Another thing to verify would be how can a variable in a group have the offset 172, when the group size is calculated to be 172 (probably something to do with pragma).

To check what's going on, I added another integer cfg variable to the last group. The memory mapping this time (without using pragma blocks) is as follows:

START
block address: 0x7fde5c80fa08
vars address: 0x7fde5c80fa0c
Total allocated size: 1168
Group number: 1
Group location from start 20
Group size: 248
Group number: 2
Group location from start 284
Group size: 56
Group number: 3
Group location from start 356
Group size: 4
Group number: 4
Group location from start 380
Group size: 4
Group number: 5
Group location from start 404
Group size: 4
Group number: 6
Group location from start 428
Group size: 176
Group number: 7
Group location from start 620
Group size: 104
Group number: 8
Group location from start 740
Group size: 200
Group number: 9
Group location from start 956
Group size: 16
Group number: 10
Group location from start 988
Group size: 180

#For the new variable added 

(gdb) print var->offset
$3 = 176
(gdb) print group
$4 = (cfg_group_t *) 0x7fdee1bca108
(gdb) print group->var_offset
$5 = 984

So, this time, the offset and group size make sense.

According to me, using pragma blocks would create more problems. The next thing to investigate would be this inconsistency between group size and variable offset.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.