Re: ntdll: Store full line status register in the internal structure, not just the TIOCSER_TEMT bit. Take 2.

[Alexandre Julliard]

Dmitry Timoshkov  writes:
> Printing complete lsr value in the log may help with diagnosing
> test failures on Alexandre's machine.

I don't see how, since TIOCSER_TEMT is the only defined bit. Do you
actually have a driver that returns other status?

ntdll: Store full line status register in the internal structure, not just the TIOCSER_TEMT bit. Take 2.

[Dmitry Timoshkov]

Alexandre Julliard  wrote:
> > Printing complete lsr value in the log may help with diagnosing
> > test failures on Alexandre's machine.
>
> I don't see how, since TIOCSER_TEMT is the only defined bit. Do you
> actually have a driver that returns other status?

I don't, but maybe yours does. The only way to check that is to store
and print the whole LSR value.

Do you happen to have something connected to COM-port (perhaps mouse?)
which may explain test failures.

ntdll: Store full line status register in the internal structure, not just the TIOCSER_TEMT bit. Take 2.

[Alexandre Julliard]

Dmitry Timoshkov  writes:
> Alexandre Julliard  wrote:
> > > Printing complete lsr value in the log may help with diagnosing
> > > test failures on Alexandre's machine.
> >
> > I don't see how, since TIOCSER_TEMT is the only defined bit. Do you
> > actually have a driver that returns other status?
>
> I don't, but maybe yours does. The only way to check that is to store
> and print the whole LSR value.

Or you could check the kernel source and see what TIOCSERGETLSR does...
> Do you happen to have something connected to COM-port (perhaps mouse?)
> which may explain test failures.

No.

ntdll: Store full line status register in the internal structure, not just the TIOCSER_TEMT bit. Take 2.

[Dmitry Timoshkov]

Alexandre Julliard  wrote:
> > > > Printing complete lsr value in the log may help with diagnosing
> > > > test failures on Alexandre's machine.
> > >
> > > I don't see how, since TIOCSER_TEMT is the only defined bit. Do you
> > > actually have a driver that returns other status?
> >
> > I don't, but maybe yours does. The only way to check that is to store
> > and print the whole LSR value.
>
> Or you could check the kernel source and see what TIOCSERGETLSR does...

Of course I did check the kernel source, and looked around in various
mailing lists, which made me think that your kernel may set UART_LSR_TEMT
instead...

ntdll: Store full line status register in the internal structure, not just the TIOCSER_TEMT bit. Take 2.

[Alexandre Julliard]

Dmitry Timoshkov  writes:
> Alexandre Julliard  wrote:
> > > > > Printing complete lsr value in the log may help with diagnosing
> > > > > test failures on Alexandre's machine.
> > > >
> > > > I don't see how, since TIOCSER_TEMT is the only defined bit. Do you
> > > > actually have a driver that returns other status?
> > >
> > > I don't, but maybe yours does. The only way to check that is to store
> > > and print the whole LSR value.
> >
> > Or you could check the kernel source and see what TIOCSERGETLSR does...
>
> Of course I did check the kernel source, and looked around in various
> mailing lists, which made me think that your kernel may set UART_LSR_TEMT
> instead...

If you find a place in the kernel that does this, it's a bug. And no, my
kernel returns TIOCSER_TEMT, like everybody else's.