- Alsa-devel - mailman.alsa-project.org

[alsa-devel] [PATCH 2/2] ca0106 playback 44100Hz support to SPDIF and playback format & rate constraints
by Ben Stanley 22 Aug '08

22 Aug '08

Signed-off-by: Ben Stanley Ben.Stanley(a)exemail.com.au This is the patch against alsa-driver. --- pci/ca0106/ca0106_main.patch | 8 ++++---- 1 files changed, 4 insertions(+), 4 deletions(-) diff --git a/pci/ca0106/ca0106_main.patch b/pci/ca0106/ca0106_main.patch index 03820b3..d661011 100644 --- a/pci/ca0106/ca0106_main.patch +++ b/pci/ca0106/ca0106_main.patch @@ -14,12 +14,12 @@ static struct snd_ca0106_details ca0106_chip_details[] = { /* Sound Blaster X-Fi Extreme Audio. This does not have an AC97. 53SB079000000 */ -@@ -1352,7 +1353,7 @@ +@@ -1542,7 +1543,7 @@ pci_read_config_word(pci, PCI_SUBSYSTEM_ID, &chip->model); #if 1 - printk(KERN_INFO "snd-ca0106: Model %04x Rev %08x Serial %08x\n", chip->model, -- pci->revision, chip->serial); -+ snd_pci_revision(pci), chip->serial); + printk(KERN_INFO DRVNAME ": Model %04x Rev %08x Serial %08x\n", +- chip->model, pci->revision, chip->serial); ++ chip->model, snd_pci_revision(pci), chip->serial); #endif strcpy(card->driver, "CA0106"); strcpy(card->shortname, "CA0106"); -- 1.5.4.3

1 0

[alsa-devel] [PATCH - asihpi 4/6] Mem buffer alloc inside adapter mutex.
by linux＠audioscience.com 22 Aug '08

22 Aug '08

From: Eliot Blennerhassett <eblennerhassett(a)audioscience.com> Signed-off-by: Eliot Blennerhassett <eblennerhassett(a)audioscience.com> diff --git a/pci/asihpi/hpioctl.c b/pci/asihpi/hpioctl.c index 8222178..cf7c702 100644 --- a/pci/asihpi/hpioctl.c +++ b/pci/asihpi/hpioctl.c @@ -139,7 +139,20 @@ long asihpi_hpi_ioctl( pa = &adapters[hm.wAdapterIndex]; hr.wSize = 0; if (hm.wObject == HPI_OBJ_SUBSYSTEM) { - HPI_MessageF(&hm, &hr, file); + switch (hm.wFunction) { + case HPI_SUBSYS_CREATE_ADAPTER: + case HPI_SUBSYS_DELETE_ADAPTER: + /* Application must not use these functions! */ + hr.wSize = sizeof(struct hpi_response_header); + hr.wError = HPI_ERROR_INVALID_OPERATION; + hr.wFunction = hm.wFunction; + uncopied_bytes = copy_to_user(phr, &hr, hr.wSize); + if (uncopied_bytes) + return -EFAULT; + return 0; + default: + HPI_MessageF(&hm, &hr, file); + } } else { u16 __user *ptr = NULL; u32 size = 0; @@ -164,16 +177,6 @@ long asihpi_hpi_ioctl( /* Dig out any pointers embedded in the message. */ switch (hm.wFunction) { - case HPI_SUBSYS_CREATE_ADAPTER: - case HPI_SUBSYS_DELETE_ADAPTER: - /* Application must not use these functions! */ - hr.wSize = sizeof(struct hpi_response_header); - hr.wError = HPI_ERROR_INVALID_OPERATION; - hr.wFunction = hm.wFunction; - uncopied_bytes = copy_to_user(phr, &hr, hr.wSize); - if (uncopied_bytes) - return -EFAULT; - return 0; case HPI_OSTREAM_WRITE: case HPI_ISTREAM_READ: /* Yes, sparse, this is correct. */ @@ -202,6 +205,9 @@ long asihpi_hpi_ioctl( "HPI could not allocate " "stream buffer size %d\n", size); + + mutex_unlock(&adapters[nAdapter]. + mutex); return -EINVAL; } @@ -216,10 +222,11 @@ long asihpi_hpi_ioctl( break; default: + size = 0; break; } - if (wrflag == 0) { + if (size && (wrflag == 0)) { uncopied_bytes = copy_from_user(pa->pBuffer, ptr, size); if (uncopied_bytes) @@ -231,7 +238,7 @@ long asihpi_hpi_ioctl( HPI_MessageF(&hm, &hr, file); - if (wrflag == 1) { + if (size && (wrflag == 1)) { uncopied_bytes = copy_to_user(ptr, pa->pBuffer, size); if (uncopied_bytes) HPI_DEBUG_LOG(WARNING, -- 1.5.4.3

