linux/include/sound, branch v4.5

ALSA: hda - Loop interrupt handling until really cleared

2016-02-26T07:50:31Z

Currently the interrupt handler of HD-audio driver assumes that no irq update is needed while processing the irq. But in reality, it has been confirmed that the HW irq is issued even during the irq handling. Since we clear the irq status at the beginning, process the interrupt, then exits from the handler, the lately issued interrupt is left untouched without being properly processed. This patch changes the interrupt handler code to loop over the check-and-process. The handler tries repeatedly as long as the IRQ status are turned on, and either stream or CORB/RIRB is handled. For checking the stream handling, snd_hdac_bus_handle_stream_irq() returns a value indicating the stream indices bits. Other than that, the change is only in the irq handler itself. Reported-by: Libin Yang Cc: Signed-off-by: Takashi Iwai

ALSA: rawmidi: Make snd_rawmidi_transmit() race-free

2016-02-03T13:51:28Z

A kernel WARNING in snd_rawmidi_transmit_ack() is triggered by syzkaller fuzzer: WARNING: CPU: 1 PID: 20739 at sound/core/rawmidi.c:1136 Call Trace: [< inline >] __dump_stack lib/dump_stack.c:15 [] dump_stack+0x6f/0xa2 lib/dump_stack.c:50 [] warn_slowpath_common+0xd9/0x140 kernel/panic.c:482 [] warn_slowpath_null+0x29/0x30 kernel/panic.c:515 [] snd_rawmidi_transmit_ack+0x275/0x400 sound/core/rawmidi.c:1136 [] snd_virmidi_output_trigger+0x4b1/0x5a0 sound/core/seq/seq_virmidi.c:163 [< inline >] snd_rawmidi_output_trigger sound/core/rawmidi.c:150 [] snd_rawmidi_kernel_write1+0x549/0x780 sound/core/rawmidi.c:1223 [] snd_rawmidi_write+0x543/0xb30 sound/core/rawmidi.c:1273 [] __vfs_write+0x113/0x480 fs/read_write.c:528 [] vfs_write+0x167/0x4a0 fs/read_write.c:577 [< inline >] SYSC_write fs/read_write.c:624 [] SyS_write+0x111/0x220 fs/read_write.c:616 [] entry_SYSCALL_64_fastpath+0x16/0x7a arch/x86/entry/entry_64.S:185 Also a similar warning is found but in another path: Call Trace: [< inline >] __dump_stack lib/dump_stack.c:15 [] dump_stack+0x6f/0xa2 lib/dump_stack.c:50 [] warn_slowpath_common+0xd9/0x140 kernel/panic.c:482 [] warn_slowpath_null+0x29/0x30 kernel/panic.c:515 [] rawmidi_transmit_ack+0x24a/0x3b0 sound/core/rawmidi.c:1133 [] snd_rawmidi_transmit_ack+0x51/0x80 sound/core/rawmidi.c:1163 [] snd_virmidi_output_trigger+0x2b6/0x570 sound/core/seq/seq_virmidi.c:185 [< inline >] snd_rawmidi_output_trigger sound/core/rawmidi.c:150 [] snd_rawmidi_kernel_write1+0x4bb/0x760 sound/core/rawmidi.c:1252 [] snd_rawmidi_write+0x543/0xb30 sound/core/rawmidi.c:1302 [] __vfs_write+0x113/0x480 fs/read_write.c:528 [] vfs_write+0x167/0x4a0 fs/read_write.c:577 [< inline >] SYSC_write fs/read_write.c:624 [] SyS_write+0x111/0x220 fs/read_write.c:616 [] entry_SYSCALL_64_fastpath+0x16/0x7a arch/x86/entry/entry_64.S:185 In the former case, the reason is that virmidi has an open code calling snd_rawmidi_transmit_ack() with the value calculated outside the spinlock. We may use snd_rawmidi_transmit() in a loop just for consuming the input data, but even there, there is a race between snd_rawmidi_transmit_peek() and snd_rawmidi_tranmit_ack(). Similarly in the latter case, it calls snd_rawmidi_transmit_peek() and snd_rawmidi_tranmit_ack() separately without protection, so they are racy as well. The patch tries to address these issues by the following ways: - Introduce the unlocked versions of snd_rawmidi_transmit_peek() and snd_rawmidi_transmit_ack() to be called inside the explicit lock. - Rewrite snd_rawmidi_transmit() to be race-free (the former case). - Make the split calls (the latter case) protected in the rawmidi spin lock. BugLink: http://lkml.kernel.org/r/CACT4Y+YPq1+cYLkadwjWa5XjzF1_Vki1eHnVn-Lm0hzhSpu5PA@mail.gmail.com BugLink: http://lkml.kernel.org/r/CACT4Y+acG4iyphdOZx47Nyq_VHGbpJQK-6xNpiqUjaZYqsXOGw@mail.gmail.com Reported-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Cc: Signed-off-by: Takashi Iwai

