From: Antonio Borneo Date: Mon, 23 Jul 2018 13:20:40 +0000 (+0200) Subject: target/cortex_a: poll all targets in SMP node after halt X-Git-Url: https://git.sur5r.net/?a=commitdiff_plain;h=56cdc3f3d54ecef9f1847045c041a2d8e723c375;p=openocd target/cortex_a: poll all targets in SMP node after halt The periodic poll scans all the targets in the same order they have been declared in the configuration file. When one target in a SMP node halts, the transition is detected in the following poll and this triggers a halt request to all the other cores of the SMP node. The targets that will be polled afterwards will be identified as "halted", but the targets already scanned will remain as "running" until the next periodic poll. This creates a race condition with GDB; GDB sets the breakpoints when runs the target and removes them as soon as the target is halted. When it receives the halt event, it starts removing the breakpoints and fails on the targets that are still reported as "running". Fixed by polling all the targets in the SMP node before informing GDB about the halt event. This implementation is almost copy/paste from the one in aarch64. Change-Id: Id2bd99f1e56b014e48e9e34ccb891b4219c518f8 Signed-off-by: Antonio Borneo Reviewed-on: http://openocd.zylin.com/4622 Tested-by: jenkins Reviewed-by: Matthias Welwarsky --- diff --git a/src/target/cortex_a.c b/src/target/cortex_a.c index 4aae5e47..bc4aa33b 100644 --- a/src/target/cortex_a.c +++ b/src/target/cortex_a.c @@ -57,6 +57,9 @@ #include "transport/transport.h" #include +#define foreach_smp_target(pos, head) \ + for (pos = head; (pos != NULL); pos = pos->next) + static int cortex_a_poll(struct target *target); static int cortex_a_debug_entry(struct target *target); static int cortex_a_restore_context(struct target *target, bool bpwp); @@ -806,12 +809,43 @@ static int cortex_a_halt_smp(struct target *target) static int update_halt_gdb(struct target *target) { + struct target *gdb_target = NULL; + struct target_list *head; + struct target *curr; int retval = 0; + if (target->gdb_service && target->gdb_service->core[0] == -1) { target->gdb_service->target = target; target->gdb_service->core[0] = target->coreid; retval += cortex_a_halt_smp(target); } + + if (target->gdb_service) + gdb_target = target->gdb_service->target; + + foreach_smp_target(head, target->head) { + curr = head->target; + /* skip calling context */ + if (curr == target) + continue; + if (!target_was_examined(curr)) + continue; + /* skip targets that were already halted */ + if (curr->state == TARGET_HALTED) + continue; + /* Skip gdb_target; it alerts GDB so has to be polled as last one */ + if (curr == gdb_target) + continue; + + /* avoid recursion in cortex_a_poll() */ + curr->smp = 0; + cortex_a_poll(curr); + curr->smp = 1; + } + + /* after all targets were updated, poll the gdb serving target */ + if (gdb_target != NULL && gdb_target != target) + cortex_a_poll(gdb_target); return retval; }