioaddr; /* Mapped address */ -- cgit v0.10.2 From 04ac2f46d6ecb995f78c9ae4e2e4707d00b5339f Mon Sep 17 00:00:00 2001 From: Nicolas Ferre Date: Tue, 16 Jun 2009 13:05:50 +0200 Subject: MAINTAINERS: add myself as atmel-mci maintainer (sd/mmc interface) Add MAINTAINERS entry for atmel-mci driver. This driver was maintained by its author: Haavard Skinnemoen. I take the maintainance of it. Signed-off-by: Nicolas Ferre Acked-by: Haavard Skinnemoen Signed-off-by: Pierre Ossman diff --git a/MAINTAINERS b/MAINTAINERS index ba7ee2b..5b526ce 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1010,6 +1010,13 @@ W: http://www.at91.com/ S: Maintained F: drivers/mmc/host/at91_mci.c +ATMEL AT91 / AT32 MCI DRIVER +P: Nicolas Ferre +M: nicolas.ferre@atmel.com +S: Maintained +F: drivers/mmc/host/atmel-mci.c +F: drivers/mmc/host/atmel-mci-regs.h + ATMEL AT91 / AT32 SERIAL DRIVER P: Haavard Skinnemoen M: hskinnemoen@atmel.com -- cgit v0.10.2 From 5fe23c7f51def59f66bc6aeee988ef1a467a2c8c Mon Sep 17 00:00:00 2001 From: Anton Vorontsov Date: Thu, 18 Jun 2009 00:14:08 +0400 Subject: sdhci: Add support for hosts that are only capable of 1-bit transfers Some hosts (hardware configurations, or particular SD/MMC slots) may not support 4-bit bus. For example, on MPC8569E-MDS boards we can switch between serial (1-bit only) and nibble (4-bit) modes, thought we have to disable more peripherals to work in 4-bit mode. Along with some small core changes, this patch modifies sdhci-of driver, so that now it looks for "sdhci,1-bit-only" property in the device-tree, and if specified we enable a proper quirk. Signed-off-by: Anton Vorontsov Acked-by: Grant Likely Signed-off-by: Pierre Ossman diff --git a/Documentation/powerpc/dts-bindings/fsl/esdhc.txt b/Documentation/powerpc/dts-bindings/fsl/esdhc.txt index 5093ddf..3ed3797 100644 --- a/Documentation/powerpc/dts-bindings/fsl/esdhc.txt +++ b/Documentation/powerpc/dts-bindings/fsl/esdhc.txt @@ -10,6 +10,8 @@ Required properties: - interrupts : should contain eSDHC interrupt. - interrupt-parent : interrupt source phandle. - clock-frequency : specifies eSDHC base clock frequency. + - sdhci,1-bit-only : (optional) specifies that a controller can + only handle 1-bit data transfers. Example: diff --git a/drivers/mmc/host/sdhci-of.c b/drivers/mmc/host/sdhci-of.c index 128c614..d79fa55 100644 --- a/drivers/mmc/host/sdhci-of.c +++ b/drivers/mmc/host/sdhci-of.c @@ -250,6 +250,9 @@ static int __devinit sdhci_of_probe(struct of_device *ofdev, host->ops = &sdhci_of_data->ops; } + if (of_get_property(np, "sdhci,1-bit-only", NULL)) + host->quirks |= SDHCI_QUIRK_FORCE_1_BIT_DATA; + clk = of_get_property(np, "clock-frequency", &size); if (clk && size == sizeof(*clk) && *clk) of_host->clock = *clk; diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index 3856669..c758673 100644 --- a/drivers/mmc/host/sdhci.c +++ b/drivers/mmc/host/sdhci.c @@ -1761,7 +1761,10 @@ int sdhci_add_host(struct sdhci_host *host) mmc->ops = &sdhci_ops; mmc->f_min = host->max_clk / 256; mmc->f_max = host->max_clk; - mmc->caps = MMC_CAP_4_BIT_DATA | MMC_CAP_SDIO_IRQ; + mmc->caps = MMC_CAP_SDIO_IRQ; + + if (!(host->quirks & SDHCI_QUIRK_FORCE_1_BIT_DATA)) + mmc->caps |= MMC_CAP_4_BIT_DATA; if (caps & SDHCI_CAN_DO_HISPD) mmc->caps |= MMC_CAP_SD_HIGHSPEED; diff --git a/drivers/mmc/host/sdhci.h b/drivers/mmc/host/sdhci.h index 84d1490..5d37dd9 100644 --- a/drivers/mmc/host/sdhci.h +++ b/drivers/mmc/host/sdhci.h @@ -228,6 +228,8 @@ struct sdhci_host { #define SDHCI_QUIRK_FORCE_BLK_SZ_2048 (1<<20) /* Controller cannot do multi-block transfers */ #define SDHCI_QUIRK_NO_MULTIBLOCK (1<<21) +/* Controller can only handle 1-bit data transfers */ +#define SDHCI_QUIRK_FORCE_1_BIT_DATA (1<<22) int irq; /* Device IRQ */ void __iomem * ioaddr; /* Mapped address */ -- cgit v0.10.2 From f0bf7f61b8405224bc52fc9a3ccd167a68126e00 Mon Sep 17 00:00:00 2001 From: Harald Welte Date: Wed, 17 Jun 2009 20:22:39 +0200 Subject: mmc: Add new via-sdmmc host controller driver This adds the via-sdmmc driver for the SD/MMC-controller of VIA, which is found in a number of recent integrated VIA chipset products. Signed-off-by: Harald Welte Signed-off-by: Pierre Ossman diff --git a/MAINTAINERS b/MAINTAINERS index 5b526ce..1d47043 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -6230,6 +6230,14 @@ S: Maintained F: Documentation/i2c/busses/i2c-viapro F: drivers/i2c/busses/i2c-viapro.c +VIA SD/MMC CARD CONTROLLER DRIVER +P: Joseph Chan +M: JosephChan@via.com.tw +P: Harald Welte +M: HaraldWelte@viatech.com +S: Maintained +F: drivers/mmc/host/via-sdmmc.c + VIA UNICHROME(PRO)/CHROME9 FRAMEBUFFER DRIVER P: Joseph Chan M: JosephChan@via.com.tw diff --git a/drivers/mmc/host/Kconfig b/drivers/mmc/host/Kconfig index 7609922..891ef18 100644 --- a/drivers/mmc/host/Kconfig +++ b/drivers/mmc/host/Kconfig @@ -290,3 +290,14 @@ config MMC_CB710 This driver can also be built as a module. If so, the module will be called cb710-mmc. +config MMC_VIA_SDMMC + tristate "VIA SD/MMC Card Reader Driver" + depends on PCI + help + This selects the VIA SD/MMC Card Reader driver, say Y or M here. + VIA provides one multi-functional card reader which integrated into + some motherboards manufactured by VIA. This card reader supports + SD/MMC/SDHC. + If you have a controller with this interface, say Y or M here. + + If unsure, say N. diff --git a/drivers/mmc/host/Makefile b/drivers/mmc/host/Makefile index e637f2f..cf153f6 100644 --- a/drivers/mmc/host/Makefile +++ b/drivers/mmc/host/Makefile @@ -32,6 +32,7 @@ obj-$(CONFIG_MMC_S3C) += s3cmci.o obj-$(CONFIG_MMC_SDRICOH_CS) += sdricoh_cs.o obj-$(CONFIG_MMC_TMIO) += tmio_mmc.o obj-$(CONFIG_MMC_CB710) += cb710-mmc.o +obj-$(CONFIG_MMC_VIA_SDMMC) += via-sdmmc.o ifeq ($(CONFIG_CB710_DEBUG),y) CFLAGS-cb710-mmc += -DDEBUG diff --git a/drivers/mmc/host/via-sdmmc.c b/drivers/mmc/host/via-sdmmc.c new file mode 100644 index 0000000..632858a --- /dev/null +++ b/drivers/mmc/host/via-sdmmc.c @@ -0,0 +1,1362 @@ +/* + * drivers/mmc/host/via-sdmmc.c - VIA SD/MMC Card Reader driver + * Copyright (c) 2008, VIA Technologies Inc. All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or (at + * your option) any later version. + */ + +#include +#include +#include +#include + +#include + +#define DRV_NAME "via_sdmmc" + +#define PCI_DEVICE_ID_VIA_9530 0x9530 + +#define VIA_CRDR_SDC_OFF 0x200 +#define VIA_CRDR_DDMA_OFF 0x400 +#define VIA_CRDR_PCICTRL_OFF 0x600 + +#define VIA_CRDR_MIN_CLOCK 375000 +#define VIA_CRDR_MAX_CLOCK 48000000 + +/* + * PCI registers + */ + +#define VIA_CRDR_PCI_WORK_MODE 0x40 +#define VIA_CRDR_PCI_DBG_MODE 0x41 + +/* + * SDC MMIO Registers + */ + +#define VIA_CRDR_SDCTRL 0x0 +#define VIA_CRDR_SDCTRL_START 0x01 +#define VIA_CRDR_SDCTRL_WRITE 0x04 +#define VIA_CRDR_SDCTRL_SINGLE_WR 0x10 +#define VIA_CRDR_SDCTRL_SINGLE_RD 0x20 +#define VIA_CRDR_SDCTRL_MULTI_WR 0x30 +#define VIA_CRDR_SDCTRL_MULTI_RD 0x40 +#define VIA_CRDR_SDCTRL_STOP 0x70 + +#define VIA_CRDR_SDCTRL_RSP_NONE 0x0 +#define VIA_CRDR_SDCTRL_RSP_R1 0x10000 +#define VIA_CRDR_SDCTRL_RSP_R2 0x20000 +#define VIA_CRDR_SDCTRL_RSP_R3 0x30000 +#define VIA_CRDR_SDCTRL_RSP_R1B 0x90000 + +#define VIA_CRDR_SDCARG 0x4 + +#define VIA_CRDR_SDBUSMODE 0x8 +#define VIA_CRDR_SDMODE_4BIT 0x02 +#define VIA_CRDR_SDMODE_CLK_ON 0x40 + +#define VIA_CRDR_SDBLKLEN 0xc +/* + * Bit 0 -Bit 10 : Block length. So, the maximum block length should be 2048. + * Bit 11 - Bit 13 : Reserved. + * GPIDET : Select GPI pin to detect card, GPI means CR_CD# in top design. + * INTEN : Enable SD host interrupt. + * Bit 16 - Bit 31 : Block count. So, the maximun block count should be 65536. + */ +#define VIA_CRDR_SDBLKLEN_GPIDET 0x2000 +#define VIA_CRDR_SDBLKLEN_INTEN 0x8000 +#define VIA_CRDR_MAX_BLOCK_COUNT 65536 +#define VIA_CRDR_MAX_BLOCK_LENGTH 2048 + +#define VIA_CRDR_SDRESP0 0x10 +#define VIA_CRDR_SDRESP1 0x14 +#define VIA_CRDR_SDRESP2 0x18 +#define VIA_CRDR_SDRESP3 0x1c + +#define VIA_CRDR_SDCURBLKCNT 0x20 + +#define VIA_CRDR_SDINTMASK 0x24 +/* + * MBDIE : Multiple Blocks transfer Done Interrupt Enable + * BDDIE : Block Data transfer Done Interrupt Enable + * CIRIE : Card Insertion or Removal Interrupt Enable + * CRDIE : Command-Response transfer Done Interrupt Enable + * CRTOIE : Command-Response response TimeOut Interrupt Enable + * ASCRDIE : Auto Stop Command-Response transfer Done Interrupt Enable + * DTIE : Data access Timeout Interrupt Enable + * SCIE : reSponse CRC error Interrupt Enable + * RCIE : Read data CRC error Interrupt Enable + * WCIE : Write data CRC error Interrupt Enable + */ +#define VIA_CRDR_SDINTMASK_MBDIE 0x10 +#define VIA_CRDR_SDINTMASK_BDDIE 0x20 +#define VIA_CRDR_SDINTMASK_CIRIE 0x80 +#define VIA_CRDR_SDINTMASK_CRDIE 0x200 +#define VIA_CRDR_SDINTMASK_CRTOIE 0x400 +#define VIA_CRDR_SDINTMASK_ASCRDIE 0x800 +#define VIA_CRDR_SDINTMASK_DTIE 0x1000 +#define VIA_CRDR_SDINTMASK_SCIE 0x2000 +#define VIA_CRDR_SDINTMASK_RCIE 0x4000 +#define VIA_CRDR_SDINTMASK_WCIE 0x8000 + +#define VIA_CRDR_SDACTIVE_INTMASK \ + (VIA_CRDR_SDINTMASK_MBDIE | VIA_CRDR_SDINTMASK_CIRIE \ + | VIA_CRDR_SDINTMASK_CRDIE | VIA_CRDR_SDINTMASK_CRTOIE \ + | VIA_CRDR_SDINTMASK_DTIE | VIA_CRDR_SDINTMASK_SCIE \ + | VIA_CRDR_SDINTMASK_RCIE | VIA_CRDR_SDINTMASK_WCIE) + +#define VIA_CRDR_SDSTATUS 0x28 +/* + * CECC : Reserved + * WP : SD card Write Protect status + * SLOTD : Reserved + * SLOTG : SD SLOT status(Gpi pin status) + * MBD : Multiple Blocks transfer Done interrupt status + * BDD : Block Data transfer Done interrupt status + * CD : Reserved + * CIR : Card Insertion or Removal interrupt detected on GPI pin + * IO : Reserved + * CRD : Command-Response transfer Done interrupt status + * CRTO : Command-Response response TimeOut interrupt status + * ASCRDIE : Auto Stop Command-Response transfer Done interrupt status + * DT : Data access Timeout interrupt status + * SC : reSponse CRC error interrupt status + * RC : Read data CRC error interrupt status + * WC : Write data CRC error interrupt status + */ +#define VIA_CRDR_SDSTS_CECC 0x01 +#define VIA_CRDR_SDSTS_WP 0x02 +#define VIA_CRDR_SDSTS_SLOTD 0x04 +#define VIA_CRDR_SDSTS_SLOTG 0x08 +#define VIA_CRDR_SDSTS_MBD 0x10 +#define VIA_CRDR_SDSTS_BDD 0x20 +#define VIA_CRDR_SDSTS_CD 0x40 +#define VIA_CRDR_SDSTS_CIR 0x80 +#define VIA_CRDR_SDSTS_IO 0x100 +#define VIA_CRDR_SDSTS_CRD 0x200 +#define VIA_CRDR_SDSTS_CRTO 0x400 +#define VIA_CRDR_SDSTS_ASCRDIE 0x800 +#define VIA_CRDR_SDSTS_DT 0x1000 +#define VIA_CRDR_SDSTS_SC 0x2000 +#define VIA_CRDR_SDSTS_RC 0x4000 +#define VIA_CRDR_SDSTS_WC 0x8000 + +#define VIA_CRDR_SDSTS_IGN_MASK\ + (VIA_CRDR_SDSTS_BDD | VIA_CRDR_SDSTS_ASCRDIE | VIA_CRDR_SDSTS_IO) +#define VIA_CRDR_SDSTS_INT_MASK \ + (VIA_CRDR_SDSTS_MBD | VIA_CRDR_SDSTS_BDD | VIA_CRDR_SDSTS_CD \ + | VIA_CRDR_SDSTS_CIR | VIA_CRDR_SDSTS_IO | VIA_CRDR_SDSTS_CRD \ + | VIA_CRDR_SDSTS_CRTO | VIA_CRDR_SDSTS_ASCRDIE | VIA_CRDR_SDSTS_DT \ + | VIA_CRDR_SDSTS_SC | VIA_CRDR_SDSTS_RC | VIA_CRDR_SDSTS_WC) +#define VIA_CRDR_SDSTS_W1C_MASK \ + (VIA_CRDR_SDSTS_CECC | VIA_CRDR_SDSTS_MBD | VIA_CRDR_SDSTS_BDD \ + | VIA_CRDR_SDSTS_CD | VIA_CRDR_SDSTS_CIR | VIA_CRDR_SDSTS_CRD \ + | VIA_CRDR_SDSTS_CRTO | VIA_CRDR_SDSTS_ASCRDIE | VIA_CRDR_SDSTS_DT \ + | VIA_CRDR_SDSTS_SC | VIA_CRDR_SDSTS_RC | VIA_CRDR_SDSTS_WC) +#define VIA_CRDR_SDSTS_CMD_MASK \ + (VIA_CRDR_SDSTS_CRD | VIA_CRDR_SDSTS_CRTO | VIA_CRDR_SDSTS_SC) +#define VIA_CRDR_SDSTS_DATA_MASK\ + (VIA_CRDR_SDSTS_MBD | VIA_CRDR_SDSTS_DT \ + | VIA_CRDR_SDSTS_RC | VIA_CRDR_SDSTS_WC) + +#define VIA_CRDR_SDSTATUS2 0x2a +/* + * CFE : Enable SD host automatic Clock FReezing + */ +#define VIA_CRDR_SDSTS_CFE 0x80 + +#define VIA_CRDR_SDRSPTMO 0x2C + +#define VIA_CRDR_SDCLKSEL 0x30 + +#define VIA_CRDR_SDEXTCTRL 0x34 +#define VIS_CRDR_SDEXTCTRL_AUTOSTOP_SD 0x01 +#define VIS_CRDR_SDEXTCTRL_SHIFT_9 0x02 +#define VIS_CRDR_SDEXTCTRL_MMC_8BIT 0x04 +#define VIS_CRDR_SDEXTCTRL_RELD_BLK 0x08 +#define VIS_CRDR_SDEXTCTRL_BAD_CMDA 0x10 +#define VIS_CRDR_SDEXTCTRL_BAD_DATA 0x20 +#define VIS_CRDR_SDEXTCTRL_AUTOSTOP_SPI 0x40 +#define VIA_CRDR_SDEXTCTRL_HISPD 0x80 +/* 0x38-0xFF reserved */ + +/* + * Data DMA Control Registers + */ + +#define VIA_CRDR_DMABASEADD 0x0 +#define VIA_CRDR_DMACOUNTER 0x4 + +#define VIA_CRDR_DMACTRL 0x8 +/* + * DIR :Transaction Direction + * 0 : From card to memory + * 1 : From memory to card + */ +#define VIA_CRDR_DMACTRL_DIR 0x100 +#define VIA_CRDR_DMACTRL_ENIRQ 0x10000 +#define VIA_CRDR_DMACTRL_SFTRST 0x1000000 + +#define VIA_CRDR_DMASTS 0xc + +#define VIA_CRDR_DMASTART 0x10 +/*0x14-0xFF reserved*/ + +/* + * PCI Control Registers + */ + +/*0x0 - 0x1 reserved*/ +#define VIA_CRDR_PCICLKGATT 0x2 +/* + * SFTRST : + * 0 : Soft reset all the controller and it will be de-asserted automatically + * 1 : Soft reset is de-asserted + */ +#define VIA_CRDR_PCICLKGATT_SFTRST 0x01 +/* + * 3V3 : Pad power select + * 0 : 1.8V + * 1 : 3.3V + * NOTE : No mater what the actual value should be, this bit always + * read as 0. This is a hardware bug. + */ +#define VIA_CRDR_PCICLKGATT_3V3 0x10 +/* + * PAD_PWRON : Pad Power on/off select + * 0 : Power off + * 1 : Power on + * NOTE : No mater what the actual value should be, this bit always + * read as 0. This is a hardware bug. + */ +#define VIA_CRDR_PCICLKGATT_PAD_PWRON 0x20 + +#define VIA_CRDR_PCISDCCLK 0x5 + +#define VIA_CRDR_PCIDMACLK 0x7 +#define VIA_CRDR_PCIDMACLK_SDC 0x2 + +#define VIA_CRDR_PCIINTCTRL 0x8 +#define VIA_CRDR_PCIINTCTRL_SDCIRQEN 0x04 + +#define VIA_CRDR_PCIINTSTATUS 0x9 +#define VIA_CRDR_PCIINTSTATUS_SDC 0x04 + +#define VIA_CRDR_PCITMOCTRL 0xa +#define VIA_CRDR_PCITMOCTRL_NO 0x0 +#define VIA_CRDR_PCITMOCTRL_32US 0x1 +#define VIA_CRDR_PCITMOCTRL_256US 0x2 +#define VIA_CRDR_PCITMOCTRL_1024US 0x3 +#define VIA_CRDR_PCITMOCTRL_256MS 0x4 +#define VIA_CRDR_PCITMOCTRL_512MS 0x5 +#define VIA_CRDR_PCITMOCTRL_1024MS 0x6 + +/*0xB-0xFF reserved*/ + +enum PCI_HOST_CLK_CONTROL { + PCI_CLK_375K = 0x03, + PCI_CLK_8M = 0x04, + PCI_CLK_12M = 0x00, + PCI_CLK_16M = 0x05, + PCI_CLK_24M = 0x01, + PCI_CLK_33M = 0x06, + PCI_CLK_48M = 0x02 +}; + +struct sdhcreg { + u32 sdcontrol_reg; + u32 sdcmdarg_reg; + u32 sdbusmode_reg; + u32 sdblklen_reg; + u32 sdresp_reg[4]; + u32 sdcurblkcnt_reg; + u32 sdintmask_reg; + u32 sdstatus_reg; + u32 sdrsptmo_reg; + u32 sdclksel_reg; + u32 sdextctrl_reg; +}; + +struct pcictrlreg { + u8 reserve[2]; + u8 pciclkgat_reg; + u8 pcinfcclk_reg; + u8 pcimscclk_reg; + u8 pcisdclk_reg; + u8 pcicaclk_reg; + u8 pcidmaclk_reg; + u8 pciintctrl_reg; + u8 pciintstatus_reg; + u8 pcitmoctrl_reg; + u8 Resv; +}; + +struct via_crdr_mmc_host { + struct mmc_host *mmc; + struct mmc_request *mrq; + struct mmc_command *cmd; + struct mmc_data *data; + + void __iomem *mmiobase; + void __iomem *sdhc_mmiobase; + void __iomem *ddma_mmiobase; + void __iomem *pcictrl_mmiobase; + + struct pcictrlreg pm_pcictrl_reg; + struct sdhcreg pm_sdhc_reg; + + struct work_struct carddet_work; + struct tasklet_struct finish_tasklet; + + struct timer_list timer; + spinlock_t lock; + u8 power; + int reject; + unsigned int quirks; +}; + +/* some devices need a very long delay for power to stabilize */ +#define VIA_CRDR_QUIRK_300MS_PWRDELAY 0x0001 + +static struct pci_device_id via_ids[] = { + {PCI_VENDOR_ID_VIA, PCI_DEVICE_ID_VIA_9530, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0,}, + {0,} +}; + +MODULE_DEVICE_TABLE(pci, via_ids); + +static void via_print_sdchc(struct via_crdr_mmc_host *host) +{ + void __iomem *addrbase = host->sdhc_mmiobase; + + pr_debug("SDC MMIO Registers:\n"); + pr_debug("SDCONTROL=%08x, SDCMDARG=%08x, SDBUSMODE=%08x\n", + readl(addrbase + VIA_CRDR_SDCTRL), + readl(addrbase + VIA_CRDR_SDCARG), + readl(addrbase + VIA_CRDR_SDBUSMODE)); + pr_debug("SDBLKLEN=%08x, SDCURBLKCNT=%08x, SDINTMASK=%08x\n", + readl(addrbase + VIA_CRDR_SDBLKLEN), + readl(addrbase + VIA_CRDR_SDCURBLKCNT), + readl(addrbase + VIA_CRDR_SDINTMASK)); + pr_debug("SDSTATUS=%08x, SDCLKSEL=%08x, SDEXTCTRL=%08x\n", + readl(addrbase + VIA_CRDR_SDSTATUS), + readl(addrbase + VIA_CRDR_SDCLKSEL), + readl(addrbase + VIA_CRDR_SDEXTCTRL)); +} + +static void via_print_pcictrl(struct via_crdr_mmc_host *host) +{ + void __iomem *addrbase = host->pcictrl_mmiobase; + + pr_debug("PCI Control Registers:\n"); + pr_debug("PCICLKGATT=%02x, PCISDCCLK=%02x, PCIDMACLK=%02x\n", + readb(addrbase + VIA_CRDR_PCICLKGATT), + readb(addrbase + VIA_CRDR_PCISDCCLK), + readb(addrbase + VIA_CRDR_PCIDMACLK)); + pr_debug("PCIINTCTRL=%02x, PCIINTSTATUS=%02x\n", + readb(addrbase + VIA_CRDR_PCIINTCTRL), + readb(addrbase + VIA_CRDR_PCIINTSTATUS)); +} + +static void via_save_pcictrlreg(struct via_crdr_mmc_host *host) +{ + struct pcictrlreg *pm_pcictrl_reg; + void __iomem *addrbase; + + pm_pcictrl_reg = &(host->pm_pcictrl_reg); + addrbase = host->pcictrl_mmiobase; + + pm_pcictrl_reg->pciclkgat_reg = readb(addrbase + VIA_CRDR_PCICLKGATT); + pm_pcictrl_reg->pciclkgat_reg |= + VIA_CRDR_PCICLKGATT_3V3 | VIA_CRDR_PCICLKGATT_PAD_PWRON; + pm_pcictrl_reg->pcisdclk_reg = readb(addrbase + VIA_CRDR_PCISDCCLK); + pm_pcictrl_reg->pcidmaclk_reg = readb(addrbase + VIA_CRDR_PCIDMACLK); + pm_pcictrl_reg->pciintctrl_reg = readb(addrbase + VIA_CRDR_PCIINTCTRL); + pm_pcictrl_reg->pciintstatus_reg = + readb(addrbase + VIA_CRDR_PCIINTSTATUS); + pm_pcictrl_reg->pcitmoctrl_reg = readb(addrbase + VIA_CRDR_PCITMOCTRL); +} + +static void via_restore_pcictrlreg(struct via_crdr_mmc_host *host) +{ + struct pcictrlreg *pm_pcictrl_reg; + void __iomem *addrbase; + + pm_pcictrl_reg = &(host->pm_pcictrl_reg); + addrbase = host->pcictrl_mmiobase; + + writeb(pm_pcictrl_reg->pciclkgat_reg, addrbase + VIA_CRDR_PCICLKGATT); + writeb(pm_pcictrl_reg->pcisdclk_reg, addrbase + VIA_CRDR_PCISDCCLK); + writeb(pm_pcictrl_reg->pcidmaclk_reg, addrbase + VIA_CRDR_PCIDMACLK); + writeb(pm_pcictrl_reg->pciintctrl_reg, addrbase + VIA_CRDR_PCIINTCTRL); + writeb(pm_pcictrl_reg->pciintstatus_reg, + addrbase + VIA_CRDR_PCIINTSTATUS); + writeb(pm_pcictrl_reg->pcitmoctrl_reg, addrbase + VIA_CRDR_PCITMOCTRL); +} + +static void via_save_sdcreg(struct via_crdr_mmc_host *host) +{ + struct sdhcreg *pm_sdhc_reg; + void __iomem *addrbase; + + pm_sdhc_reg = &(host->pm_sdhc_reg); + addrbase = host->sdhc_mmiobase; + + pm_sdhc_reg->sdcontrol_reg = readl(addrbase + VIA_CRDR_SDCTRL); + pm_sdhc_reg->sdcmdarg_reg = readl(addrbase + VIA_CRDR_SDCARG); + pm_sdhc_reg->sdbusmode_reg = readl(addrbase + VIA_CRDR_SDBUSMODE); + pm_sdhc_reg->sdblklen_reg = readl(addrbase + VIA_CRDR_SDBLKLEN); + pm_sdhc_reg->sdcurblkcnt_reg = readl(addrbase + VIA_CRDR_SDCURBLKCNT); + pm_sdhc_reg->sdintmask_reg = readl(addrbase + VIA_CRDR_SDINTMASK); + pm_sdhc_reg->sdstatus_reg = readl(addrbase + VIA_CRDR_SDSTATUS); + pm_sdhc_reg->sdrsptmo_reg = readl(addrbase + VIA_CRDR_SDRSPTMO); + pm_sdhc_reg->sdclksel_reg = readl(addrbase + VIA_CRDR_SDCLKSEL); + pm_sdhc_reg->sdextctrl_reg = readl(addrbase + VIA_CRDR_SDEXTCTRL); +} + +static void via_restore_sdcreg(struct via_crdr_mmc_host *host) +{ + struct sdhcreg *pm_sdhc_reg; + void __iomem *addrbase; + + pm_sdhc_reg = &(host->pm_sdhc_reg); + addrbase = host->sdhc_mmiobase; + + writel(pm_sdhc_reg->sdcontrol_reg, addrbase + VIA_CRDR_SDCTRL); + writel(pm_sdhc_reg->sdcmdarg_reg, addrbase + VIA_CRDR_SDCARG); + writel(pm_sdhc_reg->sdbusmode_reg, addrbase + VIA_CRDR_SDBUSMODE); + writel(pm_sdhc_reg->sdblklen_reg, addrbase + VIA_CRDR_SDBLKLEN); + writel(pm_sdhc_reg->sdcurblkcnt_reg, addrbase + VIA_CRDR_SDCURBLKCNT); + writel(pm_sdhc_reg->sdintmask_reg, addrbase + VIA_CRDR_SDINTMASK); + writel(pm_sdhc_reg->sdstatus_reg, addrbase + VIA_CRDR_SDSTATUS); + writel(pm_sdhc_reg->sdrsptmo_reg, addrbase + VIA_CRDR_SDRSPTMO); + writel(pm_sdhc_reg->sdclksel_reg, addrbase + VIA_CRDR_SDCLKSEL); + writel(pm_sdhc_reg->sdextctrl_reg, addrbase + VIA_CRDR_SDEXTCTRL); +} + +static void via_pwron_sleep(struct via_crdr_mmc_host *sdhost) +{ + if (sdhost->quirks & VIA_CRDR_QUIRK_300MS_PWRDELAY) + msleep(300); + else + msleep(3); +} + +static void via_set_ddma(struct via_crdr_mmc_host *host, + dma_addr_t dmaaddr, u32 count, int dir, int enirq) +{ + void __iomem *addrbase; + u32 ctrl_data = 0; + + if (enirq) + ctrl_data |= VIA_CRDR_DMACTRL_ENIRQ; + + if (dir) + ctrl_data |= VIA_CRDR_DMACTRL_DIR; + + addrbase = host->ddma_mmiobase; + + writel(dmaaddr, addrbase + VIA_CRDR_DMABASEADD); + writel(count, addrbase + VIA_CRDR_DMACOUNTER); + writel(ctrl_data, addrbase + VIA_CRDR_DMACTRL); + writel(0x01, addrbase + VIA_CRDR_DMASTART); + + /* It seems that our DMA can not work normally with 375kHz clock */ + /* FIXME: don't brute-force 8MHz but use PIO at 375kHz !! */ + addrbase = host->pcictrl_mmiobase; + if (readb(addrbase + VIA_CRDR_PCISDCCLK) == PCI_CLK_375K) { + dev_info(host->mmc->parent, "forcing card speed to 8MHz\n"); + writeb(PCI_CLK_8M, addrbase + VIA_CRDR_PCISDCCLK); + } +} + +static void via_sdc_preparedata(struct via_crdr_mmc_host *host, + struct mmc_data *data) +{ + void __iomem *addrbase; + u32 blk_reg; + int count; + + WARN_ON(host->data); + + /* Sanity checks */ + BUG_ON(data->blksz > host->mmc->max_blk_size); + BUG_ON(data->blocks > host->mmc->max_blk_count); + + host->data = data; + + count = dma_map_sg(mmc_dev(host->mmc), data->sg, data->sg_len, + ((data->flags & MMC_DATA_READ) ? + PCI_DMA_FROMDEVICE : PCI_DMA_TODEVICE)); + BUG_ON(count != 1); + + via_set_ddma(host, sg_dma_address(data->sg), sg_dma_len(data->sg), + (data->flags & MMC_DATA_WRITE) ? 1 : 0, 1); + + addrbase = host->sdhc_mmiobase; + + blk_reg = data->blksz - 1; + blk_reg |= VIA_CRDR_SDBLKLEN_GPIDET | VIA_CRDR_SDBLKLEN_INTEN; + blk_reg |= (data->blocks) << 16; + + writel(blk_reg, addrbase + VIA_CRDR_SDBLKLEN); +} + +static void via_sdc_get_response(struct via_crdr_mmc_host *host, + struct mmc_command *cmd) +{ + void __iomem *addrbase = host->sdhc_mmiobase; + u32 dwdata0 = readl(addrbase + VIA_CRDR_SDRESP0); + u32 dwdata1 = readl(addrbase + VIA_CRDR_SDRESP1); + u32 dwdata2 = readl(addrbase + VIA_CRDR_SDRESP2); + u32 dwdata3 = readl(addrbase + VIA_CRDR_SDRESP3); + + if (cmd->flags & MMC_RSP_136) { + cmd->resp[0] = ((u8) (dwdata1)) | + (((u8) (dwdata0 >> 24)) << 8) | + (((u8) (dwdata0 >> 16)) << 16) | + (((u8) (dwdata0 >> 8)) << 24); + + cmd->resp[1] = ((u8) (dwdata2)) | + (((u8) (dwdata1 >> 24)) << 8) | + (((u8) (dwdata1 >> 16)) << 16) | + (((u8) (dwdata1 >> 8)) << 24); + + cmd->resp[2] = ((u8) (dwdata3)) | + (((u8) (dwdata2 >> 24)) << 8) | + (((u8) (dwdata2 >> 16)) << 16) | + (((u8) (dwdata2 >> 8)) << 24); + + cmd->resp[3] = 0xff | + ((((u8) (dwdata3 >> 24))) << 8) | + (((u8) (dwdata3 >> 16)) << 16) | + (((u8) (dwdata3 >> 8)) << 24); + } else { + dwdata0 >>= 8; + cmd->resp[0] = ((dwdata0 & 0xff) << 24) | + (((dwdata0 >> 8) & 0xff) << 16) | + (((dwdata0 >> 16) & 0xff) << 8) | (dwdata1 & 0xff); + + dwdata1 >>= 8; + cmd->resp[1] = ((dwdata1 & 0xff) << 24) | + (((dwdata1 >> 8) & 0xff) << 16) | + (((dwdata1 >> 16) & 0xff) << 8); + } +} + +static void via_sdc_send_command(struct via_crdr_mmc_host *host, + struct mmc_command *cmd) +{ + void __iomem *addrbase; + struct mmc_data *data; + u32 cmdctrl = 0; + + WARN_ON(host->cmd); + + data = cmd->data; + mod_timer(&host->timer, jiffies + HZ); + host->cmd = cmd; + + /*Command index*/ + cmdctrl = cmd->opcode << 8; + + /*Response type*/ + switch (mmc_resp_type(cmd)) { + case MMC_RSP_NONE: + cmdctrl |= VIA_CRDR_SDCTRL_RSP_NONE; + break; + case MMC_RSP_R1: + cmdctrl |= VIA_CRDR_SDCTRL_RSP_R1; + break; + case MMC_RSP_R1B: + cmdctrl |= VIA_CRDR_SDCTRL_RSP_R1B; + break; + case MMC_RSP_R2: + cmdctrl |= VIA_CRDR_SDCTRL_RSP_R2; + break; + case MMC_RSP_R3: + cmdctrl |= VIA_CRDR_SDCTRL_RSP_R3; + break; + default: + pr_err("%s: cmd->flag is not valid\n", mmc_hostname(host->mmc)); + break; + } + + if (!