Commit 334c388f ("pflash_cfi: Error out if device length
isn't a power of two") aimed to finish the effort started by
commit 06f15217 ("pflash: Require backend size to match device,
improve errors"), but unfortunately we are not quite there since
various machines are still ready to accept incomplete / oversized
pflash backend images, and now fail, i.e. on Debian bullseye:

 $ qemu-system-x86_64 \
   -drive \
   if=pflash,format=raw,unit=0,readonly=on,file=/usr/share/OVMF/OVMF_CODE.fd
 qemu-system-x86_64: Device size must be a power of two.

where OVMF_CODE.fd comes from the ovmf package, which doesn't
pad the firmware images to the flash size:

 $ ls -lh /usr/share/OVMF/
 -rw-r--r-- 1 root root 3.5M Aug 19  2021 OVMF_CODE_4M.fd
 -rw-r--r-- 1 root root 1.9M Aug 19  2021 OVMF_CODE.fd
 -rw-r--r-- 1 root root 128K Aug 19  2021 OVMF_VARS.fd

Since we entered the freeze period to prepare the v7.2.0 release,
the safest is to revert commit 334c388f.

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1294


Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221108175755.95141-1-philmd@linaro.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20221108172633.860700-1-danielhb413@gmail.com>