fw_upload's poll_complete() is really intended for use with
asynchronous write() implementations - or at least those where the
write() loop may terminate without the kernel yet being aware of whether
or not the firmware upload has succeeded. For auto-update, write() is
only ever called once and will only return when uploading has completed,
be that by passing or failing. The core fw_upload code only calls
poll_complete() after the final call to write() has returned.

However, the poll_complete() implementation in the auto-update driver
was written to expect poll_complete() to be called from another context,
and it waits for a completion signalled from write(). Since
poll_complete() is actually called from the same context, after the
write() loop has terminated, wait_for_completion() never sees the
completion get signalled and always times out, causing programming to
always report a failing.

Since write() is full synchronous, and its return value will indicate
whether or not programming passed or failed, poll_complete() serves no
purpose and can be cut down to simply return FW_UPLOAD_ERR_NONE.

Cc: stable@vger.kernel.org
Fixes: ec5b0f1193ad4 ("firmware: microchip: add PolarFire SoC Auto Update support")
Reported-by: Jamie Gibbons <jamie.gibbons@microchip.com>
Tested-by: Jamie Gibbons <jamie.gibbons@microchip.com>
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>

After successfully programming the SPI flash with an MFPS auto update
image, the error sysfs attribute reports programming:timeout.
This is caused by an incorrect check on the return value from
wait_for_completion_timeout() in mpfs_auto_update_poll_complete().

Fixes: ec5b0f1193ad ("firmware: microchip: add PolarFire SoC Auto Update support")
Signed-off-by: Steve Wilkins <steve.wilkins@raymarine.com>
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>

There's a bunch of structs created and freed every time the mailbox is
used. Move them to use the scope-based cleanup infrastructure to avoid
manually tearing them down. mpfs_auto_update_available() didn't free the
memory that it used (albeit it allocated exactly once during probe) so
that gets moved over too.

Signed-off-by: Conor Dooley <conor.dooley@microchip.com>

This buffer is used exclusively by mpfs_auto_update_set_image_address(),
so move the management of it there, employing the recently added cleanup
infrastructure to avoid littering the function with gotos.

Signed-off-by: Conor Dooley <conor.dooley@microchip.com>

Updating the FPGA image might bring with it changes visible to Linux,
so it is helpful to also co-locate dt-overlays that describe the new
image contents. If these are packaged in a specific format [1] (detected
by first 4 bytes being MCHP, since FPGA images have no magic), load
the file to the reserved 1 MiB region immediately after the directory in
flash. The Beagle-V Fire's "gateware" already creates these files and
puts them in flash [2].

Link: https://github.com/polarfire-soc/polarfire-soc-documentation/blob/master/how-to/re-programming-the-fpga-from-linux.md#main-header-format
Link: https://openbeagle.org/beaglev-fire/gateware/-/blob/main/gateware_scripts/generate_gateware_overlays.py?ref_type=heads [2]
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>

As it says on the tin. It can be kinda confusing when "22830" is in hex,
so prefix the hex numbers with a "0x".

Signed-off-by: Conor Dooley <conor.dooley@microchip.com>

If validation fails, both prints are made. Skip the success one in the
failure case.

Fixes: ec5b0f1193ad ("firmware: microchip: add PolarFire SoC Auto Update support")
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>

RISC-V firmware drivers for v6.9

A single minor fix for an oversized allocation due to sizeof() misuse by
yours truly that came in since I sent my last fixes PR.

Signed-off-by: Conor Dooley <conor.dooley@microchip.com>

* tag 'riscv-firmware-for-v6.9' of https://git.kernel.org/pub/scm/linux/kernel/git/conor/linux:
  firmware: microchip: Fix over-requested allocation size

Link: https://lore.kernel.org/r/20240305-vicinity-dumpling-8943ef26f004@spud
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

cocci warnings: (new ones prefixed by >>)
>> drivers/firmware/microchip/mpfs-auto-update.c:387:72-78:
   ERROR: application of sizeof to pointer
   drivers/firmware/microchip/mpfs-auto-update.c:170:72-78:
   ERROR: application of sizeof to pointer

response_msg is a pointer to u32, so the size of element it points to is
supposed to be a multiple of sizeof(u32), rather than sizeof(u32 *).

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202403040516.CYxoWTXw-lkp@intel.com/
Signed-off-by: Dawei Li <dawei.li@shingroup.cn>
Fixes: ec5b0f1193ad ("firmware: microchip: add PolarFire SoC Auto Update support")
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>

response_msg is a pointer to an unsigned int (u32). So passing just
response_msg to sizeof would not print the size of the variable. To get
the size of response_msg we need to pass it as a pointer variable.

Fixes: ec5b0f1193ad ("firmware: microchip: add PolarFire SoC Auto Update support")
Signed-off-by: Samasth Norway Ananda <samasth.norway.ananda@oracle.com>
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>