(cmd->data)) + goto nodata; + + via_sdc_preparedata(host, data); + + /*Command control*/ + if (data->blocks > 1) { + if (data->flags & MMC_DATA_WRITE) { + cmdctrl |= VIA_CRDR_SDCTRL_WRITE; + cmdctrl |= VIA_CRDR_SDCTRL_MULTI_WR; + } else { + cmdctrl |= VIA_CRDR_SDCTRL_MULTI_RD; + } + } else { + if (data->flags & MMC_DATA_WRITE) { + cmdctrl |= VIA_CRDR_SDCTRL_WRITE; + cmdctrl |= VIA_CRDR_SDCTRL_SINGLE_WR; + } else { + cmdctrl |= VIA_CRDR_SDCTRL_SINGLE_RD; + } + } + +nodata: + if (cmd == host->mrq->stop) + cmdctrl |= VIA_CRDR_SDCTRL_STOP; + + cmdctrl |= VIA_CRDR_SDCTRL_START; + + addrbase = host->sdhc_mmiobase; + writel(cmd->arg, addrbase + VIA_CRDR_SDCARG); + writel(cmdctrl, addrbase + VIA_CRDR_SDCTRL); +} + +static void via_sdc_finish_data(struct via_crdr_mmc_host *host) +{ + struct mmc_data *data; + + BUG_ON(!host->data); + + data = host->data; + host->data = NULL; + + if (data->error) + data->bytes_xfered = 0; + else + data->bytes_xfered = data->blocks * data->blksz; + + dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len, + ((data->flags & MMC_DATA_READ) ? + PCI_DMA_FROMDEVICE : PCI_DMA_TODEVICE)); + + if (data->stop) + via_sdc_send_command(host, data->stop); + else + tasklet_schedule(&host->finish_tasklet); +} + +static void via_sdc_finish_command(struct via_crdr_mmc_host *host) +{ + via_sdc_get_response(host, host->cmd); + + host->cmd->error = 0; + + if (!host->cmd->data) + tasklet_schedule(&host->finish_tasklet); + + host->cmd = NULL; +} + +static void via_sdc_request(struct mmc_host *mmc, struct mmc_request *mrq) +{ + void __iomem *addrbase; + struct via_crdr_mmc_host *host; + unsigned long flags; + u16 status; + + host = mmc_priv(mmc); + + spin_lock_irqsave(&host->lock, flags); + + addrbase = host->pcictrl_mmiobase; + writeb(VIA_CRDR_PCIDMACLK_SDC, addrbase + VIA_CRDR_PCIDMACLK); + + status = readw(host->sdhc_mmiobase + VIA_CRDR_SDSTATUS); + status &= VIA_CRDR_SDSTS_W1C_MASK; + writew(status, host->sdhc_mmiobase + VIA_CRDR_SDSTATUS); + + WARN_ON(host->mrq != NULL); + host->mrq = mrq; + + status = readw(host->sdhc_mmiobase + VIA_CRDR_SDSTATUS); + if (!(status & VIA_CRDR_SDSTS_SLOTG) || host->reject) { + host->mrq->cmd->error = -ENOMEDIUM; + tasklet_schedule(&host->finish_tasklet); + } else { + via_sdc_send_command(host, mrq->cmd); + } + + mmiowb(); + spin_unlock_irqrestore(&host->lock, flags); +} + +static void via_sdc_set_power(struct via_crdr_mmc_host *host, + unsigned short power, unsigned int on) +{ + unsigned long flags; + u8 gatt; + + spin_lock_irqsave(&host->lock, flags); + + host->power = (1 << power); + + gatt = readb(host->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT); + if (host->power == MMC_VDD_165_195) + gatt &= ~VIA_CRDR_PCICLKGATT_3V3; + else + gatt |= VIA_CRDR_PCICLKGATT_3V3; + if (on) + gatt |= VIA_CRDR_PCICLKGATT_PAD_PWRON; + else + gatt &= ~VIA_CRDR_PCICLKGATT_PAD_PWRON; + writeb(gatt, host->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT); + + mmiowb(); + spin_unlock_irqrestore(&host->lock, flags); + + via_pwron_sleep(host); +} + +static void via_sdc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios) +{ + struct via_crdr_mmc_host *host; + unsigned long flags; + void __iomem *addrbase; + u32 org_data, sdextctrl; + u8 clock; + + host = mmc_priv(mmc); + + spin_lock_irqsave(&host->lock, flags); + + addrbase = host->sdhc_mmiobase; + org_data = readl(addrbase + VIA_CRDR_SDBUSMODE); + sdextctrl = readl(addrbase + VIA_CRDR_SDEXTCTRL); + + if (ios->bus_width == MMC_BUS_WIDTH_1) + org_data &= ~VIA_CRDR_SDMODE_4BIT; + else + org_data |= VIA_CRDR_SDMODE_4BIT; + + if (ios->power_mode == MMC_POWER_OFF) + org_data &= ~VIA_CRDR_SDMODE_CLK_ON; + else + org_data |= VIA_CRDR_SDMODE_CLK_ON; + + if (ios->timing == MMC_TIMING_SD_HS) + sdextctrl |= VIA_CRDR_SDEXTCTRL_HISPD; + else + sdextctrl &= ~VIA_CRDR_SDEXTCTRL_HISPD; + + writel(org_data, addrbase + VIA_CRDR_SDBUSMODE); + writel(sdextctrl, addrbase + VIA_CRDR_SDEXTCTRL); + + if (ios->clock >= 48000000) + clock = PCI_CLK_48M; + else if (ios->clock >= 33000000) + clock = PCI_CLK_33M; + else if (ios->clock >= 24000000) + clock = PCI_CLK_24M; + else if (ios->clock >= 16000000) + clock = PCI_CLK_16M; + else if (ios->clock >= 12000000) + clock = PCI_CLK_12M; + else if (ios->clock >= 8000000) + clock = PCI_CLK_8M; + else + clock = PCI_CLK_375K; + + addrbase = host->pcictrl_mmiobase; + if (readb(addrbase + VIA_CRDR_PCISDCCLK) != clock) + writeb(clock, addrbase + VIA_CRDR_PCISDCCLK); + + mmiowb(); + spin_unlock_irqrestore(&host->lock, flags); + + if (ios->power_mode != MMC_POWER_OFF) + via_sdc_set_power(host, ios->vdd, 1); + else + via_sdc_set_power(host, ios->vdd, 0); +} + +static int via_sdc_get_ro(struct mmc_host *mmc) +{ + struct via_crdr_mmc_host *host; + unsigned long flags; + u16 status; + + host = mmc_priv(mmc); + + spin_lock_irqsave(&host->lock, flags); + + status = readw(host->sdhc_mmiobase + VIA_CRDR_SDSTATUS); + + spin_unlock_irqrestore(&host->lock, flags); + + return !(status & VIA_CRDR_SDSTS_WP); +} + +static const struct mmc_host_ops via_sdc_ops = { + .request = via_sdc_request, + .set_ios = via_sdc_set_ios, + .get_ro = via_sdc_get_ro, +}; + +static void via_reset_pcictrl(struct via_crdr_mmc_host *host) +{ + void __iomem *addrbase; + unsigned long flags; + u8 gatt; + + addrbase = host->pcictrl_mmiobase; + + spin_lock_irqsave(&host->lock, flags); + + via_save_pcictrlreg(host); + via_save_sdcreg(host); + + spin_unlock_irqrestore(&host->lock, flags); + + gatt = VIA_CRDR_PCICLKGATT_PAD_PWRON; + if (host->power == MMC_VDD_165_195) + gatt &= VIA_CRDR_PCICLKGATT_3V3; + else + gatt |= VIA_CRDR_PCICLKGATT_3V3; + writeb(gatt, host->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT); + via_pwron_sleep(host); + gatt |= VIA_CRDR_PCICLKGATT_SFTRST; + writeb(gatt, host->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT); + msleep(3); + + spin_lock_irqsave(&host->lock, flags); + + via_restore_pcictrlreg(host); + via_restore_sdcreg(host); + + mmiowb(); + spin_unlock_irqrestore(&host->lock, flags); +} + +static void via_sdc_cmd_isr(struct via_crdr_mmc_host *host, u16 intmask) +{ + BUG_ON(intmask == 0); + + if (!host->cmd) { + pr_err("%s: Got command interrupt 0x%x even " + "though no command operation was in progress.\n", + mmc_hostname(host->mmc), intmask); + return; + } + + if (intmask & VIA_CRDR_SDSTS_CRTO) + host->cmd->error = -ETIMEDOUT; + else if (intmask & VIA_CRDR_SDSTS_SC) + host->cmd->error = -EILSEQ; + + if (host->cmd->error) + tasklet_schedule(&host->finish_tasklet); + else if (intmask & VIA_CRDR_SDSTS_CRD) + via_sdc_finish_command(host); +} + +static void via_sdc_data_isr(struct via_crdr_mmc_host *host, u16 intmask) +{ + BUG_ON(intmask == 0); + + if (intmask & VIA_CRDR_SDSTS_DT) + host->data->error = -ETIMEDOUT; + else if (intmask & (VIA_CRDR_SDSTS_RC | VIA_CRDR_SDSTS_WC)) + host->data->error = -EILSEQ; + + via_sdc_finish_data(host); +} + +static irqreturn_t via_sdc_isr(int irq, void *dev_id) +{ + struct via_crdr_mmc_host *sdhost = dev_id; + void __iomem *addrbase; + u8 pci_status; + u16 sd_status; + irqreturn_t result; + + if (!sdhost) + return IRQ_NONE; + + spin_lock(&sdhost->lock); + + addrbase = sdhost->pcictrl_mmiobase; + pci_status = readb(addrbase + VIA_CRDR_PCIINTSTATUS); + if (!(pci_status & VIA_CRDR_PCIINTSTATUS_SDC)) { + result = IRQ_NONE; + goto out; + } + + addrbase = sdhost->sdhc_mmiobase; + sd_status = readw(addrbase + VIA_CRDR_SDSTATUS); + sd_status &= VIA_CRDR_SDSTS_INT_MASK; + sd_status &= ~VIA_CRDR_SDSTS_IGN_MASK; + if (!sd_status) { + result = IRQ_NONE; + goto out; + } + + if (sd_status & VIA_CRDR_SDSTS_CIR) { + writew(sd_status & VIA_CRDR_SDSTS_CIR, + addrbase + VIA_CRDR_SDSTATUS); + + schedule_work(&sdhost->carddet_work); + } + + sd_status &= ~VIA_CRDR_SDSTS_CIR; + if (sd_status & VIA_CRDR_SDSTS_CMD_MASK) { + writew(sd_status & VIA_CRDR_SDSTS_CMD_MASK, + addrbase + VIA_CRDR_SDSTATUS); + via_sdc_cmd_isr(sdhost, sd_status & VIA_CRDR_SDSTS_CMD_MASK); + } + if (sd_status & VIA_CRDR_SDSTS_DATA_MASK) { + writew(sd_status & VIA_CRDR_SDSTS_DATA_MASK, + addrbase + VIA_CRDR_SDSTATUS); + via_sdc_data_isr(sdhost, sd_status & VIA_CRDR_SDSTS_DATA_MASK); + } + + sd_status &= ~(VIA_CRDR_SDSTS_CMD_MASK | VIA_CRDR_SDSTS_DATA_MASK); + if (sd_status) { + pr_err("%s: Unexpected interrupt 0x%x\n", + mmc_hostname(sdhost->mmc), sd_status); + writew(sd_status, addrbase + VIA_CRDR_SDSTATUS); + } + + result = IRQ_HANDLED; + + mmiowb(); +out: + spin_unlock(&sdhost->lock); + + return result; +} + +static void via_sdc_timeout(unsigned long ulongdata) +{ + struct via_crdr_mmc_host *sdhost; + unsigned long flags; + + sdhost = (struct via_crdr_mmc_host *)ulongdata; + + spin_lock_irqsave(&sdhost->lock, flags); + + if (sdhost->mrq) { + pr_err("%s: Timeout waiting for hardware interrupt." + "cmd:0x%x\n", mmc_hostname(sdhost->mmc), + sdhost->mrq->cmd->opcode); + + if (sdhost->data) { + writel(VIA_CRDR_DMACTRL_SFTRST, + sdhost->ddma_mmiobase + VIA_CRDR_DMACTRL); + sdhost->data->error = -ETIMEDOUT; + via_sdc_finish_data(sdhost); + } else { + if (sdhost->cmd) + sdhost->cmd->error = -ETIMEDOUT; + else + sdhost->mrq->cmd->error = -ETIMEDOUT; + tasklet_schedule(&sdhost->finish_tasklet); + } + } + + mmiowb(); + spin_unlock_irqrestore(&sdhost->lock, flags); +} + +static void via_sdc_tasklet_finish(unsigned long param) +{ + struct via_crdr_mmc_host *host; + unsigned long flags; + struct mmc_request *mrq; + + host = (struct via_crdr_mmc_host *)param; + + spin_lock_irqsave(&host->lock, flags); + + del_timer(&host->timer); + mrq = host->mrq; + host->mrq = NULL; + host->cmd = NULL; + host->data = NULL; + + spin_unlock_irqrestore(&host->lock, flags); + + mmc_request_done(host->mmc, mrq); +} + +static void via_sdc_card_detect(struct work_struct *work) +{ + struct via_crdr_mmc_host *host; + void __iomem *addrbase; + unsigned long flags; + u16 status; + + host = container_of(work, struct via_crdr_mmc_host, carddet_work); + + addrbase = host->ddma_mmiobase; + writel(VIA_CRDR_DMACTRL_SFTRST, addrbase + VIA_CRDR_DMACTRL); + + spin_lock_irqsave(&host->lock, flags); + + addrbase = host->pcictrl_mmiobase; + writeb(VIA_CRDR_PCIDMACLK_SDC, addrbase + VIA_CRDR_PCIDMACLK); + + addrbase = host->sdhc_mmiobase; + status = readw(addrbase + VIA_CRDR_SDSTATUS); + if (!(status & VIA_CRDR_SDSTS_SLOTG)) { + if (host->mrq) { + pr_err("%s: Card removed during transfer!\n", + mmc_hostname(host->mmc)); + host->mrq->cmd->error = -ENOMEDIUM; + tasklet_schedule(&host->finish_tasklet); + } + + mmiowb(); + spin_unlock_irqrestore(&host->lock, flags); + + via_reset_pcictrl(host); + + spin_lock_irqsave(&host->lock, flags); + } + + mmiowb(); + spin_unlock_irqrestore(&host->lock, flags); + + via_print_pcictrl(host); + via_print_sdchc(host); + + mmc_detect_change(host->mmc, msecs_to_jiffies(500)); +} + +static void via_init_mmc_host(struct via_crdr_mmc_host *host) +{ + struct mmc_host *mmc = host->mmc; + void __iomem *addrbase; + u32 lenreg; + u32 status; + + init_timer(&host->timer); + host->timer.data = (unsigned long)host; + host->timer.function = via_sdc_timeout; + + spin_lock_init(&host->lock); + + mmc->f_min = VIA_CRDR_MIN_CLOCK; + mmc->f_max = VIA_CRDR_MAX_CLOCK; + mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34 | MMC_VDD_165_195; + mmc->caps = MMC_CAP_4_BIT_DATA | MMC_CAP_SD_HIGHSPEED; + mmc->ops = &via_sdc_ops; + + /*Hardware cannot do scatter lists*/ + mmc->max_hw_segs = 1; + mmc->max_phys_segs = 1; + + mmc->max_blk_size = VIA_CRDR_MAX_BLOCK_LENGTH; + mmc->max_blk_count = VIA_CRDR_MAX_BLOCK_COUNT; + + mmc->max_seg_size = mmc->max_blk_size * mmc->max_blk_count; + mmc->max_req_size = mmc->max_seg_size; + + INIT_WORK(&host->carddet_work, via_sdc_card_detect); + + tasklet_init(&host->finish_tasklet, via_sdc_tasklet_finish, + (unsigned long)host); + + addrbase = host->sdhc_mmiobase; + writel(0x0, addrbase + VIA_CRDR_SDINTMASK); + msleep(1); + + lenreg = VIA_CRDR_SDBLKLEN_GPIDET | VIA_CRDR_SDBLKLEN_INTEN; + writel(lenreg, addrbase + VIA_CRDR_SDBLKLEN); + + status = readw(addrbase + VIA_CRDR_SDSTATUS); + status &= VIA_CRDR_SDSTS_W1C_MASK; + writew(status, addrbase + VIA_CRDR_SDSTATUS); + + status = readw(addrbase + VIA_CRDR_SDSTATUS2); + status |= VIA_CRDR_SDSTS_CFE; + writew(status, addrbase + VIA_CRDR_SDSTATUS2); + + writeb(0x0, addrbase + VIA_CRDR_SDEXTCTRL); + + writel(VIA_CRDR_SDACTIVE_INTMASK, addrbase + VIA_CRDR_SDINTMASK); + msleep(1); +} + +static int __devinit via_sd_probe(struct pci_dev *pcidev, + const struct pci_device_id *id) +{ + struct mmc_host *mmc; + struct via_crdr_mmc_host *sdhost; + u32 base, len; + u8 rev, gatt; + int ret; + + pci_read_config_byte(pcidev, PCI_CLASS_REVISION, &rev); + pr_info(DRV_NAME + ": VIA SDMMC controller found at %s [%04x:%04x] (rev %x)\n", + pci_name(pcidev), (int)pcidev->vendor, (int)pcidev->device, + (int)rev); + + ret = pci_enable_device(pcidev); + if (ret) + return ret; + + ret = pci_request_regions(pcidev, DRV_NAME); + if (ret) + goto disable; + + pci_write_config_byte(pcidev, VIA_CRDR_PCI_WORK_MODE, 0); + pci_write_config_byte(pcidev, VIA_CRDR_PCI_DBG_MODE, 0); + + mmc = mmc_alloc_host(sizeof(struct via_crdr_mmc_host), &pcidev->dev); + if (!mmc) { + ret = -ENOMEM; + goto release; + } + + sdhost = mmc_priv(mmc); + sdhost->mmc = mmc; + dev_set_drvdata(&pcidev->dev, sdhost); + + len = pci_resource_len(pcidev, 0); + base = pci_resource_start(pcidev, 0); + sdhost->mmiobase = ioremap_nocache(base, len); + if (!sdhost->mmiobase) { + ret = -ENOMEM; + goto free_mmc_host; + } + + sdhost->sdhc_mmiobase = + sdhost->mmiobase + VIA_CRDR_SDC_OFF; + sdhost->ddma_mmiobase = + sdhost->mmiobase + VIA_CRDR_DDMA_OFF; + sdhost->pcictrl_mmiobase = + sdhost->mmiobase + VIA_CRDR_PCICTRL_OFF; + + sdhost->power = MMC_VDD_165_195; + + gatt = VIA_CRDR_PCICLKGATT_3V3 | VIA_CRDR_PCICLKGATT_PAD_PWRON; + writeb(gatt, sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT); + via_pwron_sleep(sdhost); + gatt |= VIA_CRDR_PCICLKGATT_SFTRST; + writeb(gatt, sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT); + msleep(3); + + via_init_mmc_host(sdhost); + + ret = + request_irq(pcidev->irq, via_sdc_isr, IRQF_SHARED, DRV_NAME, + sdhost); + if (ret) + goto unmap; + + writeb(VIA_CRDR_PCIINTCTRL_SDCIRQEN, + sdhost->pcictrl_mmiobase + VIA_CRDR_PCIINTCTRL); + writeb(VIA_CRDR_PCITMOCTRL_1024MS, + sdhost->pcictrl_mmiobase + VIA_CRDR_PCITMOCTRL); + + /* device-specific quirks */ + if (pcidev->subsystem_vendor == PCI_VENDOR_ID_LENOVO && + pcidev->subsystem_device == 0x3891) + sdhost->quirks = VIA_CRDR_QUIRK_300MS_PWRDELAY; + + mmc_add_host(mmc); + + return 0; + +unmap: + iounmap(sdhost->mmiobase); +free_mmc_host: + dev_set_drvdata(&pcidev->dev, NULL); + mmc_free_host(mmc); +release: + pci_release_regions(pcidev); +disable: + pci_disable_device(pcidev); + + return ret; +} + +static void __devexit via_sd_remove(struct pci_dev *pcidev) +{ + struct via_crdr_mmc_host *sdhost = pci_get_drvdata(pcidev); + unsigned long flags; + u8 gatt; + + spin_lock_irqsave(&sdhost->lock, flags); + + /* Ensure we don't accept more commands from mmc layer */ + sdhost->reject = 1; + + /* Disable generating further interrupts */ + writeb(0x0, sdhost->pcictrl_mmiobase + VIA_CRDR_PCIINTCTRL); + mmiowb(); + + if (sdhost->mrq) { + printk(KERN_ERR "%s: Controller removed during " + "transfer\n", mmc_hostname(sdhost->mmc)); + + /* make sure all DMA is stopped */ + writel(VIA_CRDR_DMACTRL_SFTRST, + sdhost->ddma_mmiobase + VIA_CRDR_DMACTRL); + mmiowb(); + sdhost->mrq->cmd->error = -ENOMEDIUM; + if (sdhost->mrq->stop) + sdhost->mrq->stop->error = -ENOMEDIUM; + tasklet_schedule(&sdhost->finish_tasklet); + } + spin_unlock_irqrestore(&sdhost->lock, flags); + + mmc_remove_host(sdhost->mmc); + + free_irq(pcidev->irq, sdhost); + + del_timer_sync(&sdhost->timer); + + tasklet_kill(&sdhost->finish_tasklet); + + /* switch off power */ + gatt = readb(sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT); + gatt &= ~VIA_CRDR_PCICLKGATT_PAD_PWRON; + writeb(gatt, sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT); + + iounmap(sdhost->mmiobase); + dev_set_drvdata(&pcidev->dev, NULL); + mmc_free_host(sdhost->mmc); + pci_release_regions(pcidev); + pci_disable_device(pcidev); + + pr_info(DRV_NAME + ": VIA SDMMC controller at %s [%04x:%04x] has been removed\n", + pci_name(pcidev), (int)pcidev->vendor, (int)pcidev->device); +} + +#ifdef CONFIG_PM + +static void via_init_sdc_pm(struct via_crdr_mmc_host *host) +{ + struct sdhcreg *pm_sdhcreg; + void __iomem *addrbase; + u32 lenreg; + u16 status; + + pm_sdhcreg = &(host->pm_sdhc_reg); + addrbase = host->sdhc_mmiobase; + + writel(0x0, addrbase + VIA_CRDR_SDINTMASK); + + lenreg = VIA_CRDR_SDBLKLEN_GPIDET | VIA_CRDR_SDBLKLEN_INTEN; + writel(lenreg, addrbase + VIA_CRDR_SDBLKLEN); + + status = readw(addrbase + VIA_CRDR_SDSTATUS); + status &= VIA_CRDR_SDSTS_W1C_MASK; + writew(status, addrbase + VIA_CRDR_SDSTATUS); + + status = readw(addrbase + VIA_CRDR_SDSTATUS2); + status |= VIA_CRDR_SDSTS_CFE; + writew(status, addrbase + VIA_CRDR_SDSTATUS2); + + writel(pm_sdhcreg->sdcontrol_reg, addrbase + VIA_CRDR_SDCTRL); + writel(pm_sdhcreg->sdcmdarg_reg, addrbase + VIA_CRDR_SDCARG); + writel(pm_sdhcreg->sdintmask_reg, addrbase + VIA_CRDR_SDINTMASK); + writel(pm_sdhcreg->sdrsptmo_reg, addrbase + VIA_CRDR_SDRSPTMO); + writel(pm_sdhcreg->sdclksel_reg, addrbase + VIA_CRDR_SDCLKSEL); + writel(pm_sdhcreg->sdextctrl_reg, addrbase + VIA_CRDR_SDEXTCTRL); + + via_print_pcictrl(host); + via_print_sdchc(host); +} + +static int via_sd_suspend(struct pci_dev *pcidev, pm_message_t state) +{ + struct via_crdr_mmc_host *host; + int ret = 0; + + host = pci_get_drvdata(pcidev); + + via_save_pcictrlreg(host); + via_save_sdcreg(host); + + ret = mmc_suspend_host(host->mmc, state); + + pci_save_state(pcidev); + pci_enable_wake(pcidev, pci_choose_state(pcidev, state), 0); + pci_disable_device(pcidev); + pci_set_power_state(pcidev, pci_choose_state(pcidev, state)); + + return ret; +} + +static int via_sd_resume(struct pci_dev *pcidev) +{ + struct via_crdr_mmc_host *sdhost; + int ret = 0; + u8 gatt; + + sdhost = pci_get_drvdata(pcidev); + + gatt = VIA_CRDR_PCICLKGATT_PAD_PWRON; + if (sdhost->power == MMC_VDD_165_195) + gatt &= ~VIA_CRDR_PCICLKGATT_3V3; + else + gatt |= VIA_CRDR_PCICLKGATT_3V3; + writeb(gatt, sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT); + via_pwron_sleep(sdhost); + gatt |= VIA_CRDR_PCICLKGATT_SFTRST; + writeb(gatt, sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT); + msleep(3); + + msleep(100); + + pci_set_power_state(pcidev, PCI_D0); + pci_restore_state(pcidev); + ret = pci_enable_device(pcidev); + if (ret) + return ret; + + via_restore_pcictrlreg(sdhost); + via_init_sdc_pm(sdhost); + + ret = mmc_resume_host(sdhost->mmc); + + return ret; +} + +#else /* CONFIG_PM */ + +#define via_sd_suspend NULL +#define via_sd_resume NULL + +#endif /* CONFIG_PM */ + +static struct pci_driver via_sd_driver = { + .name = DRV_NAME, + .id_table = via_ids, + .probe = via_sd_probe, + .remove = __devexit_p(via_sd_remove), + .suspend = via_sd_suspend, + .resume = via_sd_resume, +}; + +static int __init via_sd_drv_init(void) +{ + pr_info(DRV_NAME ": VIA SD/MMC Card Reader driver " + "(C) 2008 VIA Technologies, Inc.\n"); + + return pci_register_driver(&via_sd_driver); +} + +static void __exit via_sd_drv_exit(void) +{ + pci_unregister_driver(&via_sd_driver); +} + +module_init(via_sd_drv_init); +module_exit(via_sd_drv_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("VIA Technologies Inc."); +MODULE_DESCRIPTION("VIA SD/MMC Card Interface driver"); -- cgit v0.10.2 From fe9db6cbf16ed64f882999dc0bffef0c65f70c4f Mon Sep 17 00:00:00 2001 From: Ben Dooks Date: Mon, 8 Jun 2009 23:33:56 +0100 Subject: s3cmci: fix dma configuration call This was missed in the DMA changes during the s3c24xx updates in commit 8970ef47d56fd3db28ee798b9d400caf08abd924. Signed-off-by: Ben Dooks Signed-off-by: Pierre Ossman diff --git a/drivers/mmc/host/s3cmci.c b/drivers/mmc/host/s3cmci.c index 4eb4f37..8c08cd7 100644 --- a/drivers/mmc/host/s3cmci.c +++ b/drivers/mmc/host/s3cmci.c @@ -794,7 +794,7 @@ static void s3cmci_dma_setup(struct s3cmci_host *host, host->mem->start + host->sdidata); if (!setup_ok) { - s3c2410_dma_config(host->dma, 4, 0); + s3c2410_dma_config(host->dma, 4); s3c2410_dma_set_buffdone_fn(host->dma, s3cmci_dma_done_callback); s3c2410_dma_setflags(host->dma, S3C2410_DMAF_AUTOSTART); -- cgit v0.10.2 From 557b06971b1f05cbadec2f376a305ee1954e9b0d Mon Sep 17 00:00:00 2001 From: Harald Welte Date: Thu, 18 Jun 2009 16:53:38 +0200 Subject: sdhci: Specific quirk vor VIA SDHCI controller in VX855ES The SDHCI controller found in the VX855ES requires 10ms delay between applying power and applying clock. This issue has been discovered and documented by the OLPC XO1.5 team. Signed-off-by: Harald Welte Signed-off-by: Pierre Ossman diff --git a/drivers/mmc/host/sdhci-pci.c b/drivers/mmc/host/sdhci-pci.c index 65be279..2f15cc1 100644 --- a/drivers/mmc/host/sdhci-pci.c +++ b/drivers/mmc/host/sdhci-pci.c @@ -284,6 +284,18 @@ static const struct sdhci_pci_fixes sdhci_jmicron = { .resume = jmicron_resume, }; +static int via_probe(struct sdhci_pci_chip *chip) +{ + if (chip->pdev->revision == 0x10) + chip->quirks |= SDHCI_QUIRK_DELAY_AFTER_POWER; + + return 0; +} + +static const struct sdhci_pci_fixes sdhci_via = { + .probe = via_probe, +}; + static const struct pci_device_id pci_ids[] __devinitdata = { { .vendor = PCI_VENDOR_ID_RICOH, @@ -349,6 +361,14 @@ static const struct pci_device_id pci_ids[] __devinitdata = { .driver_data = (kernel_ulong_t)&sdhci_jmicron, }, + { + .vendor = PCI_VENDOR_ID_VIA, + .device = 0x95d0, + .subvendor = PCI_ANY_ID, + .subdevice = PCI_ANY_ID, + .driver_data = (kernel_ulong_t)&sdhci_via, + }, + { /* Generic SD host controller */ PCI_DEVICE_CLASS((PCI_CLASS_SYSTEM_SDHCI << 8), 0xFFFF00) }, diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index c758673..f4066fd 100644 --- a/drivers/mmc/host/sdhci.c +++ b/drivers/mmc/host/sdhci.c @@ -1057,6 +1057,13 @@ static void sdhci_set_power(struct sdhci_host *host, unsigned short power) pwr |= SDHCI_POWER_ON; sdhci_writeb(host, pwr, SDHCI_POWER_CONTROL); + + /* + * Some controllers need an extra 10ms delay of 10ms before they + * can apply clock after applying power + */ + if ((host->quirks & SDHCI_QUIRK_DELAY_AFTER_POWER)) + mdelay(10); } /*****************************************************************************\ diff --git a/drivers/mmc/host/sdhci.h b/drivers/mmc/host/sdhci.h index 5d37dd9..831ddf7 100644 --- a/drivers/mmc/host/sdhci.h +++ b/drivers/mmc/host/sdhci.h @@ -230,6 +230,8 @@ struct sdhci_host { #define SDHCI_QUIRK_NO_MULTIBLOCK (1<<21) /* Controller can only handle 1-bit data transfers */ #define SDHCI_QUIRK_FORCE_1_BIT_DATA (1<<22) +/* Controller needs 10ms delay between applying power and clock */ +#define SDHCI_QUIRK_DELAY_AFTER_POWER (1<<23) int irq; /* Device IRQ */ void __iomem * ioaddr; /* Mapped address */ -- cgit v0.10.2 From 11a2f1b78a43d0c2bd026d79b952742c7588f529 Mon Sep 17 00:00:00 2001 From: Pierre Ossman Date: Sun, 21 Jun 2009 20:59:33 +0200 Subject: sdhci: remove needless double parenthesis Signed-off-by: Pierre Ossman diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index f4066fd..6779b4e 100644 --- a/drivers/mmc/host/sdhci.c +++ b/drivers/mmc/host/sdhci.c @@ -584,7 +584,7 @@ static u8 sdhci_calc_timeout(struct sdhci_host *host, struct mmc_data *data) * longer to time out, but that's much better than having a too-short * timeout value. */ - if ((host->quirks & SDHCI_QUIRK_BROKEN_TIMEOUT_VAL)) + if (host->quirks & SDHCI_QUIRK_BROKEN_TIMEOUT_VAL) return 0xE; /* timeout in us */ @@ -1051,7 +1051,7 @@ static void sdhci_set_power(struct sdhci_host *host, unsigned short power) * At least the Marvell CaFe chip gets confused if we set the voltage * and set turn on power at the same time, so set the voltage first. */ - if ((host->quirks & SDHCI_QUIRK_NO_SIMULT_VDD_AND_POWER)) + if (host->quirks & SDHCI_QUIRK_NO_SIMULT_VDD_AND_POWER) sdhci_writeb(host, pwr, SDHCI_POWER_CONTROL); pwr |= SDHCI_POWER_ON; @@ -1062,7 +1062,7 @@ static void sdhci_set_power(struct sdhci_host *host, unsigned short power) * Some controllers need an extra 10ms delay of 10ms before they * can apply clock after applying power */ - if ((host->quirks & SDHCI_QUIRK_DELAY_AFTER_POWER)) + if (host->quirks & SDHCI_QUIRK_DELAY_AFTER_POWER) mdelay(10); } -- cgit v0.10.2 From 232086b19964d0e13359d30d74b11ca31b0751cb Mon Sep 17 00:00:00 2001 From: Johannes Weiner Date: Sat, 20 Jun 2009 02:23:29 +0200 Subject: ipc: unbreak 32-bit shmctl/semctl/msgctl 31a985f "ipc: use __ARCH_WANT_IPC_PARSE_VERSION in ipc/util.h" would choose the implementation of ipc_parse_version() based on a symbol defined in . But it failed to also include this header and thus broke IPC_64-passing 32-bit userspace because the flag wasn't masked out properly anymore and the command not understood. Include to give the architecture a chance to ask for the no-no-op ipc_parse_version(). Signed-off-by: Johannes Weiner Acked-by: Arnd Bergmann Signed-off-by: Linus Torvalds diff --git a/ipc/util.h b/ipc/util.h index ab3ebf2..764b51a 100644 --- a/ipc/util.h +++ b/ipc/util.h @@ -10,6 +10,7 @@ #ifndef _IPC_UTIL_H #define _IPC_UTIL_H +#include #include #define SEQ_MULTIPLIER (IPCMNI) -- cgit v0.10.2 From 30c9f3a9fae79517bca595826a19c6855fbb6d32 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Fri, 10 Apr 2009 08:43:11 -0700 Subject: Remove internal use of 'write_access' in mm/memory.c The fault handling routines really want more fine-grained flags than a single "was it a write fault" boolean - the callers will want to set flags like "you can return a retry error" etc. And that's actually how the VM works internally, but right now the top-level fault handling functions in mm/memory.c all pass just the 'write_access' boolean around. This switches them over to pass around the FAULT_FLAG_xyzzy 'flags' variable instead. The 'write_access' calling convention still exists for the exported 'handle_mm_fault()' function, but that is next. Signed-off-by: Linus Torvalds diff --git a/mm/memory.c b/mm/memory.c index d5d1653..e6a9700 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -2496,7 +2496,7 @@ int vmtruncate_range(struct inode *inode, loff_t offset, loff_t end) */ static int do_swap_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pte_t *page_table, pmd_t *pmd, - int write_access, pte_t orig_pte) + unsigned int flags, pte_t orig_pte) { spinlock_t *ptl; struct page *page; @@ -2572,9 +2572,9 @@ static int do_swap_page(struct mm_struct *mm, struct vm_area_struct *vma, inc_mm_counter(mm, anon_rss); pte = mk_pte(page, vma->vm_page_prot); - if (write_access && reuse_swap_page(page)) { + if ((flags & FAULT_FLAG_WRITE) && reuse_swap_page(page)) { pte = maybe_mkwrite(pte_mkdirty(pte), vma); - write_access = 0; + flags &= ~FAULT_FLAG_WRITE; } flush_icache_page(vma, page); set_pte_at(mm, address, page_table, pte); @@ -2587,7 +2587,7 @@ static int do_swap_page(struct mm_struct *mm, struct vm_area_struct *vma, try_to_free_swap(page); unlock_page(page); - if (write_access) { + if (flags & FAULT_FLAG_WRITE) { ret |= do_wp_page(mm, vma, address, page_table, pmd, ptl, pte); if (ret & VM_FAULT_ERROR) ret &= VM_FAULT_ERROR; @@ -2616,7 +2616,7 @@ out_page: */ static int do_anonymous_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pte_t *page_table, pmd_t *pmd, - int write_access) + unsigned int flags) { struct page *page; spinlock_t *ptl; @@ -2776,7 +2776,7 @@ static int __do_fault(struct mm_struct *mm, struct vm_area_struct *vma, * due to the bad i386 page protection. But it's valid * for other architectures too. * - * Note that if write_access is true, we either now have + * Note that if FAULT_FLAG_WRITE is set, we either now have * an exclusive copy of the page, or this is a shared mapping, * so we can make it writable and dirty to avoid having to * handle that later. @@ -2847,11 +2847,10 @@ unwritable_page: static int do_linear_fault(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pte_t *page_table, pmd_t *pmd, - int write_access, pte_t orig_pte) + unsigned int flags, pte_t orig_pte) { pgoff_t pgoff = (((address & PAGE_MASK) - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff; - unsigned int flags = (write_access ? FAULT_FLAG_WRITE : 0); pte_unmap(page_table); return __do_fault(mm, vma, address, pmd, pgoff, flags, orig_pte); @@ -2868,12 +2867,12 @@ static int do_linear_fault(struct mm_struct *mm, struct vm_area_struct *vma, */ static int do_nonlinear_fault(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pte_t *page_table, pmd_t *pmd, - int write_access, pte_t orig_pte) + unsigned int flags, pte_t orig_pte) { - unsigned int flags = FAULT_FLAG_NONLINEAR | - (write_access ? FAULT_FLAG_WRITE : 0); pgoff_t pgoff; + flags |= FAULT_FLAG_NONLINEAR; + if (!pte_unmap_same(mm, pmd, page_table, orig_pte)) return 0; @@ -2904,7 +2903,7 @@ static int do_nonlinear_fault(struct mm_struct *mm, struct vm_area_struct *vma, */ static inline int handle_pte_fault(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, - pte_t *pte, pmd_t *pmd, int write_access) + pte_t *pte, pmd_t *pmd, unsigned int flags) { pte_t entry; spinlock_t *ptl; @@ -2915,30 +2914,30 @@ static inline int handle_pte_fault(struct mm_struct *mm, if (vma->vm_ops) { if (likely(vma->vm_ops->fault)) return do_linear_fault(mm, vma, address, - pte, pmd, write_access, entry); + pte, pmd, flags, entry); } return do_anonymous_page(mm, vma, address, - pte, pmd, write_access); + pte, pmd, flags); } if (pte_file(entry)) return do_nonlinear_fault(mm, vma, address, - pte, pmd, write_access, entry); + pte, pmd, flags, entry); return do_swap_page(mm, vma, address, - pte, pmd, write_access, entry); + pte, pmd, flags, entry); } ptl = pte_lockptr(mm, pmd); spin_lock(ptl); if (unlikely(!pte_same(*pte, entry))) goto unlock; - if (write_access) { + if (flags & FAULT_FLAG_WRITE) { if (!pte_write(entry)) return do_wp_page(mm, vma, address, pte, pmd, ptl, entry); entry = pte_mkdirty(entry); } entry = pte_mkyoung(entry); - if (ptep_set_access_flags(vma, address, pte, entry, write_access)) { + if (ptep_set_access_flags(vma, address, pte, entry, flags & FAULT_FLAG_WRITE)) { update_mmu_cache(vma, address, entry); } else { /* @@ -2947,7 +2946,7 @@ static inline int handle_pte_fault(struct mm_struct *mm, * This still avoids useless tlb flushes for .text page faults * with threads. */ - if (write_access) + if (flags & FAULT_FLAG_WRITE) flush_tlb_page(vma, address); } unlock: @@ -2965,13 +2964,14 @@ int handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma, pud_t *pud; pmd_t *pmd; pte_t *pte; + unsigned int flags = write_access ? FAULT_FLAG_WRITE : 0; __set_current_state(TASK_RUNNING); count_vm_event(PGFAULT); if (unlikely(is_vm_hugetlb_page(vma))) - return hugetlb_fault(mm, vma, address, write_access); + return hugetlb_fault(mm, vma, address, flags); pgd = pgd_offset(mm, address); pud = pud_alloc(mm, pgd, address); @@ -2984,7 +2984,7 @@ int handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma, if (!pte) return VM_FAULT_OOM; - return handle_pte_fault(mm, vma, address, pte, pmd, write_access); + return handle_pte_fault(mm, vma, address, pte, pmd, flags); } #ifndef __PAGETABLE_PUD_FOLDED -- cgit v0.10.2 From d06063cc221fdefcab86589e79ddfdb7c0e14b63 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Fri, 10 Apr 2009 09:01:23 -0700 Subject: Move FAULT_FLAG_xyz into handle_mm_fault() callers This allows the callers to now pass down the full set of FAULT_FLAG_xyz flags to handle_mm_fault(). All callers have been (mechanically) converted to the new calling convention, there's almost certainly room for architectures to clean up their code and then add FAULT_FLAG_RETRY when that support is added. Signed-off-by: Linus Torvalds diff --git a/arch/alpha/mm/fault.c b/arch/alpha/mm/fault.c index 4829f96..00a31de 100644 --- a/arch/alpha/mm/fault.c +++ b/arch/alpha/mm/fault.c @@ -146,7 +146,7 @@ do_page_fault(unsigned long address, unsigned long mmcsr, /* If for any reason at all we couldn't handle the fault, make sure we exit gracefully rather than endlessly redo the fault. */ - fault = handle_mm_fault(mm, vma, address, cause > 0); + fault = handle_mm_fault(mm, vma, address, cause > 0 ? FAULT_FLAG_WRITE : 0); up_read(&mm->mmap_sem); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) diff --git a/arch/arm/mm/fault.c b/arch/arm/mm/fault.c index 0455557..6fdcbb7 100644 --- a/arch/arm/mm/fault.c +++ b/arch/arm/mm/fault.c @@ -208,7 +208,7 @@ good_area: * than endlessly redo the fault. */ survive: - fault = handle_mm_fault(mm, vma, addr & PAGE_MASK, fsr & (1 << 11)); + fault = handle_mm_fault(mm, vma, addr & PAGE_MASK, (fsr & (1 << 11)) ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/avr32/mm/fault.c b/arch/avr32/mm/fault.c index 62d4abb..b61d86d 100644 --- a/arch/avr32/mm/fault.c +++ b/arch/avr32/mm/fault.c @@ -133,7 +133,7 @@ good_area: * fault. */ survive: - fault = handle_mm_fault(mm, vma, address, writeaccess); + fault = handle_mm_fault(mm, vma, address, writeaccess ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/cris/mm/fault.c b/arch/cris/mm/fault.c index c4c76db..f925115 100644 --- a/arch/cris/mm/fault.c +++ b/arch/cris/mm/fault.c @@ -163,7 +163,7 @@ do_page_fault(unsigned long address, struct pt_regs *regs, * the fault. */ - fault = handle_mm_fault(mm, vma, address, writeaccess & 1); + fault = handle_mm_fault(mm, vma, address, (writeaccess & 1) ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/frv/mm/fault.c b/arch/frv/mm/fault.c index 05093d4..30f5d10 100644 --- a/arch/frv/mm/fault.c +++ b/arch/frv/mm/fault.c @@ -163,7 +163,7 @@ asmlinkage void do_page_fault(int datammu, unsigned long esr0, unsigned long ear * make sure we exit gracefully rather than endlessly redo * the fault. */ - fault = handle_mm_fault(mm, vma, ear0, write); + fault = handle_mm_fault(mm, vma, ear0, write ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/ia64/mm/fault.c b/arch/ia64/mm/fault.c index 23088be..19261a9 100644 --- a/arch/ia64/mm/fault.c +++ b/arch/ia64/mm/fault.c @@ -154,7 +154,7 @@ ia64_do_page_fault (unsigned long address, unsigned long isr, struct pt_regs *re * sure we exit gracefully rather than endlessly redo the * fault. */ - fault = handle_mm_fault(mm, vma, address, (mask & VM_WRITE) != 0); + fault = handle_mm_fault(mm, vma, address, (mask & VM_WRITE) ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { /* * We ran out of memory, or some other thing happened diff --git a/arch/m32r/mm/fault.c b/arch/m32r/mm/fault.c index 4a71df4..7274b47 100644 --- a/arch/m32r/mm/fault.c +++ b/arch/m32r/mm/fault.c @@ -196,7 +196,7 @@ survive: */ addr = (address & PAGE_MASK); set_thread_fault_code(error_code); - fault = handle_mm_fault(mm, vma, addr, write); + fault = handle_mm_fault(mm, vma, addr, write ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/m68k/mm/fault.c b/arch/m68k/mm/fault.c index f493f03..d0e35cf 100644 --- a/arch/m68k/mm/fault.c +++ b/arch/m68k/mm/fault.c @@ -155,7 +155,7 @@ good_area: */ survive: - fault = handle_mm_fault(mm, vma, address, write); + fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0); #ifdef DEBUG printk("handle_mm_fault returns %d\n",fault); #endif diff --git a/arch/microblaze/mm/fault.c b/arch/microblaze/mm/fault.c index 5e67cd1..956607a 100644 --- a/arch/microblaze/mm/fault.c +++ b/arch/microblaze/mm/fault.c @@ -232,7 +232,7 @@ good_area: * the fault. */ survive: - fault = handle_mm_fault(mm, vma, address, is_write); + fault = handle_mm_fault(mm, vma, address, is_write ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/mips/mm/fault.c b/arch/mips/mm/fault.c index 55767ad..6751ce9 100644 --- a/arch/mips/mm/fault.c +++ b/arch/mips/mm/fault.c @@ -102,7 +102,7 @@ good_area: * make sure we exit gracefully rather than endlessly redo * the fault. */ - fault = handle_mm_fault(mm, vma, address, write); + fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/mn10300/mm/fault.c b/arch/mn10300/mm/fault.c index 33cf250..a62e1e1 100644 --- a/arch/mn10300/mm/fault.c +++ b/arch/mn10300/mm/fault.c @@ -258,7 +258,7 @@ good_area: * make sure we exit gracefully rather than endlessly redo * the fault. */ - fault = handle_mm_fault(mm, vma, address, write); + fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/parisc/mm/fault.c b/arch/parisc/mm/fault.c index 92c7fa4..bfb6dd6 100644 --- a/arch/parisc/mm/fault.c +++ b/arch/parisc/mm/fault.c @@ -202,7 +202,7 @@ good_area: * fault. */ - fault = handle_mm_fault(mm, vma, address, (acc_type & VM_WRITE) != 0); + fault = handle_mm_fault(mm, vma, address, (acc_type & VM_WRITE) ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { /* * We hit a shared mapping outside of the file, or some diff --git a/arch/powerpc/mm/fault.c b/arch/powerpc/mm/fault.c index 5beffc8..830bef0 100644 --- a/arch/powerpc/mm/fault.c +++ b/arch/powerpc/mm/fault.c @@ -302,7 +302,7 @@ good_area: * the fault. */ survive: - ret = handle_mm_fault(mm, vma, address, is_write); + ret = handle_mm_fault(mm, vma, address, is_write ? FAULT_FLAG_WRITE : 0); if (unlikely(ret & VM_FAULT_ERROR)) { if (ret & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/powerpc/platforms/cell/spu_fault.c b/arch/powerpc/platforms/cell/spu_fault.c index 95d8dad..d06ba87 100644 --- a/arch/powerpc/platforms/cell/spu_fault.c +++ b/arch/powerpc/platforms/cell/spu_fault.c @@ -70,7 +70,7 @@ int spu_handle_mm_fault(struct mm_struct *mm, unsigned long ea, } ret = 0; - *flt = handle_mm_fault(mm, vma, ea, is_write); + *flt = handle_mm_fault(mm, vma, ea, is_write ? FAULT_FLAG_WRITE : 0); if (unlikely(*flt & VM_FAULT_ERROR)) { if (*flt & VM_FAULT_OOM) { ret = -ENOMEM; diff --git a/arch/s390/lib/uaccess_pt.c b/arch/s390/lib/uaccess_pt.c index b0b84c3..cb5d59e 100644 --- a/arch/s390/lib/uaccess_pt.c +++ b/arch/s390/lib/uaccess_pt.c @@ -66,7 +66,7 @@ static int __handle_fault(struct mm_struct *mm, unsigned long address, } survive: - fault = handle_mm_fault(mm, vma, address, write_access); + fault = handle_mm_fault(mm, vma, address, write_access ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c index 220a152..74eb26b 100644 --- a/arch/s390/mm/fault.c +++ b/arch/s390/mm/fault.c @@ -352,7 +352,7 @@ good_area: * make sure we exit gracefully rather than endlessly redo * the fault. */ - fault = handle_mm_fault(mm, vma, address, write); + fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) { up_read(&mm->mmap_sem); diff --git a/arch/sh/mm/fault_32.c b/arch/sh/mm/fault_32.c index 2c50f80..cc8ddbd 100644 --- a/arch/sh/mm/fault_32.c +++ b/arch/sh/mm/fault_32.c @@ -133,7 +133,7 @@ good_area: * the fault. */ survive: - fault = handle_mm_fault(mm, vma, address, writeaccess); + fault = handle_mm_fault(mm, vma, address, writeaccess ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/sh/mm/tlbflush_64.c b/arch/sh/mm/tlbflush_64.c index 7876997..fcbb6e1 100644 --- a/arch/sh/mm/tlbflush_64.c +++ b/arch/sh/mm/tlbflush_64.c @@ -187,7 +187,7 @@ good_area: * the fault. */ survive: - fault = handle_mm_fault(mm, vma, address, writeaccess); + fault = handle_mm_fault(mm, vma, address, writeaccess ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/sparc/mm/fault_32.c b/arch/sparc/mm/fault_32.c index 12e447f..a5e30c6 100644 --- a/arch/sparc/mm/fault_32.c +++ b/arch/sparc/mm/fault_32.c @@ -241,7 +241,7 @@ good_area: * make sure we exit gracefully rather than endlessly redo * the fault. */ - fault = handle_mm_fault(mm, vma, address, write); + fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; @@ -484,7 +484,7 @@ good_area: if(!(vma->vm_flags & (VM_READ | VM_EXEC))) goto bad_area; } - switch (handle_mm_fault(mm, vma, address, write)) { + switch (handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0)) { case VM_FAULT_SIGBUS: case VM_FAULT_OOM: goto do_sigbus; diff --git a/arch/sparc/mm/fault_64.c b/arch/sparc/mm/fault_64.c index 4ab8993..e5620b2 100644 --- a/arch/sparc/mm/fault_64.c +++ b/arch/sparc/mm/fault_64.c @@ -398,7 +398,7 @@ good_area: goto bad_area; } - fault = handle_mm_fault(mm, vma, address, (fault_code & FAULT_CODE_WRITE)); + fault = handle_mm_fault(mm, vma, address, (fault_code & FAULT_CODE_WRITE) ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/arch/um/kernel/trap.c b/arch/um/kernel/trap.c index 7384d8a..637c650 100644 --- a/arch/um/kernel/trap.c +++ b/arch/um/kernel/trap.c @@ -65,7 +65,7 @@ good_area: do { int fault; - fault = handle_mm_fault(mm, vma, address, is_write); + fault = handle_mm_fault(mm, vma, address, is_write ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) { goto out_of_memory; diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index c403526..78a5fff 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -1113,7 +1113,7 @@ good_area: * make sure we exit gracefully rather than endlessly redo * the fault: */ - fault = handle_mm_fault(mm, vma, address, write); + fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { mm_fault_error(regs, error_code, address, fault); diff --git a/arch/xtensa/mm/fault.c b/arch/xtensa/mm/fault.c index bdd860d..bc07333 100644 --- a/arch/xtensa/mm/fault.c +++ b/arch/xtensa/mm/fault.c @@ -106,7 +106,7 @@ good_area: * the fault. */ survive: - fault = handle_mm_fault(mm, vma, address, is_write); + fault = handle_mm_fault(mm, vma, address, is_write ? FAULT_FLAG_WRITE : 0); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; diff --git a/include/linux/mm.h b/include/linux/mm.h index cf260d8..d006e93 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -810,11 +810,11 @@ extern int vmtruncate_range(struct inode * inode, loff_t offset, loff_t end); #ifdef CONFIG_MMU extern int handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma, - unsigned long address, int write_access); + unsigned long address, unsigned int flags); #else static inline int handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, - int write_access) + unsigned int flags) { /* should never happen if there's no MMU */ BUG(); diff --git a/mm/memory.c b/mm/memory.c index e6a9700..98bcb90 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -1310,8 +1310,9 @@ int __get_user_pages(struct task_struct *tsk, struct mm_struct *mm, cond_resched(); while (!(page = follow_page(vma, start, foll_flags))) { int ret; - ret = handle_mm_fault(mm, vma, start, - foll_flags & FOLL_WRITE); + + /* FOLL_WRITE matches FAULT_FLAG_WRITE! */ + ret = handle_mm_fault(mm, vma, start, foll_flags & FOLL_WRITE); if (ret & VM_FAULT_ERROR) { if (ret & VM_FAULT_OOM) return i ? i : -ENOMEM; @@ -2958,13 +2959,12 @@ unlock: * By the time we get here, we already hold the mm semaphore */ int handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma, - unsigned long address, int write_access) + unsigned long address, unsigned int flags) { pgd_t *pgd; pud_t *pud; pmd_t *pmd; pte_t *pte; - unsigned int flags = write_access ? FAULT_FLAG_WRITE : 0; __set_current_state(TASK_RUNNING); -- cgit v0.10.2 From 85ae87c1ad8e18a421e7448a99a42ecda183f29f Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 22 Jun 2009 11:56:23 +0900 Subject: percpu: fix too lazy vunmap cache flushing In pcpu_unmap(), flushing virtual cache on vunmap can't be delayed as the page is going to be returned to the page allocator. Only TLB flushing can be put off such that vmalloc code can handle it lazily. Fix it. [ Impact: fix subtle virtual cache flush bug ] Signed-off-by: Tejun Heo Cc: Nick Piggin Cc: Ingo Molnar diff --git a/mm/percpu.c b/mm/percpu.c index c0b2c1a..d06f474 100644 --- a/mm/percpu.c +++ b/mm/percpu.c @@ -549,14 +549,14 @@ static void pcpu_free_area(struct pcpu_chunk *chunk, int freeme) * @chunk: chunk of interest * @page_start: page index of the first page to unmap * @page_end: page index of the last page to unmap + 1 - * @flush: whether to flush cache and tlb or not + * @flush_tlb: whether to flush tlb or not * * For each cpu, unmap pages [@page_start,@page_end) out of @chunk. * If @flush is true, vcache is flushed before unmapping and tlb * after. */ static void pcpu_unmap(struct pcpu_chunk *chunk, int page_start, int page_end, - bool flush) + bool flush_tlb) { unsigned int last = num_possible_cpus() - 1; unsigned int cpu; @@ -569,9 +569,8 @@ static void pcpu_unmap(struct pcpu_chunk *chunk, int page_start, int page_end, * the whole region at once rather than doing it for each cpu. * This could be an overkill but is more scalable. */ - if (flush) - flush_cache_vunmap(pcpu_chunk_addr(chunk, 0, page_start), - pcpu_chunk_addr(chunk, last, page_end)); + flush_cache_vunmap(pcpu_chunk_addr(chunk, 0, page_start), + pcpu_chunk_addr(chunk, last, page_end)); for_each_possible_cpu(cpu) unmap_kernel_range_noflush( @@ -579,7 +578,7 @@ static void pcpu_unmap(struct pcpu_chunk *chunk, int page_start, int page_end, (page_end - page_start) << PAGE_SHIFT); /* ditto as flush_cache_vunmap() */ - if (flush) + if (flush_tlb) flush_tlb_kernel_range(pcpu_chunk_addr(chunk, 0, page_start), pcpu_chunk_addr(chunk, last, page_end)); } -- cgit v0.10.2 From c5806df9232d2a7f554b4839b57cac2e664fc256 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 22 Jun 2009 11:56:24 +0900 Subject: x86: fix duplicate free in setup_pcpu_remap() failure path In the failure path, setup_pcpu_remap() tries to free the area which has already been freed to make holes in the large page. Fix it. [ Impact: fix duplicate free in failure path ] Signed-off-by: Tejun Heo Cc: Ingo Molnar diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index 9c3f082..dfbc7e6 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -228,7 +228,7 @@ static ssize_t __init setup_pcpu_remap(size_t static_size) enomem: for_each_possible_cpu(cpu) if (pcpur_ptrs[cpu]) - free_bootmem(__pa(pcpur_ptrs[cpu]), PMD_SIZE); + free_bootmem(__pa(pcpur_ptrs[cpu]), pcpur_size); ret = -ENOMEM; out_free_ar: free_bootmem(__pa(pcpur_ptrs), ptrs_size); -- cgit v0.10.2 From 97c9bf0618cd40b05b4859c1f8a90d8ad97fefb2 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 22 Jun 2009 11:56:24 +0900 Subject: x86: rename remap percpu first chunk allocator to lpage The "remap" allocator remaps large pages to build the first chunk; however, the name isn't very good because 4k allocator remaps too and the whole point of the remap allocator is using large page mapping. The allocator will be generalized and exported outside of x86, rename it to lpage before that happens. percpu_alloc kernel parameter is updated to accept both "remap" and "lpage" for lpage allocator. [ Impact: code cleanup, kernel parameter argument updated ] Signed-off-by: Tejun Heo Cc: Ingo Molnar diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index dfbc7e6..8794c0c 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -124,7 +124,7 @@ static void * __init pcpu_alloc_bootmem(unsigned int cpu, unsigned long size, } /* - * Remap allocator + * Large page remap allocator * * This allocator uses PMD page as unit. A PMD page is allocated for * each cpu and each is remapped into vmalloc area using PMD mapping. @@ -137,20 +137,20 @@ static void * __init pcpu_alloc_bootmem(unsigned int cpu, unsigned long size, * better than only using 4k mappings while still being NUMA friendly. */ #ifdef CONFIG_NEED_MULTIPLE_NODES -static size_t pcpur_size __initdata; -static void **pcpur_ptrs __initdata; +static size_t pcpul_size __initdata; +static void **pcpul_ptrs __initdata; -static struct page * __init pcpur_get_page(unsigned int cpu, int pageno) +static struct page * __init pcpul_get_page(unsigned int cpu, int pageno) { size_t off = (size_t)pageno << PAGE_SHIFT; - if (off >= pcpur_size) + if (off >= pcpul_size) return NULL; - return virt_to_page(pcpur_ptrs[cpu] + off); + return virt_to_page(pcpul_ptrs[cpu] + off); } -static ssize_t __init setup_pcpu_remap(size_t static_size) +static ssize_t __init setup_pcpu_lpage(size_t static_size) { static struct vm_struct vm; size_t ptrs_size, dyn_size; @@ -170,36 +170,36 @@ static ssize_t __init setup_pcpu_remap(size_t static_size) * Currently supports only single page. Supporting multiple * pages won't be too difficult if it ever becomes necessary. */ - pcpur_size = PFN_ALIGN(static_size + PERCPU_MODULE_RESERVE + + pcpul_size = PFN_ALIGN(static_size + PERCPU_MODULE_RESERVE + PERCPU_DYNAMIC_RESERVE); - if (pcpur_size > PMD_SIZE) { + if (pcpul_size > PMD_SIZE) { pr_warning("PERCPU: static data is larger than large page, " "can't use large page\n"); return -EINVAL; } - dyn_size = pcpur_size - static_size - PERCPU_FIRST_CHUNK_RESERVE; + dyn_size = pcpul_size - static_size - PERCPU_FIRST_CHUNK_RESERVE; /* allocate pointer array and alloc large pages */ - ptrs_size = PFN_ALIGN(num_possible_cpus() * sizeof(pcpur_ptrs[0])); - pcpur_ptrs = alloc_bootmem(ptrs_size); + ptrs_size = PFN_ALIGN(num_possible_cpus() * sizeof(pcpul_ptrs[0])); + pcpul_ptrs = alloc_bootmem(ptrs_size); for_each_possible_cpu(cpu) { - pcpur_ptrs[cpu] = pcpu_alloc_bootmem(cpu, PMD_SIZE, PMD_SIZE); - if (!pcpur_ptrs[cpu]) + pcpul_ptrs[cpu] = pcpu_alloc_bootmem(cpu, PMD_SIZE, PMD_SIZE); + if (!pcpul_ptrs[cpu]) goto enomem; /* - * Only use pcpur_size bytes and give back the rest. + * Only use pcpul_size bytes and give back the rest. * * Ingo: The 2MB up-rounding bootmem is needed to make * sure the partial 2MB page is still fully RAM - it's * not well-specified to have a PAT-incompatible area * (unmapped RAM, device memory, etc.) in that hole. */ - free_bootmem(__pa(pcpur_ptrs[cpu] + pcpur_size), - PMD_SIZE - pcpur_size); + free_bootmem(__pa(pcpul_ptrs[cpu] + pcpul_size), + PMD_SIZE - pcpul_size); - memcpy(pcpur_ptrs[cpu], __per_cpu_load, static_size); + memcpy(pcpul_ptrs[cpu], __per_cpu_load, static_size); } /* allocate address and map */ @@ -212,7 +212,7 @@ static ssize_t __init setup_pcpu_remap(size_t static_size) pmd = populate_extra_pmd((unsigned long)vm.addr + cpu * PMD_SIZE); - set_pmd(pmd, pfn_pmd(page_to_pfn(virt_to_page(pcpur_ptrs[cpu])), + set_pmd(pmd, pfn_pmd(page_to_pfn(virt_to_page(pcpul_ptrs[cpu])), PAGE_KERNEL_LARGE)); } @@ -220,22 +220,22 @@ static ssize_t __init setup_pcpu_remap(size_t static_size) pr_info("PERCPU: Remapped at %p with large pages, static data " "%zu bytes\n", vm.addr, static_size); - ret = pcpu_setup_first_chunk(pcpur_get_page, static_size, + ret = pcpu_setup_first_chunk(pcpul_get_page, static_size, PERCPU_FIRST_CHUNK_RESERVE, dyn_size, PMD_SIZE, vm.addr, NULL); goto out_free_ar; enomem: for_each_possible_cpu(cpu) - if (pcpur_ptrs[cpu]) - free_bootmem(__pa(pcpur_ptrs[cpu]), pcpur_size); + if (pcpul_ptrs[cpu]) + free_bootmem(__pa(pcpul_ptrs[cpu]), pcpul_size); ret = -ENOMEM; out_free_ar: - free_bootmem(__pa(pcpur_ptrs), ptrs_size); + free_bootmem(__pa(pcpul_ptrs), ptrs_size); return ret; } #else -static ssize_t __init setup_pcpu_remap(size_t static_size) +static ssize_t __init setup_pcpu_lpage(size_t static_size) { return -EINVAL; } @@ -367,7 +367,7 @@ void __init setup_per_cpu_areas(void) * of large page mappings. Please read comments on top of * each allocator for details. */ - ret = setup_pcpu_remap(static_size); + ret = setup_pcpu_lpage(static_size); if (ret < 0) ret = setup_pcpu_embed(static_size); if (ret < 0) -- cgit v0.10.2 From 0ff2587fd54bd6f66bc6914ada4eb77a7e819a5b Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 22 Jun 2009 11:56:24 +0900 Subject: x86: prepare setup_pcpu_lpage() for pageattr fix Make the following changes in preparation of coming pageattr updates. * Define and use array of struct pcpul_ent instead of array of pointers. The only difference is ->cpu field which is set but unused yet. * Rename variables according to the above change. * Rename local variable vm to pcpul_vm and move it out of the function. [ Impact: no functional difference ] Signed-off-by: Tejun Heo Cc: Jan Beulich Cc: Andi Kleen Cc: Ingo Molnar diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index 8794c0c..7d38941 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -137,8 +137,14 @@ static void * __init pcpu_alloc_bootmem(unsigned int cpu, unsigned long size, * better than only using 4k mappings while still being NUMA friendly. */ #ifdef CONFIG_NEED_MULTIPLE_NODES +struct pcpul_ent { + unsigned int cpu; + void *ptr; +}; + static size_t pcpul_size __initdata; -static void **pcpul_ptrs __initdata; +static struct pcpul_ent *pcpul_map __initdata; +static struct vm_struct pcpul_vm; static struct page * __init pcpul_get_page(unsigned int cpu, int pageno) { @@ -147,13 +153,12 @@ static struct page * __init pcpul_get_page(unsigned int cpu, int pageno) if (off >= pcpul_size) return NULL; - return virt_to_page(pcpul_ptrs[cpu] + off); + return virt_to_page(pcpul_map[cpu].ptr + off); } static ssize_t __init setup_pcpu_lpage(size_t static_size) { - static struct vm_struct vm; - size_t ptrs_size, dyn_size; + size_t map_size, dyn_size; unsigned int cpu; ssize_t ret; @@ -180,12 +185,14 @@ static ssize_t __init setup_pcpu_lpage(size_t static_size) dyn_size = pcpul_size - static_size - PERCPU_FIRST_CHUNK_RESERVE; /* allocate pointer array and alloc large pages */ - ptrs_size = PFN_ALIGN(num_possible_cpus() * sizeof(pcpul_ptrs[0])); - pcpul_ptrs = alloc_bootmem(ptrs_size); + map_size = PFN_ALIGN(num_possible_cpus() * sizeof(pcpul_map[0])); + pcpul_map = alloc_bootmem(map_size); for_each_possible_cpu(cpu) { - pcpul_ptrs[cpu] = pcpu_alloc_bootmem(cpu, PMD_SIZE, PMD_SIZE); - if (!pcpul_ptrs[cpu]) + pcpul_map[cpu].cpu = cpu; + pcpul_map[cpu].ptr = pcpu_alloc_bootmem(cpu, PMD_SIZE, + PMD_SIZE); + if (!pcpul_map[cpu].ptr) goto enomem; /* @@ -196,42 +203,43 @@ static ssize_t __init setup_pcpu_lpage(size_t static_size) * not well-specified to have a PAT-incompatible area * (unmapped RAM, device memory, etc.) in that hole. */ - free_bootmem(__pa(pcpul_ptrs[cpu] + pcpul_size), + free_bootmem(__pa(pcpul_map[cpu].ptr + pcpul_size), PMD_SIZE - pcpul_size); - memcpy(pcpul_ptrs[cpu], __per_cpu_load, static_size); + memcpy(pcpul_map[cpu].ptr, __per_cpu_load, static_size); } /* allocate address and map */ - vm.flags = VM_ALLOC; - vm.size = num_possible_cpus() * PMD_SIZE; - vm_area_register_early(&vm, PMD_SIZE); + pcpul_vm.flags = VM_ALLOC; + pcpul_vm.size = num_possible_cpus() * PMD_SIZE; + vm_area_register_early(&pcpul_vm, PMD_SIZE); for_each_possible_cpu(cpu) { - pmd_t *pmd; + pmd_t *pmd, pmd_v; - pmd = populate_extra_pmd((unsigned long)vm.addr - + cpu * PMD_SIZE); - set_pmd(pmd, pfn_pmd(page_to_pfn(virt_to_page(pcpul_ptrs[cpu])), - PAGE_KERNEL_LARGE)); + pmd = populate_extra_pmd((unsigned long)pcpul_vm.addr + + cpu * PMD_SIZE); + pmd_v = pfn_pmd(page_to_pfn(virt_to_page(pcpul_map[cpu].ptr)), + PAGE_KERNEL_LARGE); + set_pmd(pmd, pmd_v); } /* we're ready, commit */ pr_info("PERCPU: Remapped at %p with large pages, static data " - "%zu bytes\n", vm.addr, static_size); + "%zu bytes\n", pcpul_vm.addr, static_size); ret = pcpu_setup_first_chunk(pcpul_get_page, static_size, PERCPU_FIRST_CHUNK_RESERVE, dyn_size, - PMD_SIZE, vm.addr, NULL); - goto out_free_ar; + PMD_SIZE, pcpul_vm.addr, NULL); + goto out_free_map; enomem: for_each_possible_cpu(cpu) - if (pcpul_ptrs[cpu]) - free_bootmem(__pa(pcpul_ptrs[cpu]), pcpul_size); + if (pcpul_map[cpu].ptr) + free_bootmem(__pa(pcpul_map[cpu].ptr), pcpul_size); ret = -ENOMEM; -out_free_ar: - free_bootmem(__pa(pcpul_ptrs), ptrs_size); +out_free_map: + free_bootmem(__pa(pcpul_map), map_size); return ret; } #else -- cgit v0.10.2 From 992f4c1c2c1583cef3296ec4bf5205843a9a5f3d Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 22 Jun 2009 11:56:24 +0900 Subject: x86: reorganize cpa_process_alias() Reorganize cpa_process_alias() so that new alias condition can be added easily. Jan Beulich spotted problem in the original cleanup thread which incorrectly assumed the two existing conditions were mutially exclusive. [ Impact: code reorganization ] Signed-off-by: Tejun Heo Cc: Jan Beulich Cc: Andi Kleen Cc: Ingo Molnar diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c index 3cfe9ce..2ab058b 100644 --- a/arch/x86/mm/pageattr.c +++ b/arch/x86/mm/pageattr.c @@ -681,8 +681,9 @@ static int __change_page_attr_set_clr(struct cpa_data *cpa, int checkalias); static int cpa_process_alias(struct cpa_data *cpa) { struct cpa_data alias_cpa; - int ret = 0; - unsigned long temp_cpa_vaddr, vaddr; + unsigned long laddr = (unsigned long)__va(cpa->pfn << PAGE_SHIFT); + unsigned long vaddr; + int ret; if (cpa->pfn >= max_pfn_mapped) return 0; @@ -706,42 +707,37 @@ static int cpa_process_alias(struct cpa_data *cpa) PAGE_OFFSET + (max_pfn_mapped << PAGE_SHIFT)))) { alias_cpa = *cpa; - temp_cpa_vaddr = (unsigned long) __va(cpa->pfn << PAGE_SHIFT); - alias_cpa.vaddr = &temp_cpa_vaddr; + alias_cpa.vaddr = &laddr; alias_cpa.flags &= ~(CPA_PAGES_ARRAY | CPA_ARRAY); - ret = __change_page_attr_set_clr(&alias_cpa, 0); + if (ret) + return ret; } #ifdef CONFIG_X86_64 - if (ret) - return ret; /* - * No need to redo, when the primary call touched the high - * mapping already: - */ - if (within(vaddr, (unsigned long) _text, _brk_end)) - return 0; - - /* - * If the physical address is inside the kernel map, we need + * If the primary call didn't touch the high mapping already + * and the physical address is inside the kernel map, we need * to touch the high mapped kernel as well: */ - if (!within(cpa->pfn, highmap_start_pfn(), highmap_end_pfn())) - return 0; - - alias_cpa = *cpa; - temp_cpa_vaddr = (cpa->pfn << PAGE_SHIFT) + __START_KERNEL_map - phys_base; - alias_cpa.vaddr = &temp_cpa_vaddr; - alias_cpa.flags &= ~(CPA_PAGES_ARRAY | CPA_ARRAY); + if (!within(vaddr, (unsigned long)_text, _brk_end) && + within(cpa->pfn, highmap_start_pfn(), highmap_end_pfn())) { + unsigned long temp_cpa_vaddr = (cpa->pfn << PAGE_SHIFT) + + __START_KERNEL_map - phys_base; + alias_cpa = *cpa; + alias_cpa.vaddr = &temp_cpa_vaddr; + alias_cpa.flags &= ~(CPA_PAGES_ARRAY | CPA_ARRAY); - /* - * The high mapping range is imprecise, so ignore the return value. - */ - __change_page_attr_set_clr(&alias_cpa, 0); + /* + * The high mapping range is imprecise, so ignore the + * return value. + */ + __change_page_attr_set_clr(&alias_cpa, 0); + } #endif - return ret; + + return 0; } static int __change_page_attr_set_clr(struct cpa_data *cpa, int checkalias) -- cgit v0.10.2 From e59a1bb2fdfb745c685f5b40ffbed126331d3223 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 22 Jun 2009 11:56:24 +0900 Subject: x86: fix pageattr handling for lpage percpu allocator and re-enable it lpage allocator aliases a PMD page for each cpu and returns whatever is unused to the page allocator. When the pageattr of the recycled pages are changed, this makes the two aliases point to the overlapping regions with different attributes which isn't allowed and known to cause subtle data corruption in certain cases. This can be handled in simliar manner to the x86_64 highmap alias. pageattr code should detect if the target pages have PMD alias and split the PMD alias and synchronize the attributes. pcpur allocator is updated to keep the allocated PMD pages map sorted in ascending address order and provide pcpu_lpage_remapped() function which binary searches the array to determine whether the given address is aliased and if so to which address. pageattr is updated to use pcpu_lpage_remapped() to detect the PMD alias and split it up as necessary from cpa_process_alias(). Jan Beulich spotted the original problem and incorrect usage of vaddr instead of laddr for lookup. With this, lpage percpu allocator should work correctly. Re-enable it. [ Impact: fix subtle lpage pageattr bug and re-enable lpage ] Signed-off-by: Tejun Heo Reported-by: Jan Beulich Cc: Andi Kleen Cc: Ingo Molnar diff --git a/arch/x86/include/asm/percpu.h b/arch/x86/include/asm/percpu.h index 02ecb30..103f1dd 100644 --- a/arch/x86/include/asm/percpu.h +++ b/arch/x86/include/asm/percpu.h @@ -42,6 +42,7 @@ #else /* ...!ASSEMBLY */ +#include #include #ifdef CONFIG_SMP @@ -155,6 +156,15 @@ do { \ /* We can use this directly for local CPU (faster). */ DECLARE_PER_CPU(unsigned long, this_cpu_off); +#ifdef CONFIG_NEED_MULTIPLE_NODES +void *pcpu_lpage_remapped(void *kaddr); +#else +static inline void *pcpu_lpage_remapped(void *kaddr) +{ + return NULL; +} +#endif + #endif /* !