January 2018 - Alsa-devel - mailman.alsa-project.org

[alsa-devel] [PATCH 0/5] ASoC: (minor) atmel cleanup
by Ladislav Michl 22 Jan '18

22 Jan '18

Hi there! this patchset is a side product of reading ALSA code while trying to support new hardware. Nothing special here, just an attempt to make those three drivers to look more similar. Only patch 4 fixes actual problem, but: 1) this particular issue is present almost everywhere (as far as I looked) where of_node_put is used in ALSA. Perhaps coccinelle people could help? 2) seems noone bothered as OF_DYNAMIC is mostly disabled Compile tested only. Ladislav Michl (5): ASoC: sam9g20_wm8731: use dev_*() logging functions ASoC: sam9x5_wm8731: Drop 'ASoC' prefix from error messages ASoC: sam9x5_wm8731: Return -ENODEV when probe does not find OF node ASoC: sam9x5_wm8731: Fix OF node reference counting ASoC: atmel_wm8904: Return -ENODEV when probe does not find OF node sound/soc/atmel/atmel_wm8904.c | 6 ++---- sound/soc/atmel/sam9g20_wm8731.c | 17 +++++++-------- sound/soc/atmel/sam9x5_wm8731.c | 45 +++++++++++++++++----------------------- 3 files changed, 29 insertions(+), 39 deletions(-) -- 2.15.1

3 12

[alsa-devel] [bug report] ASoC: mediatek: modify MT2701 AFE driver to adapt mfd device
by Dan Carpenter 22 Jan '18

22 Jan '18

Hello Ryder Lee, The patch dfa3cbb83e09: "ASoC: mediatek: modify MT2701 AFE driver to adapt mfd device" from Jan 4, 2018, leads to the following static checker warning: sound/soc/mediatek/mt2701/mt2701-afe-pcm.c:1535 mt2701_afe_pcm_dev_probe() error: double free of 'component' sound/soc/mediatek/mt2701/mt2701-afe-pcm.c 1509 pm_runtime_get_sync(dev); 1510 1511 ret = snd_soc_register_platform(dev, &mtk_afe_pcm_platform); 1512 if (ret) { 1513 dev_warn(dev, "err_platform\n"); 1514 goto err_platform; 1515 } 1516 1517 ret = snd_soc_add_component(dev, component, ^^^^^^^^^ snd_soc_add_component() frees "component" on error. This seems like a layering violation to me, and it naturally leads to bugs like this. 1518 &mt2701_afe_pcm_dai_component, 1519 mt2701_afe_pcm_dais, 1520 ARRAY_SIZE(mt2701_afe_pcm_dais)); 1521 if (ret) { 1522 dev_warn(dev, "err_dai_component\n"); 1523 goto err_dai_component; 1524 } 1525 1526 return 0; 1527 1528 err_dai_component: 1529 snd_soc_unregister_platform(dev); 1530 err_platform: 1531 pm_runtime_put_sync(dev); 1532 err_pm_disable: 1533 pm_runtime_disable(dev); 1534 err_init_clock: 1535 kfree(component); ^^^^^^^^^ Double free 1536 1537 return ret; 1538 } regards, dan carpenter

2 1

[alsa-devel] [PATCH 3/5] ALSA: emu10k1: add optional debug printouts with DMA addresses
by Maciej S. Szmigiero 21 Jan '18