ALSA: timer: Introduce disconnect op to snd_timer_instance

2016-01-21T16:51:42Z

Instead of the previous ugly hack, introduce a new op, disconnect, to snd_timer_instance object for handling the wake up of pending tasks more cleanly. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=109431 Signed-off-by: Takashi Iwai

ALSA: hda_intel: add card number to irq description

2016-01-12T20:05:16Z

Currently the info in /proc/interrupts doesn't allow to figure out which interrupt belongs to which card (HDMI, PCH, ..). Therefore add card details to the interrupt description. With the patch the info in /proc/interrupts looks like this: PCI-MSI 442368-edge snd_hda_intel:card1 PCI-MSI 49152-edge snd_hda_intel:card0 NOTE: this patch adds the new irq_descr field snd_card struct that is filled automatically at a card object creation. This can be used generically for other drivers as well. The changes for others will follow later -- tiwai Signed-off-by: Heiner Kallweit Signed-off-by: Takashi Iwai

Merge tag 'asoc-v4.4-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into for-next

2016-01-11T16:50:15Z

ASoC: Last updates for v4.5 A bunch more updates for v4.5, mainly driver work: - More topology API enhancements from Mengdong Lin working towards making everything more component based and being able to specify PCM links via topology. - Large sets driver updates from Cirrus, Intel (mainly more Skylake support) and Renesas. - New driver for AMD ACP - Rename PCM1792a driver to be generic pcm179x

Merge remote-tracking branches 'asoc/topic/adsp' and 'asoc/topic/amd' into asoc-next

2016-01-11T13:54:32Z

Merge remote-tracking branch 'asoc/topic/pcm-list' into asoc-next

2016-01-11T13:54:31Z

Merge tag 'asoc-v4.5' into asoc-next

2016-01-11T13:54:29Z

ASoC: Updates for v4.5 This is quite a busy release on the driver front with a lot of new drivers being added but comparatively quiet on the core side with only one big change going in and that a fairly straightforward refactoring. - Conversion of the array of DAI links to a list by Mengdong Lin, supporting dynamically adding and removing DAI links. - Some more fixes for the topology code, though it is still not final and ready for enabling in production. We really need to get to the point where that can be done. - A pile of changes for Intel SkyLake drivers which hopefully deliver some useful initial functionality for systems with this chipset, though there is more work still to come. - New drivers for a number of Imagination Technologies IPs. - Lots of new features and cleanups for the Renesas drivers. - ANC support for WM5110. - New driver for Atmel class D speaker drivers. - New drivers for Cirrus CS47L24 and WM1831. - New driver for Dialog DA7128. - New drivers for Realtek RT5659 and RT56156. - New driver for Rockchip RK3036. - New driver for TI PC3168A # gpg: Signature made Wed 23 Dec 2015 00:42:40 GMT using RSA key ID 5D5487D0 # gpg: Good signature from "Mark Brown " # gpg: aka "Mark Brown " # gpg: aka "Mark Brown " # gpg: aka "Mark Brown " # gpg: aka "Mark Brown " # gpg: aka "Mark Brown " # gpg: WARNING: This key is not certified with a trusted signature! # gpg: There is no indication that the signature belongs to the owner. # Primary key fingerprint: 3F25 68AA C269 98F9 E813 A1C5 C3F4 36CA 30F5 D8EB # Subkey fingerprint: ADE6 68AA 6757 18B5 9FE2 9FEA 24D6 8B72 5D54 87D0

Merge remote-tracking branch 'asoc/topic/intel' into asoc-next

2016-01-11T13:54:28Z

ASoC: Make aux_dev more like a generic component

2016-01-10T12:46:55Z

aux_dev is mainly used by the machine driver to specify analog devices, which are registered as codecs. Making it more like a generic component can help the machine driver to use it to specify any component with topology info by name. Details: - Remove the stub 'rtd_aux' array from the soc card. - Add a list 'aux_comp_list' to store the components of aux_devs. And add a list head 'list_aux' to struct snd_soc_component, for adding such components to the above list. - Add a 'init' ops to a component for machine specific init. soc_bind_aux_dev() will set it to be aux_dev's init. And it will be called when probing the component. - soc_bind_aux_dev() will also search components by name of an aux_dev, since it may not be a codec. - Move probing of aux_devs before checking new DAI links brought by topology. - Move removal of aux_devs later than removal of links. Because topology of aux components may register DAIs and the DAI drivers will go with removal of the aux components, we want soc_remove_link_dais() to remove the DAIs at first. Signed-off-by: Mengdong Lin Signed-off-by: Mark Brown