[Bug remote/17028] GDB+GDBserver hangs on Windows waiting for stop event since target-async on by default

["brobecker at adacore dot com" <sourceware-bugzilla at sourceware dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=17028

--- Comment #3 from brobecker at adacore dot com ---
> I have to say I'm quite surprised by this, given remote non-stop (+async) has
> been known to work on Windows hosts for a long while.

I don't know, never tried it before.

One of the differences that "set debug serial 1" seems to highlight
is the fact that, in target-async mode, GDB seems to be making
adjustments to the serial descriptor. I see things like this
in the output (the "[fd468->...]" notifications).

| Sending packet: $vCont;s:2930#c0...[
| w $][v][C][o][n][t][;][s][:][2][9][3][0][#][c][0][fd468->asynchronous]
| [fd468->fd-scheduled]
| [fd468->timer-scheduled]
| [
| r
$][T][0][5][0][5][:][c][8][f][e][2][8][0][0][;][0][4][:][a][0][f][e][2][8][0][0][;][0][8][:][1][4][1][6][4][0][0][0][;][t][h][r][e][a][d][:][2][9][3][0][;][#][e][fd468->fd-scheduled]
| [a]Packet received (fd-468):
T0505:c8fe2800;04:a0fe2800;08:14164000;thread:2930;

On a vCont that does not work, I see:

| Sending packet: $vCont;s:114c;c#89...[
| w $][v][C][o][n][t][;][s][:][1][1][4][c][;][c][#][8][9][fd468->asynchronous]
| [fd468->fd-scheduled]

Note the absence of changing the state to timer-scheduled.

I don't know this code much. You? I am not sure whether this code
is in fact relevant to this issue or not; I just started looking
at it because I wanted to find a way to get to the file descriptor
number for the remote connection.

-- 
You are receiving this mail because:
You are on the CC list for the bug.