21 Jan '18

When we get a IOMMU page fault for a emu10k1 device it is very hard to discover which of chip many DMA allocations triggered it (since on a IOMMU system the DMA address space is often very different from the CPU one). Let's add optional debug printouts providing this information. These debug printouts are only enabled on an explicit request via the kernel dynamic debug mechanism. Signed-off-by: Maciej S. Szmigiero <mail(a)maciej.szmigiero.name> --- sound/pci/emu10k1/emu10k1_main.c | 8 ++++++++ sound/pci/emu10k1/memory.c | 11 +++++++++-- 2 files changed, 17 insertions(+), 2 deletions(-) diff --git a/sound/pci/emu10k1/emu10k1_main.c b/sound/pci/emu10k1/emu10k1_main.c index 11e868f569d6..8decd2a7a404 100644 --- a/sound/pci/emu10k1/emu10k1_main.c +++ b/sound/pci/emu10k1/emu10k1_main.c @@ -1898,6 +1898,9 @@ int snd_emu10k1_create(struct snd_card *card, err = -ENOMEM; goto error; } + dev_dbg(card->dev, "page table address range is %.8lx:%.8lx\n", + (unsigned long)emu->ptb_pages.addr, + (unsigned long)(emu->ptb_pages.addr + emu->ptb_pages.bytes)); emu->page_ptr_table = vmalloc(emu->max_cache_pages * sizeof(void *)); emu->page_addr_table = vmalloc(emu->max_cache_pages * @@ -1912,6 +1915,11 @@ int snd_emu10k1_create(struct snd_card *card, err = -ENOMEM; goto error; } + dev_dbg(card->dev, "silent page range is %.8lx:%.8lx\n", + (unsigned long)emu->silent_page.addr, + (unsigned long)(emu->silent_page.addr + + emu->silent_page.bytes)); + emu->memhdr = snd_util_memhdr_new(emu->max_cache_pages * PAGE_SIZE); if (emu->memhdr == NULL) { err = -ENOMEM; diff --git a/sound/pci/emu10k1/memory.c b/sound/pci/emu10k1/memory.c index eaa61024ac7f..39afe199e1c7 100644 --- a/sound/pci/emu10k1/memory.c +++ b/sound/pci/emu10k1/memory.c @@ -35,6 +35,8 @@ */ #define __set_ptb_entry(emu,page,addr) \ (((u32 *)(emu)->ptb_pages.area)[page] = cpu_to_le32(((addr) << (emu->address_mode)) | (page))) +#define __get_ptb_entry(emu, page) \ + (le32_to_cpu(((u32 *)(emu)->ptb_pages.area)[page])) #define UNIT_PAGES (PAGE_SIZE / EMUPAGESIZE) #define MAX_ALIGN_PAGES0 (MAXPAGES0 / UNIT_PAGES) @@ -44,7 +46,7 @@ /* get offset address from aligned page */ #define aligned_page_offset(page) ((page) << PAGE_SHIFT) -#if PAGE_SIZE == 4096 +#if PAGE_SIZE == 4096 && !IS_ENABLED(CONFIG_DYNAMIC_DEBUG) /* page size == EMUPAGESIZE */ /* fill PTB entrie(s) corresponding to page with addr */ #define set_ptb_entry(emu,page,addr) __set_ptb_entry(emu,page,addr) @@ -58,6 +60,8 @@ static inline void set_ptb_entry(struct snd_emu10k1 *emu, int page, dma_addr_t a page *= UNIT_PAGES; for (i = 0; i < UNIT_PAGES; i++, page++) { __set_ptb_entry(emu, page, addr); + dev_dbg(emu->card->dev, "mapped page %d to entry %.8x\n", page, + (unsigned int)__get_ptb_entry(emu, page)); addr += EMUPAGESIZE; } } @@ -65,9 +69,12 @@ static inline void set_silent_ptb(struct snd_emu10k1 *emu, int page) { int i; page *= UNIT_PAGES; - for (i = 0; i < UNIT_PAGES; i++, page++) + for (i = 0; i < UNIT_PAGES; i++, page++) { /* do not increment ptr */ __set_ptb_entry(emu, page, emu->silent_page.addr); + dev_dbg(emu->card->dev, "mapped silent page %d to entry %.8x\n", + page, (unsigned int)__get_ptb_entry(emu, page)); + } } #endif /* PAGE_SIZE */

2 1

[alsa-devel] [PATCH 5/5] ALSA: emu10k1: add a IOMMU workaround
by Maciej S. Szmigiero 21 Jan '18

