On Fri, Dec 03, 2021 at 10:37:07PM +0200, Andy Shevchenko wrote:
On Fri, Dec 03, 2021 at 07:46:04AM -0600, Pierre-Louis Bossart wrote:
On 12/3/21 4:13 AM, Dan Carpenter wrote:
Hello Vinod Koul,
The patch f533a035e4da: "ASoC: Intel: mrfld - create separate module for pci part" from Nov 4, 2014, leads to the following Smatch static checker warning:
...
Surely there should be a "return 0;" before the do_release_regions: label? How does this code work?
Thanks for reporting this Dan. Yes this doesn't look good at all.
This PCI part is only used on Merrifield/Tangier, and I am not sure if anyone ever managed to make audio work with the upstream version of this driver. It's my understanding that this platform is no longer maintained by Intel, and the Edison Yocto code uses the SOF driver.
The Kconfig updated in 2018 hints at those limitations:
config SND_SST_ATOM_HIFI2_PLATFORM_PCI tristate "PCI HiFi2 (Merrifield) Platforms" depends on X86 && PCI select SND_SST_ATOM_HIFI2_PLATFORM help If you have a Intel Merrifield/Edison platform, then enable this option by saying Y or m. Distros will typically not enable this option: while Merrifield/Edison can run a mainline kernel with limited functionality it will require a firmware file which is not in the standard firmware tree
I would guess that indeed a return 0; is missing, but maybe it's time to remove this PCI code completely. I can't think of any user of the PCI parts of this driver.
Andy, Hans, Mark, Takashi, what do you think?
The Edison platform and actually some more based on Intel Merrifield are still alive and on the (second hand) market. But yes, I would rather focus on making SOF working there, but via PCI bus (or with ACPI, ASL code for which one should actually write down, currently it's a device with PCI interface only).
That said, Pierre, have you been able to setup your Edison to see anything on I2S with SOF?