2 3

[alsa-devel] [PATCH] ALS4000 driver work, step 2
by Andreas Mohr 22 Aug '08

22 Aug '08

Hi, - more register naming work - finally figured out that weird CR register stuff (and did I mention that I hate _really_ undecipherable open-coded values?) - fix handling of IRQ sharing in interrupt handler (hopefully properly, otherwise I'd be grateful to hear your pedantic comments ;) - add handy SPECS_PAGE references wherever useful - comments, cleanup - add me as module author Patch created on 2.6.27-rc1-git4 (diffed against my earlier submission), pushed through checkpatch.pl, verified on my ALS4000 card (listening to it right now of course). Thanks! (especially for your very fast handling) Next patch _will_ contain more useful things, I promise ;) Oh, and apologies to several people for my previous asbestos-requiring comments! Seems Bart's mail address is still valid, hopefully (and found that he's also a Winchip freak, heh). Signed-off-by: Andreas Mohr <andi(a)lisas.de> --- sound/pci/als4000.c.sent 2008-08-18 22:12:27.000000000 +0200 +++ sound/pci/als4000.c 2008-08-21 21:30:10.000000000 +0200 @@ -2,7 +2,7 @@ * card-als4000.c - driver for Avance Logic ALS4000 based soundcards. * Copyright (C) 2000 by Bart Hartgers <bart(a)etpmod.phys.tue.nl>, * Jaroslav Kysela <perex(a)perex.cz> - * Copyright (C) 2002 by Andreas Mohr <hw7oshyuv3001(a)sneakemail.com> + * Copyright (C) 2002, 2008 by Andreas Mohr <hw7oshyuv3001(a)sneakemail.com> * * Framework borrowed from Massimo Piccioni's card-als100.c. * @@ -27,8 +27,10 @@ * bought an ALS4000 based soundcard, I was forced to base this driver * on reverse engineering. * - * Note: this is no longer true. Pretty verbose chip docu (ALS4000a.PDF) - * can be found on the ALSA web site. + * Note: this is no longer true (thank you!): + * pretty verbose chip docu (ALS4000a.PDF) can be found on the ALSA web site. + * Page numbers stated anywhere below with the "SPECS_PAGE:" tag + * refer to: ALS4000a.PDF specs Ver 1.0, May 28th, 1998. * * The ALS4000 seems to be the PCI-cousin of the ALS100. It contains an * ALS100-like SB DSP/mixer, an OPL3 synth, a MPU401 and a gameport @@ -59,7 +61,6 @@ * - value -> some port 0x0c0d * * ToDo: - * - Proper shared IRQ handling? * - by default, don't enable legacy game and use PCI game I/O * - power management? (card can do voice wakeup according to datasheet!!) */ @@ -79,7 +80,7 @@ #include <sound/sb.h> #include <sound/initval.h> -MODULE_AUTHOR("Bart Hartgers <bart(a)etpmod.phys.tue.nl>"); +MODULE_AUTHOR("Bart Hartgers <bart(a)etpmod.phys.tue.nl>, Andreas Mohr"); MODULE_DESCRIPTION("Avance Logic ALS4000"); MODULE_LICENSE("GPL"); MODULE_SUPPORTED_DEVICE("{{Avance Logic,ALS4000}}"); @@ -124,22 +125,22 @@ MODULE_DEVICE_TABLE(pci, snd_als4000_ids); enum als4k_iobase_t { - /* IOx: B == Byte, W = Word, D = DWord */ + /* IOx: B == Byte, W = Word, D = DWord; SPECS_PAGE: 37 */ ALS4K_IOD_00_AC97_ACCESS = 0x00, ALS4K_IOW_04_AC97_READ = 0x04, ALS4K_IOB_06_AC97_STATUS = 0x06, ALS4K_IOB_07_IRQSTATUS = 0x07, ALS4K_IOD_08_GCR_DATA = 0x08, ALS4K_IOB_0C_GCR_INDEX = 0x0c, - ALS4K_IOB_0E_SB_MPU_IRQ = 0x0e, + ALS4K_IOB_0E_IRQTYPE_SB_CR1E_MPU = 0x0e, ALS4K_IOB_10_ADLIB_ADDR0 = 0x10, ALS4K_IOB_11_ADLIB_ADDR1 = 0x11, ALS4K_IOB_12_ADLIB_ADDR2 = 0x12, ALS4K_IOB_13_ADLIB_ADDR3 = 0x13, ALS4K_IOB_14_MIXER_INDEX = 0x14, ALS4K_IOB_15_MIXER_DATA = 0x15, - ALS4K_IOB_16_ESP_RST_PORT = 0x16, - ALS4K_IOB_16_CR1E_ACK_PORT = 0x16, /* 2nd function */ + ALS4K_IOB_16_ESP_RESET = 0x16, + ALS4K_IOB_16_ACK_FOR_CR1E = 0x16, /* 2nd function */ ALS4K_IOB_18_OPL_ADDR0 = 0x18, ALS4K_IOB_19_OPL_ADDR1 = 0x19, ALS4K_IOB_1A_ESP_RD_DATA = 0x1a, @@ -154,62 +155,137 @@ ALS4K_IOB_31_MIDI_COMMAND = 0x31, /* 2nd function */ }; -enum als4k_gcr_t { - /* all registers 32bit wide */ - ALS4K_GCR_8C_MISC_CTRL = 0x8c, - ALS4K_GCR_90_TEST_MODE_REG = 0x90, - ALS4K_GCR_91_DMA0_ADDR = 0x91, - ALS4K_GCR_92_DMA0_MODE_COUNT = 0x92, - ALS4K_GCR_93_DMA1_ADDR = 0x93, - ALS4K_GCR_94_DMA1_MODE_COUNT = 0x94, - ALS4K_GCR_95_DMA3_ADDR = 