21 Jan '18

The Audigy 2 CA0102 chip (but most likely others from the emu10k1 family, too) has a problem that from time to time it likes to do few DMA reads a bit beyond its normal allocation and gets very confused if these reads get blocked by a IOMMU. For the first (reserved) page this happens multiple times at every playback, for various synth pages it happens randomly, rarely for the page table memory itself. All these reads seem to follow a similar pattern, observed read offsets beyond the allocation end were 0x00, 0x40, 0x80 and 0xc0 (PCI cache line multiples), so it looks like the device tries to accesses up to 256 extra bytes. As a workaround let's widen these DMA allocations by an extra page if we detect that the device is behind a non-passthrough IOMMU (the DMA memory should be relatively plenty on IOMMU systems). Signed-off-by: Maciej S. Szmigiero <mail(a)maciej.szmigiero.name> --- include/sound/emu10k1.h | 1 + sound/pci/emu10k1/emu10k1_main.c | 49 +++++++++++++++++++++++++++++++++++++--- sound/pci/emu10k1/memory.c | 16 ++++++++++--- 3 files changed, 60 insertions(+), 6 deletions(-) diff --git a/include/sound/emu10k1.h b/include/sound/emu10k1.h index db32b7de52e0..ba27abf65408 100644 --- a/include/sound/emu10k1.h +++ b/include/sound/emu10k1.h @@ -1710,6 +1710,7 @@ struct snd_emu10k1 { unsigned int ecard_ctrl; /* ecard control bits */ unsigned int address_mode; /* address mode */ unsigned long dma_mask; /* PCI DMA mask */ + bool iommu_workaround; /* IOMMU workaround needed */ unsigned int delay_pcm_irq; /* in samples */ int max_cache_pages; /* max memory size / PAGE_SIZE */ struct snd_dma_buffer silent_page; /* silent page */ diff --git a/sound/pci/emu10k1/emu10k1_main.c b/sound/pci/emu10k1/emu10k1_main.c index 8decd2a7a404..3d1b57546379 100644 --- a/sound/pci/emu10k1/emu10k1_main.c +++ b/sound/pci/emu10k1/emu10k1_main.c @@ -36,6 +36,7 @@ #include <linux/init.h> #include <linux/module.h> #include <linux/interrupt.h> +#include <linux/iommu.h> #include <linux/pci.h> #include <linux/slab.h> #include <linux/vmalloc.h> @@ -1758,6 +1759,37 @@ static struct snd_emu_chip_details emu_chip_details[] = { { } /* terminator */ }; +/* + * The chip (at least the Audigy 2 CA0102 chip, but most likely others, too) + * has a problem that from time to time it likes to do few DMA reads a bit + * beyond its normal allocation and gets very confused if these reads get + * blocked by a IOMMU. + * + * This behaviour has been observed for the first (reserved) page + * (for which it happens multiple times at every playback), often for various + * synth pages and sometimes for the page table memory itself. + * + * As a workaround let's widen these DMA allocations by an extra page if we + * detect that the device is behind a non-passthrough IOMMU. + */ +static void snd_emu10k1_detect_iommu(struct snd_emu10k1 *emu) +{ + struct iommu_domain *domain; + + emu->iommu_workaround = false; + + if (!iommu_present(emu->card->dev->bus)) + return; + + domain = iommu_get_domain_for_dev(emu->card->dev); + if (domain && domain->type == IOMMU_DOMAIN_IDENTITY) + return; + + dev_notice(emu->card->dev, + "non-passthrough IOMMU detected, widening DMA allocations"); + emu->iommu_workaround = true; +} + int snd_emu10k1_create(struct snd_card *card, struct pci_dev *pci, unsigned short extin_mask, @@ -1770,6 +1802,7 @@ int snd_emu10k1_create(struct snd_card *card, struct snd_emu10k1 *emu; int idx, err; int is_audigy; + size_t page_table_size, silent_page_size; unsigned int silent_page; const struct snd_emu_chip_details *c; static struct snd_device_ops ops = { @@ -1867,6 +1900,8 @@ int snd_emu10k1_create(struct snd_card *card, is_audigy = emu->audigy = c->emu10k2_chip; + snd_emu10k1_detect_iommu(emu); + /* set addressing mode */ emu->address_mode = is_audigy ? 0 : 1; /* set the DMA transfer mask */ @@ -1893,8 +1928,13 @@ int snd_emu10k1_create(struct snd_card *card, emu->port = pci_resource_start(pci, 0); emu->max_cache_pages = max_cache_bytes >> PAGE_SHIFT; + + page_table_size = sizeof(u32) * (emu->address_mode ? MAXPAGES1 : + MAXPAGES0); + if (emu->iommu_workaround) + page_table_size += PAGE_SIZE; if (snd_dma_alloc_pages(SNDRV_DMA_TYPE_DEV, snd_dma_pci_data(pci), - (emu->address_mode ? 32 : 16) * 1024, &emu->ptb_pages) < 0) { + page_table_size, &emu->ptb_pages) < 0) { err = -ENOMEM; goto error; } @@ -1910,8 +1950,11 @@ int snd_emu10k1_create(struct snd_card *card, goto error; } + silent_page_size = EMUPAGESIZE; + if (emu->iommu_workaround) + silent_page_size *= 2; if (snd_dma_alloc_pages(SNDRV_DMA_TYPE_DEV, snd_dma_pci_data(pci), - EMUPAGESIZE, &emu->silent_page) < 0) { + silent_page_size, &emu->silent_page) < 0) { err = -ENOMEM; goto error; } @@ -1995,7 +2038,7 @@ int snd_emu10k1_create(struct snd_card *card, 0x00000000 | SPCS_EMPHASIS_NONE | SPCS_COPYRIGHT; /* Clear silent pages and set up pointers */ - memset(emu->silent_page.area, 0, PAGE_SIZE); + memset(emu->silent_page.area, 0, silent_page_size); silent_page = emu->silent_page.addr << emu->address_mode; for (idx = 0; idx < (emu->address_mode ? MAXPAGES1 : MAXPAGES0); idx++) ((u32 *)emu->ptb_pages.area)[idx] = cpu_to_le32(silent_page | idx); diff --git a/sound/pci/emu10k1/memory.c b/sound/pci/emu10k1/memory.c index a23eba540132..85414852c3db 100644 --- a/sound/pci/emu10k1/memory.c +++ b/sound/pci/emu10k1/memory.c @@ -457,6 +457,16 @@ static void get_single_page_range(struct snd_util_memhdr *hdr, *last_page_ret = last_page; } +static size_t synth_get_alloc_size(struct snd_emu10k1 *emu) +{ + size_t alloc_size = PAGE_SIZE; + + if (emu->iommu_workaround) + alloc_size *= 2; + + return alloc_size; +} + /* * allocate kernel pages */ @@ -471,7 +481,7 @@ static int synth_alloc_pages(struct snd_emu10k1 *emu, struct snd_emu10k1_memblk for (page = first_page; page <= last_page; page++) { if (snd_dma_alloc_pages(SNDRV_DMA_TYPE_DEV, snd_dma_pci_data(emu->pci), - PAGE_SIZE, &dmab) < 0) + synth_get_alloc_size(emu), &dmab) < 0) goto __fail; if (!is_valid_page(emu, dmab.addr)) { snd_dma_free_pages(&dmab); @@ -488,7 +498,7 @@ static int synth_alloc_pages(struct snd_emu10k1 *emu, struct snd_emu10k1_memblk for (page = first_page; page <= last_page; page++) { dmab.area = emu->page_ptr_table[page]; dmab.addr = emu->page_addr_table[page]; - dmab.bytes = PAGE_SIZE; + dmab.bytes = synth_get_alloc_size(emu); snd_dma_free_pages(&dmab); emu->page_addr_table[page] = 0; emu->page_ptr_table[page] = NULL; @@ -513,7 +523,7 @@ static int synth_free_pages(struct snd_emu10k1 *emu, struct snd_emu10k1_memblk * continue; dmab.area = emu->page_ptr_table[page]; dmab.addr = emu->page_addr_table[page]; - dmab.bytes = PAGE_SIZE; + dmab.bytes = synth_get_alloc_size(emu); snd_dma_free_pages(&dmab); emu->page_addr_table[page] = 0; emu->page_ptr_table[page] = NULL;

