Re: How are watchpoints handled? (remote serial)

[Juha Aaltonen <turbopultti at gmail dot com>]

Maybe that's a bug in gdb-multiarch 7.7.1?
Remote-debugging an ARM (watch g_testloc):

(This is as it should be.)
gdbarch_dump: have_nonsteppable_watchpoint = 1


w \x00]0x00008180 in __delay_169 () at ../blinky.c:211
211                g_testloc = 1;
(gdb) cont
Continuing.
Sending packet:
$m81b0,4#c8...[\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][
r +]Ack
[$][0][0][0][0][0][0][0][0][#][8][0]Packet received: 00000000
[
w \x00]Sending packet:
$Z2,81b0,4#13...[\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][
r +]Ack
[$][O][K][#][9][a]Packet received: OK
[
w \x00]Sending packet: $c#63...[\x00][\x00][\x00][\x00][\x00][
r +]Ack
[$][T][0][5][#][b][9]Packet received: T05
[
w \x00]Sending packet: $g#67...[\x00][\x00][\x00][\x00][\x00][
r +]Ack
[$][8][8][1][3][0][0][0][0][2][c][0][1][0][0][0][0][8][8][1][3][0][0][0][0][0][1][0][0][0][0][0][0][0][a][0][0][0][0][0][0][0][3][0][0][0][0][0][0][e][8][0][3][0][0][0][0][b][0][8][1][0][0][0][0][2][c][0][a][0][2][1][f][2][3][0][0][0][0][0][0][2][8][0][a][0][2][1][f][2][4][0][a][0][2][1][f][4][3][0][c][0][1][1][f][a][0][9][7][0][0][0][0][8][0][8][1][0][0][0][0][8][0][8][1][0][0][0][0][0][9][0][0][0][0][9][0][0][9][0][0][0][0][9][0][0][9][0][0][0][0][9][0][0][9][0][0][0][0][9][0][0][9][0][0][0][0][9][0][0][9][0][0][0][0][9][0][0][9][0][0][0][0][9][0][0][9][0][0][0][0][9][0][0][9][0][0][0][0][9][0][9][7][0][1][0][0][6][0][#][2][4]Packet
received: 881300002c01000088130000010000000a00000003000000e8030000b08100002c0a021f23000000280a021f240a021f430c011fa0970000808100008081000009000090090000900900009009000090090000900900009009000090090000900900009097010060
[
w \x00]Sending packet:
$m8180,4#9e...[\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][
r +]Ack
[$][0][0][3][0][8][7][e][5][#][c][c]Packet received: 003087e5
[
... (lots of memory reads) ...
[
w \x00]Sending packet:
$m8180,4#9e...[\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][
r +]Ack
[$][0][0][3][0][8][7][e][5][#][c][c]Packet received: 003087e5
[
w \x00]
Program received signal SIGTRAP, Trace/breakpoint trap.
Sending packet:
$z2,81b0,4#33...[\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][\x00][
r +]Ack
[$][O][K][#][9][a]Packet received: OK
[
w \x00]0x00008180 in __delay_169 () at ../blinky.c:211
211                g_testloc = 1;
(gdb)


No stepping over watchpoint.


On Tue, Nov 10, 2015 at 12:57 AM, Juha Aaltonen <turbopultti@gmail.com> wrote:
> How should the watchpoints be handled by remote stub/server?
> It seems that when 'cont'-command is given, gdb re-installs watchpoint
> and sends 'c'-packet.
>
> If the return address is the trap address, the watchpoint trigs again,
> but if the address is advanced,
> the access is not done at all, because the watchpoint makes exception
> before the access is made.