0x95, - ALS4K_GCR_96_DMA3_MODE_COUNT = 0x96, - ALS4K_GCR_99_DMA_EMULATION_CTRL = 0x99, - ALS4K_GCR_A0_FIFO1_CURRENT_ADDR = 0xa0, - ALS4K_GCR_A1_FIFO1_STATUS_BYTECOUNT = 0xa1, - ALS4K_GCR_A2_FIFO2_PCIADDR = 0xa2, - ALS4K_GCR_A3_FIFO2_COUNT = 0xa3, - ALS4K_GCR_A4_FIFO2_CURRENT_ADDR = 0xa4, - ALS4K_GCR_A5_FIFO1_STATUS_BYTECOUNT = 0xa5, - ALS4K_GCR_A6_PM_CTRL = 0xa6, - ALS4K_GCR_A7_PCI_ACCESS_STORAGE = 0xa7, - ALS4K_GCR_A8_LEGACY_CFG1 = 0xa8, - ALS4K_GCR_A9_LEGACY_CFG2 = 0xa9, - ALS4K_GCR_FF_DUMMY_SCRATCH = 0xff, -}; - -enum als4k_gcr_8c_t { - ALS4K_GCR_8C_IRQ_MASK_CTRL_ENABLE = 0x8000, - ALS4K_GCR_8C_CHIP_REV_MASK = 0xf0000 +enum als4k_iobase_0e_t { + ALS4K_IOB_0E_MPU_IRQ = 0x10, + ALS4K_IOB_0E_CR1E_IRQ = 0x40, + ALS4K_IOB_0E_SB_DMA_IRQ = 0x80, }; -static inline void snd_als4000_gcr_write_addr(unsigned long iobase, - enum als4k_gcr_t reg, - u32 val) +enum als4k_gcr_t { /* all registers 32bit wide; SPECS_PAGE: 38 to 42 */ + ALS4K_GCR8C_MISC_CTRL = 0x8c, + ALS4K_GCR90_TEST_MODE_REG = 0x90, + ALS4K_GCR91_DMA0_ADDR = 0x91, + ALS4K_GCR92_DMA0_MODE_COUNT = 0x92, + ALS4K_GCR93_DMA1_ADDR = 0x93, + ALS4K_GCR94_DMA1_MODE_COUNT = 0x94, + ALS4K_GCR95_DMA3_ADDR = 0x95, + ALS4K_GCR96_DMA3_MODE_COUNT = 0x96, + ALS4K_GCR99_DMA_EMULATION_CTRL = 0x99, + ALS4K_GCRA0_FIFO1_CURRENT_ADDR = 0xa0, + ALS4K_GCRA1_FIFO1_STATUS_BYTECOUNT = 0xa1, + ALS4K_GCRA2_FIFO2_PCIADDR = 0xa2, + ALS4K_GCRA3_FIFO2_COUNT = 0xa3, + ALS4K_GCRA4_FIFO2_CURRENT_ADDR = 0xa4, + ALS4K_GCRA5_FIFO1_STATUS_BYTECOUNT = 0xa5, + ALS4K_GCRA6_PM_CTRL = 0xa6, + ALS4K_GCRA7_PCI_ACCESS_STORAGE = 0xa7, + ALS4K_GCRA8_LEGACY_CFG1 = 0xa8, + ALS4K_GCRA9_LEGACY_CFG2 = 0xa9, + ALS4K_GCRFF_DUMMY_SCRATCH = 0xff, +}; + +enum als4k_gcr8c_t { + ALS4K_GCR8C_IRQ_MASK_CTRL_ENABLE = 0x8000, + ALS4K_GCR8C_CHIP_REV_MASK = 0xf0000 +}; + +static inline void snd_als4k_iobase_writeb(unsigned long iobase, + enum als4k_iobase_t reg, + u8 val) +{ + outb(val, iobase + reg); +} + +static inline void snd_als4k_iobase_writel(unsigned long iobase, + enum als4k_iobase_t reg, + u32 val) +{ + outl(val, iobase + reg); +} + +static inline u8 snd_als4k_iobase_readb(unsigned long iobase, + enum als4k_iobase_t reg) +{ + return inb(iobase + reg); +} + +static inline u32 snd_als4k_iobase_readl(unsigned long iobase, + enum als4k_iobase_t reg) { - outb(reg, iobase + ALS4K_IOB_0C_GCR_INDEX); - outl(val, iobase + ALS4K_IOD_08_GCR_DATA); + return inl(iobase + reg); } -static inline void snd_als4000_gcr_write(struct snd_sb *sb, +static inline void snd_als4k_gcr_write_addr(unsigned long iobase, enum als4k_gcr_t reg, u32 val) { - snd_als4000_gcr_write_addr(sb->alt_port, reg, val); + snd_als4k_iobase_writeb(iobase, ALS4K_IOB_0C_GCR_INDEX, reg); + snd_als4k_iobase_writel(iobase, ALS4K_IOD_08_GCR_DATA, val); +} + +static inline void snd_als4k_gcr_write(struct snd_sb *sb, + enum als4k_gcr_t reg, + u32 val) +{ + snd_als4k_gcr_write_addr(sb->alt_port, reg, val); } -static inline u32 snd_als4000_gcr_read_addr(unsigned long iobase, +static inline u32 snd_als4k_gcr_read_addr(unsigned long iobase, enum als4k_gcr_t reg) { - outb(reg, iobase + ALS4K_IOB_0C_GCR_INDEX); - return inl(iobase + ALS4K_IOD_08_GCR_DATA); + /* SPECS_PAGE: 37/38 */ + snd_als4k_iobase_writeb(iobase, ALS4K_IOB_0C_GCR_INDEX, reg); + return snd_als4k_iobase_readl(iobase, ALS4K_IOD_08_GCR_DATA); +} + +static inline u32 snd_als4k_gcr_read(struct snd_sb *sb, enum als4k_gcr_t reg) +{ + return snd_als4k_gcr_read_addr(sb->alt_port, reg); +} + +enum als4k_cr_t { /* all registers 8bit wide; SPECS_PAGE: 20 to 23 */ + ALS4K_CR0_SB_CONFIG = 0x00, + ALS4K_CR2_MISC_CONTROL = 0x02, + ALS4K_CR3_CONFIGURATION = 0x03, + ALS4K_CR17_FIFO_STATUS = 0x17, + ALS4K_CR18_ESP_MAJOR_VERSION = 0x18, + ALS4K_CR19_ESP_MINOR_VERSION = 0x19, + ALS4K_CR1A_MPU401_UART_MODE_CONTROL = 0x1a, + ALS4K_CR1C_FIFO2_BLOCK_LENGTH_LO = 0x1c, + ALS4K_CR1D_FIFO2_BLOCK_LENGTH_HI = 0x1d, + ALS4K_CR1E_FIFO2_CONTROL = 0x1e, /* secondary PCM FIFO (recording) */ + ALS4K_CR3A_MISC_CONTROL = 0x3a, + ALS4K_CR3B_CRC32_BYTE0 = 