1 0

[alsa-devel] [PATCH 4/5] ALSA: emu10k1: make sure synth DMA pages are allocated with DMA functions
by Maciej S. Szmigiero 21 Jan '18

21 Jan '18

Commit a5003fc04113 ("[ALSA] emu10k1 - simplify page allocation for synth") switched from using the DMA allocator for synth DMA pages to manually calling alloc_page(). However, this usage has an implicit assumption that the DMA address space for the emu10k1-family chip is the same as the CPU physical address space which is not true for a system with a IOMMU. Since this made the synth part of the driver non-functional on such systems let's effectively revert that commit. Signed-off-by: Maciej S. Szmigiero <mail(a)maciej.szmigiero.name> --- sound/pci/emu10k1/memory.c | 70 ++++++++++++++++++++++++++-------------------- 1 file changed, 39 insertions(+), 31 deletions(-) diff --git a/sound/pci/emu10k1/memory.c b/sound/pci/emu10k1/memory.c index 39afe199e1c7..a23eba540132 100644 --- a/sound/pci/emu10k1/memory.c +++ b/sound/pci/emu10k1/memory.c @@ -457,49 +457,44 @@ static void get_single_page_range(struct snd_util_memhdr *hdr, *last_page_ret = last_page; } -/* release allocated pages */ -static void __synth_free_pages(struct snd_emu10k1 *emu, int first_page, - int last_page) -{ - int page; - - for (page = first_page; page <= last_page; page++) { - free_page((unsigned long)emu->page_ptr_table[page]); - emu->page_addr_table[page] = 0; - emu->page_ptr_table[page] = NULL; - } -} - /* * allocate kernel pages */ static int synth_alloc_pages(struct snd_emu10k1 *emu, struct snd_emu10k1_memblk *blk) { int page, first_page, last_page; + struct snd_dma_buffer dmab; emu10k1_memblk_init(blk); get_single_page_range(emu->memhdr, blk, &first_page, &last_page); /* allocate kernel pages */ for (page = first_page; page <= last_page; page++) { - /* first try to allocate from <4GB zone */ - struct page *p = alloc_page(GFP_KERNEL | GFP_DMA32 | - __GFP_NOWARN); - if (!p || (page_to_pfn(p) & ~(emu->dma_mask >> PAGE_SHIFT))) { - if (p) - __free_page(p); - /* try to allocate from <16MB zone */ - p = alloc_page(GFP_ATOMIC | GFP_DMA | - __GFP_NORETRY | /* no OOM-killer */ - __GFP_NOWARN); - } - if (!p) { - __synth_free_pages(emu, first_page, page - 1); - return -ENOMEM; + if (snd_dma_alloc_pages(SNDRV_DMA_TYPE_DEV, + snd_dma_pci_data(emu->pci), + PAGE_SIZE, &dmab) < 0) + goto __fail; + if (!is_valid_page(emu, dmab.addr)) { + snd_dma_free_pages(&dmab); + goto __fail; } - emu->page_addr_table[page] = page_to_phys(p); - emu->page_ptr_table[page] = page_address(p); + emu->page_addr_table[page] = dmab.addr; + emu->page_ptr_table[page] = dmab.area; } return 0; + +__fail: + /* release allocated pages */ + last_page = page - 1; + for (page = first_page; page <= last_page; page++) { + dmab.area = emu->page_ptr_table[page]; + dmab.addr = emu->page_addr_table[page]; + dmab.bytes = PAGE_SIZE; + snd_dma_free_pages(&dmab); + emu->page_addr_table[page] = 0; + emu->page_ptr_table[page] = NULL; + } + + return -ENOMEM; } /* @@ -507,10 +502,23 @@ static int synth_alloc_pages(struct snd_emu10k1 *emu, struct snd_emu10k1_memblk */ static int synth_free_pages(struct snd_emu10k1 *emu, struct snd_emu10k1_memblk *blk) { - int first_page, last_page; + int page, first_page, last_page; + struct snd_dma_buffer dmab; get_single_page_range(emu->memhdr, blk, &first_page, &last_page); - __synth_free_pages(emu, first_page, last_page); + dmab.dev.type = SNDRV_DMA_TYPE_DEV; + dmab.dev.dev = snd_dma_pci_data(emu->pci); + for (page = first_page; page <= last_page; page++) { + if (emu->page_ptr_table[page] == NULL) + continue; + dmab.area = emu->page_ptr_table[page]; + dmab.addr = emu->page_addr_table[page]; + dmab.bytes = PAGE_SIZE; + snd_dma_free_pages(&dmab); + emu->page_addr_table[page] = 0; + emu->page_ptr_table[page] = NULL; + } + return 0; }

