On Tue, Aug 15, 2017 at 4:55 AM, Richard Hughes <hughsient@gmail.com> wrote:

Hi all,

I'm the maintainer of fwupd, which is a daemon for doing firmware
updates in Linux. Using fwupd about 200,000 people update firmware
every month. At the moment I have an out-of-tree patch to use the
flashrom CLI for updating coreboot on Purism laptops, which exec's a
flashrom binary with the correct arguments and then screen scrapes the
output. This is less than ideal. I saw the appearance of a libflashrom
in the staging branch and got very excited. I'm assuming the long term
goal here is to install a shared with stable API/ABI library and a
pkg-config file. If so, please keep reading.

One thing that is really important for fwupd is progress completion;
as flashing the firmware is an inherently scary thing to do users do
like to see a progress bar move from 0% to 100% in a gradual way as
it's very re-assuring. This is even more important for the GUI, where
people that just see a spinner don't know if the process is going to
take 10 seconds, or 10 minutes.

Agreed! Especially since a lot of users have large chips, slow programmers, or sometimes both. Simple read operations can take a worryingly long time on some common hardware these days.

Would it be possible to have a process callback for
flashrom_image_read(), flashrom_image_write() and
flashrom_image_verify()? It'd need to have some kind of callback
userdata too, for instance:

int flashrom_image_verify(struct flashrom_flashctx *, const void
*buffer, size_t buffer_len, flashrom_progress_callback callback_fn,
void *callback_user_data);

and also:

typedef void(flashrom_progress_callback)(uint64 completed, unit64
total, void *user_data)

The short answer is yes. Nico also briefly mentioned an idea on IRC to use a mechanism similar to flashrom_set_log_callback() to avoid complicating function signatures for the read, write, and verify functions.

I'm not familiar at all with flashrom internals, but could try to
produce a patch if required -- although to propagate the progress
through operations would mean patching a lot of chip->read() callbacks
which probably requires a more experienced hand. Ideas welcome,
thanks.

Richard.

_______________________________________________
flashrom mailing list
flashrom@flashrom.org
https://mail.coreboot.org/mailman/listinfo/flashrom