0x3b, /* for testing, activate via CR3A */ + ALS4K_CR3C_CRC32_BYTE1 = 0x3c, + ALS4K_CR3D_CRC32_BYTE2 = 0x3d, + ALS4K_CR3E_CRC32_BYTE3 = 0x3e, +}; + +enum als4k_cr0_t { + ALS4K_CR0_DMA_CONTIN_MODE_CTRL = 0x02, /* IRQ/FIFO controlled for 0/1 */ + ALS4K_CR0_DMA_90H_MODE_CTRL = 0x04, /* IRQ/FIFO controlled for 0/1 */ + ALS4K_CR0_MX80_81_REG_WRITE_ENABLE = 0x80, +}; + +static inline void snd_als4_cr_write(struct snd_sb *chip, + enum als4k_cr_t reg, + u8 data) +{ + /* Control Register is reg | 0xc0 (bit 7, 6 set) on sbmixer_index + * NOTE: assumes chip->mixer_lock to be locked externally already! + * SPECS_PAGE: 6 */ + snd_sbmixer_write(chip, reg | 0xc0, data); } -static inline u32 snd_als4000_gcr_read(struct snd_sb *sb, enum als4k_gcr_t reg) +static inline u8 snd_als4_cr_read(struct snd_sb *chip, + enum als4k_cr_t reg) { - return snd_als4000_gcr_read_addr(sb->alt_port, reg); + /* NOTE: assumes chip->mixer_lock to be locked externally already! */ + return snd_sbmixer_read(chip, reg | 0xc0); } + + static void snd_als4000_set_rate(struct snd_sb *chip, unsigned int rate) { if (!(chip->mode & SB_RATE_LOCK)) { @@ -222,16 +298,18 @@ static inline void snd_als4000_set_capture_dma(struct snd_sb *chip, dma_addr_t addr, unsigned size) { - snd_als4000_gcr_write(chip, ALS4K_GCR_A2_FIFO2_PCIADDR, addr); - snd_als4000_gcr_write(chip, ALS4K_GCR_A3_FIFO2_COUNT, (size-1)); + /* SPECS_PAGE: 40 */ + snd_als4k_gcr_write(chip, ALS4K_GCRA2_FIFO2_PCIADDR, addr); + snd_als4k_gcr_write(chip, ALS4K_GCRA3_FIFO2_COUNT, (size-1)); } static inline void snd_als4000_set_playback_dma(struct snd_sb *chip, dma_addr_t addr, unsigned size) { - snd_als4000_gcr_write(chip, ALS4K_GCR_91_DMA0_ADDR, addr); - snd_als4000_gcr_write(chip, ALS4K_GCR_92_DMA0_MODE_COUNT, + /* SPECS_PAGE: 38 */ + snd_als4k_gcr_write(chip, ALS4K_GCR91_DMA0_ADDR, addr); + snd_als4k_gcr_write(chip, ALS4K_GCR92_DMA0_MODE_COUNT, (size-1)|0x180000); } @@ -316,7 +394,7 @@ count = snd_pcm_lib_period_bytes(substream); if (chip->capture_format & ALS4000_FORMAT_16BIT) - count >>=1; + count >>= 1; count--; spin_lock_irq(&chip->reg_lock); @@ -324,8 +402,8 @@ snd_als4000_set_capture_dma(chip, runtime->dma_addr, size); spin_unlock_irq(&chip->reg_lock); spin_lock_irq(&chip->mixer_lock); - snd_sbmixer_write(chip, 0xdc, count); - snd_sbmixer_write(chip, 0xdd, count>>8); + snd_als4_cr_write(chip, ALS4K_CR1C_FIFO2_BLOCK_LENGTH_LO, count & 0xff); + snd_als4_cr_write(chip, ALS4K_CR1D_FIFO2_BLOCK_LENGTH_HI, count >> 8); spin_unlock_irq(&chip->mixer_lock); return 0; } @@ -343,7 +421,7 @@ count = snd_pcm_lib_period_bytes(substream); if (chip->playback_format & ALS4000_FORMAT_16BIT) - count >>=1; + count >>= 1; count--; /* FIXME: from second playback on, there's a lot more clicks and pops @@ -360,8 +438,8 @@ /* snd_sbdsp_command(chip, SB_DSP_SPEAKER_ON); */ snd_sbdsp_command(chip, playback_cmd(chip).dsp_cmd); snd_sbdsp_command(chip, playback_cmd(chip).format); - snd_sbdsp_command(chip, count); - snd_sbdsp_command(chip, count>>8); + snd_sbdsp_command(chip, count & 0xff); + snd_sbdsp_command(chip, count >> 8); snd_sbdsp_command(chip, playback_cmd(chip).dma_off); spin_unlock_irq(&chip->reg_lock); @@ -384,12 +462,14 @@ case SNDRV_PCM_TRIGGER_START: case SNDRV_PCM_TRIGGER_RESUME: chip->mode |= SB_RATE_LOCK_CAPTURE; - snd_sbmixer_write(chip, 0xde, capture_cmd(chip)); + snd_als4_cr_write(chip, ALS4K_CR1E_FIFO2_CONTROL, + capture_cmd(chip)); break; case SNDRV_PCM_TRIGGER_STOP: case SNDRV_PCM_TRIGGER_SUSPEND: chip->mode &= ~SB_RATE_LOCK_CAPTURE; - snd_sbmixer_write(chip, 0xde, 0); + snd_als4_cr_write(chip, ALS4K_CR1E_FIFO2_CONTROL, + capture_cmd(chip)); break; default: result = -EINVAL; @@ -430,9 +510,9 @@ unsigned int result; spin_lock(&chip->reg_lock); - result = snd_als4000_gcr_read(chip, ALS4K_GCR_A4_FIFO2_CURRENT_ADDR); - result &= 0xffff; + result = snd_als4k_gcr_read(chip, ALS4K_GCRA4_FIFO2_CURRENT_ADDR); spin_unlock(&chip->reg_lock); + result &= 0xffff; return bytes_to_frames( substream->runtime, result ); } @@ -442,9 +522,9 @@ unsigned result; spin_lock(&chip->reg_lock); - result = snd_als4000_gcr_read(chip, ALS4K_GCR_A0_FIFO1_CURRENT_ADDR); - result &= 0xffff; + result = snd_als4k_gcr_read(chip, ALS4K_GCRA0_FIFO1_CURRENT_ADDR); spin_unlock(&chip->reg_lock); + result &= 0xffff; return bytes_to_frames( substream->runtime, result ); } @@ -452,46 +532,63 @@ * return IRQ_HANDLED no matter whether we actually had an IRQ flag or not). * ALS4000a.PDF writes that while ACKing IRQ in PCI block will *not* ACK * the IRQ in the SB core, ACKing IRQ in SB block *will* ACK the PCI IRQ - * register (alt_port + ALS4K_IOB_0E_SB_MPU_IRQ). Probably something + * register (alt_port + ALS4K_IOB_0E_IRQTYPE_SB_CR1E_MPU). Probably something * could be optimized here to query/write one register only... * And even if both registers need to be queried, then there's still the * question of whether it's actually correct to ACK PCI IRQ before reading * SB IRQ like we do now, since ALS4000a.PDF mentions that PCI IRQ will *clear* * SB IRQ status. - * (hmm, page 38 mentions it the other way around!) + * (hmm, SPECS_PAGE: 38 mentions it the other way around!) * And do we *really* need the lock here for *reading* SB_DSP4_IRQSTATUS?? * */ static irqreturn_t snd_als4000_interrupt(int irq, void *dev_id) { struct snd_sb *chip = dev_id; - unsigned gcr_status; - unsigned sb_status; - - /* find out which bit of the ALS4000 produced the interrupt */ - gcr_status = inb(chip->alt_port + ALS4K_IOB_0E_SB_MPU_IRQ); + unsigned pci_irqstatus; + unsigned sb_irqstatus; - if ((gcr_status & 0x80) && (chip->playback_substream)) /* playback */ + /* find out which bit of the ALS4000 PCI block produced the interrupt, + SPECS_PAGE: 38, 5 */ + pci_irqstatus = snd_als4k_iobase_readb(chip->alt_port, + ALS4K_IOB_0E_IRQTYPE_SB_CR1E_MPU); + if ((pci_irqstatus & ALS4K_IOB_0E_SB_DMA_IRQ) + && (chip->playback_substream)) /* playback */ snd_pcm_period_elapsed(chip->playback_substream); - if ((gcr_status & 0x40) && (chip->capture_substream)) /* capturing */ + if ((pci_irqstatus & ALS4K_IOB_0E_CR1E_IRQ) + && (chip->capture_substream)) /* capturing */ snd_pcm_period_elapsed(chip->capture_substream); - if ((gcr_status & 0x10) && (chip->rmidi)) /* MPU401 interrupt */ + if ((pci_irqstatus & ALS4K_IOB_0E_MPU_IRQ) + && (chip->rmidi)) /* MPU401 interrupt */ snd_mpu401_uart_interrupt(irq, chip->rmidi->private_data); - /* release the gcr */ - outb(gcr_status, chip->alt_port + ALS4K_IOB_0E_SB_MPU_IRQ); + /* ACK the PCI block IRQ */ + snd_als4k_iobase_writeb(chip->alt_port, + ALS4K_IOB_0E_IRQTYPE_SB_CR1E_MPU, pci_irqstatus); spin_lock(&chip->mixer_lock); - sb_status = snd_sbmixer_read(chip, SB_DSP4_IRQSTATUS); + /* SPECS_PAGE: 20 */ + sb_irqstatus = snd_sbmixer_read(chip, SB_DSP4_IRQSTATUS); spin_unlock(&chip->mixer_lock); - if (sb_status & SB_IRQTYPE_8BIT) + if (sb_irqstatus & SB_IRQTYPE_8BIT) snd_sb_ack_8bit(chip); - if (sb_status & SB_IRQTYPE_16BIT) + if (sb_irqstatus & SB_IRQTYPE_16BIT) snd_sb_ack_16bit(chip); - if (sb_status & SB_IRQTYPE_MPUIN) + if (sb_irqstatus & SB_IRQTYPE_MPUIN) inb(chip->mpu_port); - if (sb_status & 0x20) - inb(SBP(chip, RESET)); - return IRQ_HANDLED; + if (sb_irqstatus & ALS4K_IRQTYPE_CR1E_DMA) + snd_als4k_iobase_readb(chip->alt_port, + ALS4K_IOB_16_ACK_FOR_CR1E); + + /* printk(KERN_INFO "als4000: irq 0x%04x 0x%04x\n", + pci_irqstatus, sb_irqstatus); */ + + /* only ack the things we actually handled above */ + return IRQ_RETVAL( + (pci_irqstatus & (ALS4K_IOB_0E_SB_DMA_IRQ|ALS4K_IOB_0E_CR1E_IRQ| + ALS4K_IOB_0E_MPU_IRQ)) + || (sb_irqstatus & (SB_IRQTYPE_8BIT|SB_IRQTYPE_16BIT| + SB_IRQTYPE_MPUIN|ALS4K_IRQTYPE_CR1E_DMA)) + ); } /*****************************************************************/ @@ -603,7 +700,8 @@ struct snd_pcm *pcm; int err; - if ((err = snd_pcm_new(chip->card, "ALS4000 DSP", device, 1, 1, &pcm)) < 0) + err = snd_pcm_new(chip->card, "ALS4000 DSP", device, 1, 1, &pcm); + if (err < 0) return err; pcm->private_data = chip; pcm->info_flags = SNDRV_PCM_INFO_JOINT_DUPLEX; @@ -637,34 +735,38 @@ cfg1 |= (game_io | 1) << 16; if (opl_io > 0) cfg1 |= (opl_io | 1); - snd_als4000_gcr_write_addr(iobase, ALS4K_GCR_A8_LEGACY_CFG1, cfg1); - snd_als4000_gcr_write_addr(iobase, ALS4K_GCR_A9_LEGACY_CFG2, cfg2); + snd_als4k_gcr_write_addr(iobase, ALS4K_GCRA8_LEGACY_CFG1, cfg1); + snd_als4k_gcr_write_addr(iobase, ALS4K_GCRA9_LEGACY_CFG2, cfg2); } static void snd_als4000_configure(struct snd_sb *chip) { - unsigned tmp; + u8 tmp; int i; /* do some more configuration */ spin_lock_irq(&chip->mixer_lock); - tmp = snd_sbmixer_read(chip, 0xc0); - snd_sbmixer_write(chip, 0xc0, tmp|0x80); - /* always select DMA channel 0, since we do not actually use DMA */ + tmp = snd_als4_cr_read(chip, ALS4K_CR0_SB_CONFIG); + snd_als4_cr_write(chip, ALS4K_CR0_SB_CONFIG, + tmp|ALS4K_CR0_MX80_81_REG_WRITE_ENABLE); + /* always select DMA channel 0, since we do not actually use DMA + * SPECS_PAGE: 19/20 */ snd_sbmixer_write(chip, SB_DSP4_DMASETUP, SB_DMASETUP_DMA0); - snd_sbmixer_write(chip, 0xc0, tmp&0x7f); + snd_als4_cr_write(chip, ALS4K_CR0_SB_CONFIG, + tmp & ~ALS4K_CR0_MX80_81_REG_WRITE_ENABLE); spin_unlock_irq(&chip->mixer_lock); spin_lock_irq(&chip->reg_lock); /* enable interrupts */ - snd_als4000_gcr_write(chip, ALS4K_GCR_8C_MISC_CTRL, - ALS4K_GCR_8C_IRQ_MASK_CTRL_ENABLE); + snd_als4k_gcr_write(chip, ALS4K_GCR8C_MISC_CTRL, + ALS4K_GCR8C_IRQ_MASK_CTRL_ENABLE); - for (i = ALS4K_GCR_91_DMA0_ADDR; i <= ALS4K_GCR_96_DMA3_MODE_COUNT; ++i) - snd_als4000_gcr_write(chip, i, 0); + /* SPECS_PAGE: 39 */ + for (i = ALS4K_GCR91_DMA0_ADDR; i <= ALS4K_GCR96_DMA3_MODE_COUNT; ++i) + snd_als4k_gcr_write(chip, i, 0); - snd_als4000_gcr_write(chip, ALS4K_GCR_99_DMA_EMULATION_CTRL, - snd_als4000_gcr_read(chip, ALS4K_GCR_99_DMA_EMULATION_CTRL)); + snd_als4k_gcr_write(chip, ALS4K_GCR99_DMA_EMULATION_CTRL, + snd_als4k_gcr_read(chip, ALS4K_GCR99_DMA_EMULATION_CTRL)); spin_unlock_irq(&chip->reg_lock); } @@ -739,7 +841,7 @@ struct snd_card_als4000 *acard = card->private_data; /* make sure that interrupts are disabled */ - snd_als4000_gcr_write_addr(acard->iobase, ALS4K_GCR_8C_MISC_CTRL, 0); + snd_als4k_gcr_write_addr(acard->iobase, ALS4K_GCR8C_MISC_CTRL, 0); /* free resources */ snd_als4000_free_gameport(acard); pci_release_regions(acard->pci); @@ -788,7 +890,7 @@ pci_set_master(pci); card = snd_card_new(index[dev], id[dev], THIS_MODULE, - sizeof( struct snd_card_als4000 ) ); + sizeof(*acard) /* private_data: acard */); if (card == NULL) { pci_release_regions(pci); pci_disable_device(pci); @@ -806,6 +908,7 @@ if ((err = snd_sbdsp_create(card, iobase + ALS4K_IOB_10_ADLIB_ADDR0, pci->irq, + /* internally registered as IRQF_SHARED in case of ALS4000 SB */ snd_als4000_interrupt, -1, -1, @@ -835,8 +938,10 @@ goto out_err; } /* FIXME: ALS4000 has interesting MPU401 configuration features - * at CR 0x1A (pass-thru / UART switching, fast MIDI clock, etc.), - * however there doesn't seem to be an ALSA API for this... */ + * at ALS4K_CR1A_MPU401_UART_MODE_CONTROL + * (pass-thru / UART switching, fast MIDI clock, etc.), + * however there doesn't seem to be an ALSA API for this... + * SPECS_PAGE: 21 */ if ((err = snd_als4000_pcm(chip, 0)) < 0) { goto out_err; --- include/sound/sb.h.orig 2008-08-21 21:33:01.000000000 +0200 +++ include/sound/sb.h 2008-08-20 20:22:45.000000000 +0200 @@ -240,11 +240,15 @@ #define SB_DT019X_CAP_MAIN 0x07 #define SB_ALS4000_MONO_IO_CTRL 0x4b +#define SB_ALS4000_OUT_MIXER_CTRL_2 0x4c #define SB_ALS4000_MIC_IN_GAIN 0x4d +#define SB_ALS4000_ANALOG_REFRNC_VOLT_CTRL 0x4e #define SB_ALS4000_FMDAC 0x4f #define SB_ALS4000_3D_SND_FX 0x50 #define SB_ALS4000_3D_TIME_DELAY 0x51 #define SB_ALS4000_3D_AUTO_MUTE 0x52 +#define SB_ALS4000_ANALOG_BLOCK_CTRL 0x53 +#define SB_ALS4000_3D_DELAYLINE_PATTERN 0x54 #define SB_ALS4000_QSOUND 0xdb /* IRQ setting bitmap */ @@ -257,6 +261,7 @@ #define SB_IRQTYPE_8BIT 0x01 #define SB_IRQTYPE_16BIT 0x02 #define SB_IRQTYPE_MPUIN 0x04 +#define ALS4K_IRQTYPE_CR1E_DMA 0x20 /* DMA setting bitmap */ #define SB_DMASETUP_DMA0 0x01

2 1

[alsa-devel] [PATCH] ALSA: snd-ad1816a: add different TerraTec EWS64 S model.
by Rene Herman 22 Aug '08

22 Aug '08

2 1

[alsa-devel] [PATCH - asihpi 7/6] Turn off verbose logging
by linux＠audioscience.com 22 Aug '08

22 Aug '08

From: Eliot Blennerhassett <eblennerhassett(a)audioscience.com> Signed-off-by: Eliot Blennerhassett <eblennerhassett(a)audioscience.com> diff --git a/pci/asihpi/asihpi.c b/pci/asihpi/asihpi.c index fa0718f..0f49fbb 100644 --- a/pci/asihpi/asihpi.c +++ b/pci/asihpi/asihpi.c @@ -22,7 +22,7 @@ * for any purpose including commercial applications. */ /* >0: print Hw params, timer vars. >1: print stream write/copy sizes */ -#define REALLY_VERBOSE_LOGGING 2 +#define REALLY_VERBOSE_LOGGING 0 #if REALLY_VERBOSE_LOGGING #define VPRINTK1 printk -- 1.5.4.3

1 0

[PATCH] ALSA: snd-ad1816a: add different TerraTec EWS64 S model.
by Rene Herman 22 Aug '08

22 Aug '08

Signed-off-by: Rene Herman <rene.herman(a)gmail.com> --- sound/isa/ad1816a/ad1816a.c | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/sound/isa/ad1816a/ad1816a.c b/sound/isa/ad1816a/ad1816a.c index 68f1260..7752424 100644 --- a/sound/isa/ad1816a/ad1816a.c +++ b/sound/isa/ad1816a/ad1816a.c @@ -83,8 +83,10 @@ static struct pnp_card_device_id snd_ad1816a_pnpids[] = { { .id = "MDK1605", .devs = { { .id = "ADS7180" }, { .id = "ADS7181" } } }, /* Shark Predator ISA - added by Ken Arromdee */ { .id = "SMM7180", .devs = { { .id = "ADS7180" }, { .id = "ADS7181" } } }, - /* Analog Devices AD1816A - Terratec AudioSystem EWS64S */ + /* Analog Devices AD1816A - Terratec AudioSystem EWS64 S */ { .id = "TER1112", .devs = { { .id = "ADS7180" }, { .id = "ADS7181" } } }, + /* Analog Devices AD1816A - Terratec AudioSystem EWS64 S */ + { .id = "TER1112", .devs = { { .id = "TER1100" }, { .id = "TER1101" } } }, /* Analog Devices AD1816A - Terratec Base 64 */ { .id = "TER1411", .devs = { { .id = "ADS7180" }, { .id = "ADS7181" } } }, /* end */ -- 1.5.5

1 0

[alsa-devel] About ipc_key, ipc_gid and ipc_perm
by Simon Tremblay 21 Aug '08

21 Aug '08

Hi, someone knows to what is related ipc_key, ipc_gid and ipc_perm settings in /usr/share/alsa/alsa.conf file? I ask that because even if I don't have an audio group created on my embedded system, I can play a wave file with aplay. But since I want to use -D plug:dmix option, I can't. It asks me to create audio group. But after that, I have a strange error I can't explain: audio open error: Operation not permitted. I think ipc_perm has something related to chmod 0660 file? But what is ipc_key ? Does I have to set it somewhere since I installed manually alsa on my system ? Also, On SuSE, I see a + at the end of permission listing of each device file. What does it mean? Like this: crw-rw----+ 1 dax audio 116, 7 2008-08-13 14:06 controlC0 crw-rw----+ 1 dax audio 116, 3 2008-08-13 14:06 midiC0D0 crw-rw----+ 1 dax audio 116, 6 2008-08-13 14:06 pcmC0D0c crw-rw----+ 1 dax audio 116, 5 2008-08-13 14:06 pcmC0D0p crw-rw----+ 1 dax audio 116, 4 2008-08-13 14:06 pcmC0D1p I don't have that on my embedded system, is it something related only to SuSE ? Thanks, Simon

1 0

[alsa-devel] alsa driver on latest Mac Pro (MacPro3,1)
by S. Aguinaga 21 Aug '08

21 Aug '08

Additionally, I noticed that the results file from 'alsa-info.sh' the section about Module is does now show "macpro" under the model section. Might I be doing something wrong? < snippet > .... !!Module: snd_hda_intel bdl_pos_adj : 1,-1,-1,-1,-1,-1,-1,-1 enable : Y,Y,Y,Y,Y,Y,Y,Y enable_msi : 0 id : <NULL>,<NULL>,<NULL>,<NULL>,<NULL>,<NULL>,<NULL>,<NULL> index : -1,-1,-1,-1,-1,-1,-1,-1 model : <NULL>,<NULL>,<NULL>,<NULL>,<NULL>,<NULL>,<NULL>,<NULL> position_fix : 0,0,0,0,0,0,0,0 power_save : 0 power_save_controller : Y probe_mask : -1,-1,-1,-1,-1,-1,-1,-1 single_cmd : N .... < snippet > // Salvador // Northwestern Univeristy ----- Original Message ---- From: Takashi Iwai <tiwai(a)suse.de> To: S. Aguinaga <sa_jr(a)ieee.org> Cc: alsa-devel(a)alsa-project.org Sent: Tuesday, August 19, 2008 5:48:47 AM Subject: Re: [alsa-devel] Alsa-devel Digest, Vol 18, Issue 45 At Mon, 18 Aug 2008 10:26:42 -0700 (PDT), S. Aguinaga wrote: > > ============================================== > Trying to figure out where to add this "model=macpro" > and it looks like it should be in a file called modprobe.conf or an options file under /etc/modprobe.d/ > > but I don't have (with my installation -- fedora {9} ) those files, do you have any input on this? You can create any file, typically, /etc/modprobe.d/sound or so. Put the following line there options snd-hda-intel model=macpro Then reload the module, check the sysfs file /sys/modules/snd_hda_intel/parameters/macpro to see whether the specified option was really passed properly. Takashi > below are the only files I have. > > /etc/modprobe.d/ > > anaconda > blacklist-visor > i915modeset > blacklist > floppy-pnp > modprobe.conf.dist > > > ============================================== > Subject: Re: [alsa-devel] Alsa-devel Digest, Vol 18, Issue 45 > > At Sat, 16 Aug 2008 06:15:46 -0700 (PDT), > S. Aguinaga wrote: > > > > S. Aguinaga wrote: > > > > > > Hello Stan, > > > > > > This is awesome I did find the problem it's listed in the /var/log/messages under ALSA. > > > > > > 1. When I bring up the "Pulse Volume Control" gui, the first tab "Playback" comes up empty, and grayed-out is "No Streams Available" > > > 2. aplay does not produce sound, period! > > > 3. Amrok's welcome sample does not produce sound out of the std tower speaker or the headphone jack. > > > > > > I didn't notice anything strange on the output of the alsa-info.sh file, but when I checked the var/log/messages, this is what I found: > > > > > > sudo cat /var/log/messages | grep ALSA > > > Aug 11 07:52:10 macprolinx kernel: ALSA sound/pci/hda/hda_intel.c:1810: chipset global capabilities = 0x4401 > > > Aug 11 07:52:10 macprolinx kernel: ALSA sound/pci/hda/hda_intel.c:749: codec_mask = 0x1 > > > Aug 11 07:52:10 macprolinx kernel: ALSA sound/pci/hda/hda_codec.c:1073: Cannot find slave Surround Playback Volume, skipped > (snip) > > These messages are harmless. It's just information from the driver, > enabled only when you build it with CONFIG_SND_DEBUG_VERBOSE=y. > > > I think your problem is a misidentification of your sound chip. The > > model being used for your 889A appears to be incompatible, thus all > > the errors above. It appears that the macpro ALC885 is being selected > > before it gets to the choice for the ALC889A. Perhaps a developer can > > look into this. > > Well, ALC882, ALC885 and ALC889A are almost compatible. In the > driver, they are all handled in the same routine. > > There are a few models for intel macs with ALC88x. Try to specify > either model=macpro, model=mbp3 or model=imac24 as the module option > of snd-hda-intel driver. > > > Takashi > > > > >