1 0

[alsa-devel] [PATCH 2/5] ALSA: emu10k1: use dma_set_mask_and_coherent()
by Maciej S. Szmigiero 21 Jan '18

21 Jan '18

We have been calling dma_set_mask() and then dma_set_coherent_mask() with the same value, but there is a dma_set_mask_and_coherent() function that does exactly that so let's use it instead. Signed-off-by: Maciej S. Szmigiero <mail(a)maciej.szmigiero.name> --- sound/pci/emu10k1/emu10k1_main.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/sound/pci/emu10k1/emu10k1_main.c b/sound/pci/emu10k1/emu10k1_main.c index a0a4d31ded53..11e868f569d6 100644 --- a/sound/pci/emu10k1/emu10k1_main.c +++ b/sound/pci/emu10k1/emu10k1_main.c @@ -1871,8 +1871,7 @@ int snd_emu10k1_create(struct snd_card *card, emu->address_mode = is_audigy ? 0 : 1; /* set the DMA transfer mask */ emu->dma_mask = emu->address_mode ? EMU10K1_DMA_MASK : AUDIGY_DMA_MASK; - if (dma_set_mask(&pci->dev, emu->dma_mask) < 0 || - dma_set_coherent_mask(&pci->dev, emu->dma_mask) < 0) { + if (dma_set_mask_and_coherent(&pci->dev, emu->dma_mask) < 0) { dev_err(card->dev, "architecture does not support PCI busmaster DMA with mask 0x%lx\n", emu->dma_mask);

1 0

[alsa-devel] [PATCH 1/5] ALSA: emu10k1: remove reserved_page
by Maciej S. Szmigiero 21 Jan '18

