mail archive of the rauc mailing list
 help / color / mirror / Atom feed
From: Enrico Joerns <ejo@pengutronix.de>
To: "Middelschulte, Leif" <Leif.Middelschulte@klsmartin.com>
Cc: "rauc@pengutronix.de" <rauc@pengutronix.de>
Subject: Re: [RAUC] Progress indication
Date: Tue, 16 May 2017 11:21:02 +0200	[thread overview]
Message-ID: <b93d8481-7aa0-1539-9825-2c87952fe585@pengutronix.de> (raw)
In-Reply-To: <1494498580.6958.1.camel@klsmartin.com>

Hi Leif,

On 05/11/2017 12:29 PM, Middelschulte, Leif wrote:
> Hello Enrico,
>
> Am Mittwoch, den 10.05.2017, 15:22 +0200 schrieb Enrico Joerns:
>> Hi Leif,
>>
>> On 05/04/2017 04:33 PM, Middelschulte, Leif wrote:
>>>
>>> Now, about my second question: How does a custom hook script
>>> communicate its progress back to the install-handler (i.e. rauc
>>> service)?
>>>
>>> I'd go modify rauc's update_handler.c, set the
>>> G_SUBPROCESS_FLAGS_STDOUT_PIPE flag for the spawned slot
>>> script-process and read back the progress and propagate it to the
>>> DBus interface using r_context_send_progress. Is that the right
>>> direction here?
>>
>> in RAUC, the progress is generated by adding a 'step' for an action
>> you perform. When starting the action you call
>> r_context_begin_step(), when terminating the action you call
>> r_context_end_step(). It is important to know that steps have a
>> nesting depth that allows visualizers to control the granularity of
>> information they show. A step can also be marked as either
>> successful or failed.
>>
>>
>> An example is the "Determining slot states" step:
>>
>> r_context_begin_step("determine_slot_states", "Determining slot
>> states", 0); [...] r_context_end_step("determine_slot_states",
>> res);
>>
>>
>> Thus, this should be the interface to use to add new steps.
>>
>>
>> The current granularity we have does not go deeper than "Copying
>> image". This will be called for both a default install handler as
>> well as for an install hook.
>>
>> If you ask for progress from the install hook, you intend to be
>> more fine-grained with your messages?
> Yes, since some updates will take quite a while, we'd like to display
> an approximate progress (e.g. bytes written/byte to be written).
>
> I saw that there's a "send_progress" callback one can use from
> *within* the installer.

Not sure if I understand you right, but this is what I described above, 
isn't it? A `git grep send_progress` only gives me 
`r_context_send_progress`.

> But currently it's not even used to indicate any progress of raw (stream) copies, right?

No, there is currently no progress indication for raw stream copies. For
this it would work, but most other copy operations performed by binaries
(such as nandwrite) do not provide any progress, thus we did not put too
much effort on it, yet.

> I'm trying to stick to upstream code as much as possible and,
> eventually, contribute what's necessary to communicate progress
> back.

That sounds like the perfect way for me :)

> If I understood the code right, I'd go and do the following:

> 1.) make my hook script write its percentual progress to stdout
> 2.) I'd fork RAUC Service's update_handler code to read back STDOUT from the
> spawned's process (i.e. hook.sh)
 > 3.) use send_progress to publish the information to the "dumb" GUI

Basically, yes, but what you should use for indicating progress (had to 
dig in the code for that, too ;) ) is `r_context_set_step_percentage()`. 
You can read from its documentation string that its designed for such 
purposes you have, where you have a step that you want to give a number 
of nonspecific substeps. Thus like a copy progress inside a `installing 
slot` step.

There you can provide percentage steps that will also propagate to the 
global progress, meaning that you will have something like

   50% Updating rootfs0...
   51% Updating rootfs0...
   52% Updating rootfs0...
   53% Updating rootfs0...
   ...

Unfortunately, the only place in the code where this is used, already is 
the "/progress/test_explicit_percentage" test in test/progress.c.

Hope that helps you.


Best regards, Enrico

-- 
Pengutronix e.K.                           | Enrico Jörns                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-5080 |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |


_______________________________________________
RAUC mailing list

      reply	other threads:[~2017-05-16  9:21 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-04-26 16:03 Middelschulte, Leif
2017-05-04 10:29 ` Enrico Joerns
2017-05-04 14:33   ` Middelschulte, Leif
2017-05-10 13:22     ` Enrico Joerns
2017-05-11 10:29       ` Middelschulte, Leif
2017-05-16  9:21         ` Enrico Joerns [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=b93d8481-7aa0-1539-9825-2c87952fe585@pengutronix.de \
    --to=ejo@pengutronix.de \
    --cc=Leif.Middelschulte@klsmartin.com \
    --cc=rauc@pengutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox