Re: [RFC PATCH 0/5] Extend FW framework for user FW uploads

[Luis Chamberlain]

On Thu, Feb 03, 2022 at 02:30:47PM -0700, Russ Weight wrote:
> Extend the firmware loader subsystem to support a persistent sysfs
> interface that userspace may use to initiate a firmware update. For
> example, FPGA based PCIe cards automatically load firmware and FPGA images
> from local FLASH when the card boots. The images in FLASH may be updated
> with new images that are uploaded by the user.
> 
> A device driver may call fw_upload_register() to expose persistent
> "loading" and "data" sysfs files at /sys/class/firmare/<NAME>/*. These
> files are used in the same way as the fallback sysfs "loading" and "data"
> files. However, when 0 is written to "loading" to complete the write of
> firmware data, the data is also transferred to the lower-level driver
> using pre-registered call-back functions. The data transfer is done in
> the context of a kernel worker thread.
> 
> Additional sysfs nodes are added in the same location as "loading" and
> "data" to monitor the transfer of the image data to the device using
> callback functions provided by the lower-level device driver and to allow
> the data transfer to be cancelled.
> 
> Example usage:
> 
> $ pwd
> /sys/class/firmware/n3000bmc-sec-update.8
> $ ls
> cancel  device  loading  remaining_size  subsystem
> data    error   power    status          uevent
> $ echo 1 > loading
> $ cat /tmp/firmware.bin > data
> $ echo 0 > loading
> $ while :; do cat status; cat remaining_size ; sleep 3; done
> preparing
> 44590080
> <--snip-->
> transferring
> 44459008
> transferring
> 44311552
> <--snip-->
> <snip>
> transferring
> 173056
> <--snip-->
> programming
> 0
> <--snip-->
> idle
> 0
> ^C

Nice. Please extend lib/test_firmware.c while at it and try to break
and ensure your stuff is not regressed with future changes.

  Luis