can't download Oceanic VT3 on linux
Jef Driesen
jef at libdivecomputer.org
Wed Mar 12 07:32:09 PDT 2014
On 2014-03-11 16:59, Linus Torvalds wrote:
> On Tue, Mar 11, 2014 at 6:27 AM, Jef Driesen <jef at libdivecomputer.org>
> wrote:
>>
>> I wonder if this is another case where the tcdrain() function isn't
>> waiting
>> properly until all bytes have been send out.
>
> Assuming it is ftdi, we should have at least *tried* to do it since
> 3.8, but I really won't guarantee that it works. I'm assuming Hamish
> has something newer than that, but who knows.. As usual, I think
> Debian "stable" (also known as "bat shit crazy") is some ancient 3.2
> kernel still, but all other distros should be much more up-to-date.
>
>> We had a similar problem with Suunto devices before.
>
> As you know (but perhaps others don't) this isn't a Suunto problem per
> se - the actual *real* legacy serial cable should work fine, because
> on a real UART we generally have an easy way to check whether the send
> buffer has actually flushed out.
>
> But the USB people originally (consciously) decided to not make USB ->
> serial be a standard, because their argument was that USB should
> replace serial lines, and USB serial dongles were thus "invalid".
> Morons with an agenda. As a result, there is no actual standard USB
> serial conversion model, and every single USB serial chip is a "hack"
> pretty much by definition.
>
> And with simplex being *so* unusual, it doesn't get a lot of attention
> and testing, and since it's not some generic standard, it's always
> per-dongle-chip-driver.
With USB being the Universal *Serial* Bus, it's kind of funny that one
thing it doesn't do very well is serial communication :-)
Isn't USB CDC-ACM supposed to be the standardized way for emulating
legacy serial communication over USB? There are a few dive computer
using it (Mares Icon HD and Dive Rite Nitek Q), but these are all using
a microchip internally. So although they do present themselves to the
outside as a legacy serial port, at no point there is any conversion to
real serial communication necessary (because that microchip speaks
native usb). So that's quite different from using a usb-serial
conversion chip from ftdi or prolific, which really has to emulate
serial communication on one end.
Jef
More information about the devel
mailing list