21 Jan '18

The emu10k1-family chips need the first page (index 0) reserved in their page tables for some reason (every emu10k1 driver I've checked does this without much of an explanation). Using the first page for normal samples results in a broken playback. However, we already have a dummy page allocated - so called "silent page" and, in fact, had always been setting it as the first page in the chip page table because an initialization of every entry of the page table to point to a silent page happens after and overwrites the reserved_page allocation. So the only thing remaining to remove the reserved_page allocation is a trivial change to the page allocation logic to ignore the first page entry and start its allocations from the second entry (index 1). Signed-off-by: Maciej S. Szmigiero <mail(a)maciej.szmigiero.name> --- include/sound/emu10k1.h | 1 - sound/pci/emu10k1/emu10k1_main.c | 11 ----------- sound/pci/emu10k1/memory.c | 8 ++++++-- 3 files changed, 6 insertions(+), 14 deletions(-) diff --git a/include/sound/emu10k1.h b/include/sound/emu10k1.h index 4f42affe777c..db32b7de52e0 100644 --- a/include/sound/emu10k1.h +++ b/include/sound/emu10k1.h @@ -1718,7 +1718,6 @@ struct snd_emu10k1 { struct snd_dma_buffer p16v_buffer; struct snd_util_memhdr *memhdr; /* page allocation list */ - struct snd_emu10k1_memblk *reserved_page; /* reserved page */ struct list_head mapped_link_head; struct list_head mapped_order_link_head; diff --git a/sound/pci/emu10k1/emu10k1_main.c b/sound/pci/emu10k1/emu10k1_main.c index ccf4415a1c7b..a0a4d31ded53 100644 --- a/sound/pci/emu10k1/emu10k1_main.c +++ b/sound/pci/emu10k1/emu10k1_main.c @@ -1272,12 +1272,6 @@ static int snd_emu10k1_free(struct snd_emu10k1 *emu) release_firmware(emu->dock_fw); if (emu->irq >= 0) free_irq(emu->irq, emu); - /* remove reserved page */ - if (emu->reserved_page) { - snd_emu10k1_synth_free(emu, - (struct snd_util_memblk *)emu->reserved_page); - emu->reserved_page = NULL; - } snd_util_memhdr_free(emu->memhdr); if (emu->silent_page.area) snd_dma_free_pages(&emu->silent_page); @@ -1993,11 +1987,6 @@ int snd_emu10k1_create(struct snd_card *card, SPCS_GENERATIONSTATUS | 0x00001200 | 0x00000000 | SPCS_EMPHASIS_NONE | SPCS_COPYRIGHT; - emu->reserved_page = (struct snd_emu10k1_memblk *) - snd_emu10k1_synth_alloc(emu, 4096); - if (emu->reserved_page) - emu->reserved_page->map_locked = 1; - /* Clear silent pages and set up pointers */ memset(emu->silent_page.area, 0, PAGE_SIZE); silent_page = emu->silent_page.addr << emu->address_mode; diff --git a/sound/pci/emu10k1/memory.c b/sound/pci/emu10k1/memory.c index 4f1f69be1865..eaa61024ac7f 100644 --- a/sound/pci/emu10k1/memory.c +++ b/sound/pci/emu10k1/memory.c @@ -102,7 +102,7 @@ static void emu10k1_memblk_init(struct snd_emu10k1_memblk *blk) */ static int search_empty_map_area(struct snd_emu10k1 *emu, int npages, struct list_head **nextp) { - int page = 0, found_page = -ENOMEM; + int page = 1, found_page = -ENOMEM; int max_size = npages; int size; struct list_head *candidate = &emu->mapped_link_head; @@ -147,6 +147,10 @@ static int map_memblk(struct snd_emu10k1 *emu, struct snd_emu10k1_memblk *blk) page = search_empty_map_area(emu, blk->pages, &next); if (page < 0) /* not found */ return page; + if (page == 0) { + dev_err(emu->card->dev, "trying to map zero (reserved) page\n"); + return -EINVAL; + } /* insert this block in the proper position of mapped list */ list_add_tail(&blk->mapped_link, next); /* append this as a newest block in order list */ @@ -177,7 +181,7 @@ static int unmap_memblk(struct snd_emu10k1 *emu, struct snd_emu10k1_memblk *blk) q = get_emu10k1_memblk(p, mapped_link); start_page = q->mapped_page + q->pages; } else - start_page = 0; + start_page = 1; if ((p = blk->mapped_link.next) != &emu->mapped_link_head) { q = get_emu10k1_memblk(p, mapped_link); end_page = q->mapped_page;

1 0

Re: [alsa-devel] BUG: unable to handle kernel paging request in memset_erms (2)
by Andrew Morton 20 Jan '18