__ASSEMBLY__ */ #ifdef CONFIG_SMP diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index 7d38941..bad2fd2 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -142,8 +142,8 @@ struct pcpul_ent { void *ptr; }; -static size_t pcpul_size __initdata; -static struct pcpul_ent *pcpul_map __initdata; +static size_t pcpul_size; +static struct pcpul_ent *pcpul_map; static struct vm_struct pcpul_vm; static struct page * __init pcpul_get_page(unsigned int cpu, int pageno) @@ -160,15 +160,14 @@ static ssize_t __init setup_pcpu_lpage(size_t static_size) { size_t map_size, dyn_size; unsigned int cpu; + int i, j; ssize_t ret; /* * If large page isn't supported, there's no benefit in doing * this. Also, on non-NUMA, embedding is better. - * - * NOTE: disabled for now. */ - if (true || !cpu_has_pse || !pcpu_need_numa()) + if (!cpu_has_pse || !pcpu_need_numa()) return -EINVAL; /* @@ -231,16 +230,71 @@ static ssize_t __init setup_pcpu_lpage(size_t static_size) ret = pcpu_setup_first_chunk(pcpul_get_page, static_size, PERCPU_FIRST_CHUNK_RESERVE, dyn_size, PMD_SIZE, pcpul_vm.addr, NULL); - goto out_free_map; + + /* sort pcpul_map array for pcpu_lpage_remapped() */ + for (i = 0; i < num_possible_cpus() - 1; i++) + for (j = i + 1; j < num_possible_cpus(); j++) + if (pcpul_map[i].ptr > pcpul_map[j].ptr) { + struct pcpul_ent tmp = pcpul_map[i]; + pcpul_map[i] = pcpul_map[j]; + pcpul_map[j] = tmp; + } + + return ret; enomem: for_each_possible_cpu(cpu) if (pcpul_map[cpu].ptr) free_bootmem(__pa(pcpul_map[cpu].ptr), pcpul_size); - ret = -ENOMEM; -out_free_map: free_bootmem(__pa(pcpul_map), map_size); - return ret; + return -ENOMEM; +} + +/** + * pcpu_lpage_remapped - determine whether a kaddr is in pcpul recycled area + * @kaddr: the kernel address in question + * + * Determine whether @kaddr falls in the pcpul recycled area. This is + * used by pageattr to detect VM aliases and break up the pcpu PMD + * mapping such that the same physical page is not mapped under + * different attributes. + * + * The recycled area is always at the tail of a partially used PMD + * page. + * + * RETURNS: + * Address of corresponding remapped pcpu address if match is found; + * otherwise, NULL. + */ +void *pcpu_lpage_remapped(void *kaddr) +{ + void *pmd_addr = (void *)((unsigned long)kaddr & PMD_MASK); + unsigned long offset = (unsigned long)kaddr & ~PMD_MASK; + int left = 0, right = num_possible_cpus() - 1; + int pos; + + /* pcpul in use at all? */ + if (!pcpul_map) + return NULL; + + /* okay, perform binary search */ + while (left <= right) { + pos = (left + right) / 2; + + if (pcpul_map[pos].ptr < pmd_addr) + left = pos + 1; + else if (pcpul_map[pos].ptr > pmd_addr) + right = pos - 1; + else { + /* it shouldn't be in the area for the first chunk */ + WARN_ON(offset < pcpul_size); + + return pcpul_vm.addr + + pcpul_map[pos].cpu * PMD_SIZE + offset; + } + } + + return NULL; } #else static ssize_t __init setup_pcpu_lpage(size_t static_size) diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c index 2ab058b..1b734d7 100644 --- a/arch/x86/mm/pageattr.c +++ b/arch/x86/mm/pageattr.c @@ -11,6 +11,7 @@ #include #include #include +#include #include #include @@ -682,7 +683,7 @@ static int cpa_process_alias(struct cpa_data *cpa) { struct cpa_data alias_cpa; unsigned long laddr = (unsigned long)__va(cpa->pfn << PAGE_SHIFT); - unsigned long vaddr; + unsigned long vaddr, remapped; int ret; if (cpa->pfn >= max_pfn_mapped) @@ -737,6 +738,24 @@ static int cpa_process_alias(struct cpa_data *cpa) } #endif + /* + * If the PMD page was partially used for per-cpu remapping, + * the recycled area needs to be split and modified. Because + * the area is always proper subset of a PMD page + * cpa->numpages is guaranteed to be 1 for these areas, so + * there's no need to loop over and check for further remaps. + */ + remapped = (unsigned long)pcpu_lpage_remapped((void *)laddr); + if (remapped) { + WARN_ON(cpa->numpages > 1); + alias_cpa = *cpa; + alias_cpa.vaddr = &remapped; + alias_cpa.flags &= ~(CPA_PAGES_ARRAY | CPA_ARRAY); + ret = __change_page_attr_set_clr(&alias_cpa, 0); + if (ret) + return ret; + } + return 0; } -- cgit v0.10.2 From fa8a7094ba1679b4b9b443e0ac9f5e046c79ee8d Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 22 Jun 2009 11:56:24 +0900 Subject: x86: implement percpu_alloc kernel parameter According to Andi, it isn't clear whether lpage allocator is worth the trouble as there are many processors where PMD TLB is far scarcer than PTE TLB. The advantage or disadvantage probably depends on the actual size of percpu area and specific processor. As performance degradation due to TLB pressure tends to be highly workload specific and subtle, it is difficult to decide which way to go without more data. This patch implements percpu_alloc kernel parameter to allow selecting which first chunk allocator to use to ease debugging and testing. While at it, make sure all the failure paths report why something failed to help determining why certain allocator isn't working. Also, kill the "Great future plan" comment which had already been realized quite some time ago. [ Impact: allow explicit percpu first chunk allocator selection ] Signed-off-by: Tejun Heo Reported-by: Jan Beulich Cc: Andi Kleen Cc: Ingo Molnar diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 08def8d..ecad946 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -1882,6 +1882,12 @@ and is between 256 and 4096 characters. It is defined in the file Format: { 0 | 1 } See arch/parisc/kernel/pdc_chassis.c + percpu_alloc= [X86] Select which percpu first chunk allocator to use. + Allowed values are one of "lpage", "embed" and "4k". + See comments in arch/x86/kernel/setup_percpu.c for + details on each allocator. This parameter is primarily + for debugging and performance comparison. + pf. [PARIDE] See Documentation/blockdev/paride.txt. diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index bad2fd2..165ebd5 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -156,20 +156,23 @@ static struct page * __init pcpul_get_page(unsigned int cpu, int pageno) return virt_to_page(pcpul_map[cpu].ptr + off); } -static ssize_t __init setup_pcpu_lpage(size_t static_size) +static ssize_t __init setup_pcpu_lpage(size_t static_size, bool chosen) { size_t map_size, dyn_size; unsigned int cpu; int i, j; ssize_t ret; - /* - * If large page isn't supported, there's no benefit in doing - * this. Also, on non-NUMA, embedding is better. - */ - if (!cpu_has_pse || !pcpu_need_numa()) + /* on non-NUMA, embedding is better */ + if (!chosen && !pcpu_need_numa()) return -EINVAL; + /* need PSE */ + if (!cpu_has_pse) { + pr_warning("PERCPU: lpage allocator requires PSE\n"); + return -EINVAL; + } + /* * Currently supports only single page. Supporting multiple * pages won't be too difficult if it ever becomes necessary. @@ -191,8 +194,11 @@ static ssize_t __init setup_pcpu_lpage(size_t static_size) pcpul_map[cpu].cpu = cpu; pcpul_map[cpu].ptr = pcpu_alloc_bootmem(cpu, PMD_SIZE, PMD_SIZE); - if (!pcpul_map[cpu].ptr) + if (!pcpul_map[cpu].ptr) { + pr_warning("PERCPU: failed to allocate large page " + "for cpu%u\n", cpu); goto enomem; + } /* * Only use pcpul_size bytes and give back the rest. @@ -297,7 +303,7 @@ void *pcpu_lpage_remapped(void *kaddr) return NULL; } #else -static ssize_t __init setup_pcpu_lpage(size_t static_size) +static ssize_t __init setup_pcpu_lpage(size_t static_size, bool chosen) { return -EINVAL; } @@ -311,7 +317,7 @@ static ssize_t __init setup_pcpu_lpage(size_t static_size) * mapping so that it can use PMD mapping without additional TLB * pressure. */ -static ssize_t __init setup_pcpu_embed(size_t static_size) +static ssize_t __init setup_pcpu_embed(size_t static_size, bool chosen) { size_t reserve = PERCPU_MODULE_RESERVE + PERCPU_DYNAMIC_RESERVE; @@ -320,7 +326,7 @@ static ssize_t __init setup_pcpu_embed(size_t static_size) * this. Also, embedding allocation doesn't play well with * NUMA. */ - if (!cpu_has_pse || pcpu_need_numa()) + if (!chosen && (!cpu_has_pse || pcpu_need_numa())) return -EINVAL; return pcpu_embed_first_chunk(static_size, PERCPU_FIRST_CHUNK_RESERVE, @@ -370,8 +376,11 @@ static ssize_t __init setup_pcpu_4k(size_t static_size) void *ptr; ptr = pcpu_alloc_bootmem(cpu, PAGE_SIZE, PAGE_SIZE); - if (!ptr) + if (!ptr) { + pr_warning("PERCPU: failed to allocate " + "4k page for cpu%u\n", cpu); goto enomem; + } memcpy(ptr, __per_cpu_load + i * PAGE_SIZE, PAGE_SIZE); pcpu4k_pages[j++] = virt_to_page(ptr); @@ -395,6 +404,16 @@ out_free_ar: return ret; } +/* for explicit first chunk allocator selection */ +static char pcpu_chosen_alloc[16] __initdata; + +static int __init percpu_alloc_setup(char *str) +{ + strncpy(pcpu_chosen_alloc, str, sizeof(pcpu_chosen_alloc) - 1); + return 0; +} +early_param("percpu_alloc", percpu_alloc_setup); + static inline void setup_percpu_segment(int cpu) { #ifdef CONFIG_X86_32 @@ -408,11 +427,6 @@ static inline void setup_percpu_segment(int cpu) #endif } -/* - * Great future plan: - * Declare PDA itself and support (irqstack,tss,pgd) as per cpu data. - * Always point %gs to its beginning - */ void __init setup_per_cpu_areas(void) { size_t static_size = __per_cpu_end - __per_cpu_start; @@ -429,9 +443,26 @@ void __init setup_per_cpu_areas(void) * of large page mappings. Please read comments on top of * each allocator for details. */ - ret = setup_pcpu_lpage(static_size); - if (ret < 0) - ret = setup_pcpu_embed(static_size); + ret = -EINVAL; + if (strlen(pcpu_chosen_alloc)) { + if (strcmp(pcpu_chosen_alloc, "4k")) { + if (!strcmp(pcpu_chosen_alloc, "lpage")) + ret = setup_pcpu_lpage(static_size, true); + else if (!strcmp(pcpu_chosen_alloc, "embed")) + ret = setup_pcpu_embed(static_size, true); + else + pr_warning("PERCPU: unknown allocator %s " + "specified\n", pcpu_chosen_alloc); + if (ret < 0) + pr_warning("PERCPU: %s allocator failed (%zd), " + "falling back to 4k\n", + pcpu_chosen_alloc, ret); + } + } else { + ret = setup_pcpu_lpage(static_size, false); + if (ret < 0) + ret = setup_pcpu_embed(static_size, false); + } if (ret < 0) ret = setup_pcpu_4k(static_size); if (ret < 0) diff --git a/mm/percpu.c b/mm/percpu.c index d06f474..b70f2ac 100644 --- a/mm/percpu.c +++ b/mm/percpu.c @@ -1233,6 +1233,7 @@ static struct page * __init pcpue_get_page(unsigned int cpu, int pageno) ssize_t __init pcpu_embed_first_chunk(size_t static_size, size_t reserved_size, ssize_t dyn_size, ssize_t unit_size) { + size_t chunk_size; unsigned int cpu; /* determine parameters and allocate */ @@ -1247,11 +1248,15 @@ ssize_t __init pcpu_embed_first_chunk(size_t static_size, size_t reserved_size, } else pcpue_unit_size = max_t(size_t, pcpue_size, PCPU_MIN_UNIT_SIZE); - pcpue_ptr = __alloc_bootmem_nopanic( - num_possible_cpus() * pcpue_unit_size, - PAGE_SIZE, __pa(MAX_DMA_ADDRESS)); - if (!pcpue_ptr) + chunk_size = pcpue_unit_size * num_possible_cpus(); + + pcpue_ptr = __alloc_bootmem_nopanic(chunk_size, PAGE_SIZE, + __pa(MAX_DMA_ADDRESS)); + if (!pcpue_ptr) { + pr_warning("PERCPU: failed to allocate %zu bytes for " + "embedding\n", chunk_size); return -ENOMEM; + } /* return the leftover and copy */ for_each_possible_cpu(cpu) { -- cgit v0.10.2 From 0017c869ddcb73069905d09f9e98e68627466237 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 22 Jun 2009 11:56:24 +0900 Subject: x86: ensure percpu lpage doesn't consume too much vmalloc space On extreme configuration (e.g. 32bit 32-way NUMA machine), lpage percpu first chunk allocator can consume too much of vmalloc space. Make it fall back to 4k allocator if the consumption goes over 20%. [ Impact: add sanity check for lpage percpu first chunk allocator ] Signed-off-by: Tejun Heo Reported-by: Jan Beulich Cc: Andi Kleen Cc: Ingo Molnar diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index 165ebd5..29a3eef 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -163,9 +163,21 @@ static ssize_t __init setup_pcpu_lpage(size_t static_size, bool chosen) int i, j; ssize_t ret; - /* on non-NUMA, embedding is better */ - if (!chosen && !pcpu_need_numa()) - return -EINVAL; + if (!chosen) { + size_t vm_size = VMALLOC_END - VMALLOC_START; + size_t tot_size = num_possible_cpus() * PMD_SIZE; + + /* on non-NUMA, embedding is better */ + if (!pcpu_need_numa()) + return -EINVAL; + + /* don't consume more than 20% of vmalloc area */ + if (tot_size > vm_size / 5) { + pr_info("PERCPU: too large chunk size %zuMB for " + "large page remap\n", tot_size >> 20); + return -EINVAL; + } + } /* need PSE */ if (!cpu_has_pse) { -- cgit v0.10.2