4.8.1 planning

[Linus Torvalds]

[ I notice that I lost the subsurface list, because Miika just sent
the strace to me in private. So when I replied with a test patch, it
only went to Miika and Dirk ]

On Thu, Jul 12, 2018 at 1:31 PM Linus Torvalds
<torvalds at linux-foundation.org> wrote:
>
> So considering that this clearly works for *some* people, maybe it's
> literally some timing issue.

.. and looking around in libdivecomputer history, this *used* do be
dc_serial_sleep(), and that just did a nanosleep() internally - and we
never overrode it in our custom serial implementation, because why
would we?

So as libdivecomputer switched "dc_serial_sleep()" to
"dc_iostream_sleep()", we subtly lost the sleep entirely unless the
custom IO implemented it.

In other words, my test patch looks like it really could be very
relevant. Any divecomputer that actually depended on the delay
suddenly got none.

So here's the patch attached again, this time for the whole list in
case somebody else can test it too (eg Anton).

Still entirely untested, but it *did* compile at least once by pure
luck. And this time I added the ftdi case too - which also didn't have
the actual sleep callback, although it did implement the function.

So this could have affected ftdi in addition to bluetooth.

                   Linus
-------------- next part --------------
A non-text attachment was scrubbed...
Name: patch.diff
Type: text/x-patch
Size: 2362 bytes
Desc: not available
URL: <http://lists.subsurface-divelog.org/pipermail/subsurface/attachments/20180712/cd27eb38/attachment.bin>