20 Jan '18

On Fri, 19 Jan 2018 13:58:01 -0800 syzbot <syzbot+29f08ad5cb6820798dfe(a)syzkaller.appspotmail.com> wrote: > Hello, > > syzbot hit the following crash on mmots commit > 2164355612187e55e8d60a28d2cc6b2337841a7e (Fri Jan 19 01:07:54 2018 +0000) > pci: test for unexpectedly disabled bridges > > So far this crash happened 2 times on mmots. > C reproducer is attached. > syzkaller reproducer is attached. > Raw console output is attached. > compiler: gcc (GCC) 7.1.1 20170620 > .config is attached. > > IMPORTANT: if you fix the bug, please add the following tag to the commit: > Reported-by: syzbot+29f08ad5cb6820798dfe(a)syzkaller.appspotmail.com > It will help syzbot understand when the bug is fixed. See footer for > details. > If you forward the report, please keep this part and the footer. > > BUG: unable to handle kernel paging request at ffffc90001691000 > IP: memset_erms+0x9/0x10 arch/x86/lib/memset_64.S:65 > PGD 1dad2c067 P4D 1dad2c067 PUD 1dad2d067 PMD 1c6a8f067 PTE 0 > Oops: 0002 [#1] SMP KASAN > Dumping ftrace buffer: > (ftrace buffer empty) > Modules linked in: > CPU: 1 PID: 5739 Comm: syzkaller592073 Not tainted 4.15.0-rc8-mm1+ #57 > Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS > Google 01/01/2011 > RIP: 0010:memset_erms+0x9/0x10 arch/x86/lib/memset_64.S:65 > RSP: 0018:ffff8801cbbdfb78 EFLAGS: 00010246 > RAX: fffff520002d3f00 RBX: ffffc90001691000 RCX: 000000000000ee51 > RDX: 000000000000ee51 RSI: 0000000000000000 RDI: ffffc90001691000 > RBP: ffff8801cbbdfb98 R08: fffff520002d3fcb R09: ffffc90001691000 > R10: 0000000000001dcb R11: fffff520002d3fca R12: 000000000000ee51 > R13: 0000000000000000 R14: 00007ffffffff000 R15: 000000002001be51 > FS: 00007f88ae7d7700(0000) GS:ffff8801db300000(0000) knlGS:0000000000000000 > CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 > CR2: ffffc90001691000 CR3: 00000001ccefa005 CR4: 00000000001606e0 > DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 > DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 > Call Trace: > memset include/linux/string.h:329 [inline] > _copy_from_user+0xe9/0x110 lib/usercopy.c:16 > copy_from_user include/linux/uaccess.h:147 [inline] > snd_pcm_oss_write1 sound/core/oss/pcm_oss.c:1347 [inline] > snd_pcm_oss_write+0x438/0x880 sound/core/oss/pcm_oss.c:2659 > __vfs_write+0xef/0x970 fs/read_write.c:480 > vfs_write+0x189/0x510 fs/read_write.c:544 > SYSC_write fs/read_write.c:589 [inline] > SyS_write+0xef/0x220 fs/read_write.c:581 > entry_SYSCALL_64_fastpath+0x29/0xa0 > RIP: 0033:0x44a559 > RSP: 002b:00007f88ae7d6da8 EFLAGS: 00000293 ORIG_RAX: 0000000000000001 > RAX: ffffffffffffffda RBX: 00000000006dcc24 RCX: 000000000044a559 > RDX: 000000000000fe51 RSI: 000000002000c000 RDI: 0000000000000003 > RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000 > R10: 0000000000000000 R11: 0000000000000293 R12: 00000000006dcc20 > R13: 7073642f7665642f R14: 00800000c0045006 R15: 0000000000000001 > Code: 48 c1 e9 03 40 0f b6 f6 48 b8 01 01 01 01 01 01 01 01 48 0f af c6 f3 > 48 ab 89 d1 f3 aa 4c 89 c8 c3 90 49 89 f9 40 88 f0 48 89 d1 <f3> aa 4c 89 > c8 c3 90 49 89 fa 40 0f b6 ce 48 b8 01 01 01 01 01 > RIP: memset_erms+0x9/0x10 arch/x86/lib/memset_64.S:65 RSP: ffff8801cbbdfb78 > CR2: ffffc90001691000 > ---[ end trace 8f421641f3e10f44 ]--- > Kernel panic - not syncing: Fatal exception > Dumping ftrace buffer: > (ftrace buffer empty) > Kernel Offset: disabled > Rebooting in 86400 seconds.. It's hard to believe that the (four year old) workaround-for-a-pci-restoring-bug.patch could cause this. From: Linus Torvalds <torvalds(a)linux-foundation.org> Subject: pci: test for unexpectedly disabled bridges The all-ones value is not just a "device didn't exist" case, it's also potentially a quite valid value, so not restoring it would be wrong. What *would* be interesting is to hear where the bad values came from in the first place. It sounds like the device state is saved after the PCI bus controller in front of the device has been crapped on, resulting in the PCI config cycles never reaching the device at all. Something along this patch (together with suspend/resume debugging output) migth help pinpoint it. But it really sounds like something totally brokenly turned off the PCI bridge (some ACPI shutdown crud? I wouldn't be entirely surprised) Cc: Greg KH <greg(a)kroah.com> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> --- drivers/pci/pci.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff -puN drivers/pci/pci.c~workaround-for-a-pci-restoring-bug drivers/pci/pci.c --- a/drivers/pci/pci.c~workaround-for-a-pci-restoring-bug +++ a/drivers/pci/pci.c @@ -1094,6 +1094,15 @@ static void pci_restore_pcix_state(struc int pci_save_state(struct pci_dev *dev) { int i; + u32 val; + + /* Unable to read PCI device/manufacturer state? Something is seriously wrong! */ + if (pci_read_config_dword(dev, 0, &val) || val == 0xffffffff) { + printk("Broken read from PCI device %s\n", pci_name(dev)); + WARN_ON(1); + return -1; + } + /* XXX: 100% dword access ok here? */ for (i = 0; i < 16; i++) pci_read_config_dword(dev, i * 4, &dev->saved_config_space[i]); _

2 1

[alsa-devel] Bug report for: [P7xxDM2(-G), Realtek ALC898, issue: Headphone Jack, External] No sound at all
by Dennis Mungai 20 Jan '18