1 0

[alsa-devel] alsa drivers running under latest Mac Pro hw (MacPro 3, 1, quad core)
by S. Aguinaga 21 Aug '08

21 Aug '08

Stan & Takashi, Still no luck when adding the mode=macpro option to the file: cat /etc/modprobe.d/sound options snd-hda-intel model=macpro The sys module: /sys/module/snd_hda_intel/parameters/model shows: cat model macpro,<NULL>,<NULL>,<NULL>,<NULL>,<NULL>,<NULL>,<NULL> Attached are screen-captures of the volume meter showing no options and the Volume control showing no streams available. // Salvador ----- Original Message ---- From: Takashi Iwai <tiwai(a)suse.de> To: S. Aguinaga <sa_jr(a)ieee.org> Cc: alsa-devel(a)alsa-project.org Sent: Tuesday, August 19, 2008 5:48:47 AM Subject: Re: [alsa-devel] Alsa-devel Digest, Vol 18, Issue 45 At Mon, 18 Aug 2008 10:26:42 -0700 (PDT), S. Aguinaga wrote: > > ============================================== > Trying to figure out where to add this "model=macpro" > and it looks like it should be in a file called modprobe.conf or an options file under /etc/modprobe.d/ > > but I don't have (with my installation -- fedora {9} ) those files, do you have any input on this? You can create any file, typically, /etc/modprobe.d/sound or so. Put the following line there options snd-hda-intel model=macpro Then reload the module, check the sysfs file /sys/modules/snd_hda_intel/parameters/macpro to see whether the specified option was really passed properly. Takashi > below are the only files I have. > > /etc/modprobe.d/ > > anaconda > blacklist-visor > i915modeset > blacklist > floppy-pnp > modprobe.conf.dist > > > ============================================== > Subject: Re: [alsa-devel] Alsa-devel Digest, Vol 18, Issue 45 > > At Sat, 16 Aug 2008 06:15:46 -0700 (PDT), > S. Aguinaga wrote: > > > > S. Aguinaga wrote: > > > > > > Hello Stan, > > > > > > This is awesome I did find the problem it's listed in the /var/log/messages under ALSA. > > > > > > 1. When I bring up the "Pulse Volume Control" gui, the first tab "Playback" comes up empty, and grayed-out is "No Streams Available" > > > 2. aplay does not produce sound, period! > > > 3. Amrok's welcome sample does not produce sound out of the std tower speaker or the headphone jack. > > > > > > I didn't notice anything strange on the output of the alsa-info.sh file, but when I checked the var/log/messages, this is what I found: > > > > > > sudo cat /var/log/messages | grep ALSA > > > Aug 11 07:52:10 macprolinx kernel: ALSA sound/pci/hda/hda_intel.c:1810: chipset global capabilities = 0x4401 > > > Aug 11 07:52:10 macprolinx kernel: ALSA sound/pci/hda/hda_intel.c:749: codec_mask = 0x1 > > > Aug 11 07:52:10 macprolinx kernel: ALSA sound/pci/hda/hda_codec.c:1073: Cannot find slave Surround Playback Volume, skipped > (snip) > > These messages are harmless. It's just information from the driver, > enabled only when you build it with CONFIG_SND_DEBUG_VERBOSE=y. > > > I think your problem is a misidentification of your sound chip. The > > model being used for your 889A appears to be incompatible, thus all > > the errors above. It appears that the macpro ALC885 is being selected > > before it gets to the choice for the ALC889A. Perhaps a developer can > > look into this. > > Well, ALC882, ALC885 and ALC889A are almost compatible. In the > driver, they are all handled in the same routine. > > There are a few models for intel macs with ALC88x. Try to specify > either model=macpro, model=mbp3 or model=imac24 as the module option > of snd-hda-intel driver. > > > Takashi > > > > >

1 0

Re: [alsa-devel] [snd-bt87x]: seems cannot powerup after powerdown...
by Takashi Iwai 21 Aug '08

21 Aug '08

Added Clemens and alsa-devel to Cc. At Mon, 18 Aug 2008 19:22:20 +0400, Dmitry Butskoy wrote: > > Hi, > > Recently snd-bt87x driver has started to use CTL_A_PWRDN bit, see > http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.26.y.git;a=commi… > > Unfortunately, it produces a regression, at least in my case. > > I have "AVerMedia TVPhone 98" card, lspci is: > > 03:05.0 Multimedia video controller: Brooktree Corporation Bt878 Video > > Capture (rev 11) > > 03:05.1 Multimedia controller: Brooktree Corporation Bt878 Audio > > Capture (rev 11) > My card has no additional chip for digital sound, hence only analog is > available. > > After upgrading from 2.6.23 to 2.6.25, I have discovered no more sound > from my snd-bt87x device. Considering the code changes in the dirver, > I've guessed that it could be some power issues. > > Compiling the newest driver without CTL_A_PWRDN (ie. with that changes > reverted back) I got the sound back again. > > Note, that it is not enough to rmmod the old module and insmod the new. > I have to reboot computer completely with the new driver to take changes > in effect. This is a nasty problem. Reloading the module doesn't fix means that the PCI enable/disable process also doesn't suffice... Do you see any difference in PCI register space between the fresh boot and the silent state? thanks, Takashi > > > I can perform any tests, debug etc. if needed. > > > Regards, > Dmitry Butskoy > http://www.fedoraproject.org/wiki/DmitryButskoy >

2 1