20 Jan '18

Hello there, I have Ubuntu 16.04LTS installed on the Clevo P751DM2-G, also marketed as the Origin Eon 15-X (Late 2016 model), the System76's Serval WS (2017 model), the Eurocom Sky X4E2, among others. The sound card is the Realtek ALC898 with the following output configuration: 1. One SPDIF/headphone combo jack (to the right) 2. One Line Out. 3. One Line-In. 4. One Mic-In. Background information: On Windows, one can use the Realtek HD Audio control panel pop-up menu on jack detection to set the output type when an audio device is plugged in to an appropriate jack, combined with Sound Blaster X-Fi MB5 software to enable post-processing effects such as Bass, noise cancellation, etc. What I have tried so far: 1. Setting the model definition at /etc/modprobe.d/sound.conf with the following entries: options snd-hda-intel model=no-primary-hp enable=1 index=0 Using such a line worked before on an MSI GS43VR 6RE Phantom Pro that also has the same (or a similar) ESS Sabre Audio DAC component. Expected behavior: When plugged into the microphone jack, the ESS Sabre DAC should be activated and sound should be routed to the headset. Observed anomaly: Sound comes out through the speakers instead. However, the line-out output works. In that case, when plugged into the line out, Audio works as expected. Platform information: DistroRelease: Ubuntu 16.04 Package: alsa-base 1.0.25+dfsg-0ubuntu5 Uname: Linux 4.14.0 x86_64 NonfreeKernelModules: nvidia_uvm nvidia_drm nvidia_modeset nvidia ApportVersion: 2.20.1-0ubuntu2.10 Architecture: amd64 AudioDevicesInUse: USER PID ACCESS COMMAND /dev/snd/controlC1: lin 2199 F.... pulseaudio /dev/snd/controlC0: lin 2199 F.... pulseaudio CurrentDesktop: GNOME InstallationDate: Installed on 2017-07-25 InstallationMedia: Ubuntu 16.04.2 LTS "Xenial Xerus" - Release amd64 (20170215.2) PackageArchitecture: all SourcePackage: alsa-driver Symptom: audio Symptom_AlsaPlaybackTest: ALSA playback test through plughw:PCH failed Symptom_Card: Built-in Audio - HDA Intel PCH Symptom_DevicesInUse: USER PID ACCESS COMMAND /dev/snd/controlC1: gdm 1812 F.... pulseaudio lin 2199 F.... pulseaudio /dev/snd/controlC0: gdm 1812 F.... pulseaudio lin 2199 F.... pulseaudio Symptom_Jack: Speaker, Internal Symptom_Type: No sound at all Title: [P7xxDM2(-G), Realtek ALC898, Speaker, Internal] No sound at all UpgradeStatus: No upgrade log present (probably fresh install) dmi.bios.date: 12/21/2016 dmi.bios.vendor: American Megatrends Inc. dmi.bios.version: 1.06.04 dmi.board.asset.tag: Tag 12345 dmi.board.name: P7xxDM2(-G) dmi.board.vendor: Notebook dmi.board.version: Not Applicable dmi.chassis.asset.tag: No Asset Tag dmi.chassis.type: 10 dmi.chassis.vendor: No Enclosure dmi.chassis.version: N/A dmi.modalias: dmi:bvnAmericanMegatrendsInc.:bvr1.06.04:bd12/21/2016:svnNotebook:pnP7xxDM2(-G):pvrNotApplicable:rvnNotebook:rnP7xxDM2(-G):rvrNotApplicable:cvnNoEnclosure:ct10:cvrN/A: dmi.product.family: Not Applicable dmi.product.name: P7xxDM2(-G) dmi.product.version: Not Applicable dmi.sys.vendor: Notebook See the attachments below for the alsa-info.sh output, etc. Regards, Dennis.

2 4

[alsa-devel] [PATCH] ALSA: hda - Reduce the suspend time consumption for ALC256
by Takashi Iwai 19 Jan '18

19 Jan '18

ALC256 has its own quirk to override the shutup call, and it contains the COEF update for pulling down the headset jack control. Currently, the COEF update is called after clearing the headphone pin, and this seems triggering a stall of the codec communication, and results in a long delay over a second at suspend. A quick resolution is to swap the calls: at first with the COEF update, then clear the headphone pin. Fixes: 4a219ef8f370 ("ALSA: hda/realtek - Add ALC256 HP depop function") Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=198503 Reported-by: Paul Menzel <pmenzel(a)molgen.mpg.de> Cc: <stable(a)vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- sound/pci/hda/patch_realtek.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index 0004e282a837..23475888192b 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -3154,11 +3154,13 @@ static void alc256_shutup(struct hda_codec *codec) if (hp_pin_sense) msleep(85); + /* 3k pull low control for Headset jack. */ + /* NOTE: call this before clearing the pin, otherwise codec stalls */ + alc_update_coef_idx(codec, 0x46, 0, 3 << 12); + snd_hda_codec_write(codec, hp_pin, 0, AC_VERB_SET_PIN_WIDGET_CONTROL, 0x0); - alc_update_coef_idx(codec, 0x46, 0, 3 << 12); /* 3k pull low control for Headset jack. */ - if (hp_pin_sense) msleep(100); -- 2.15.1

1 0