[PATCH v5 5/7] iflib.c: Deactivate use of ifc_cpus
Jan Sommer
jan.sommer at dlr.de
Fri Apr 17 08:24:06 UTC 2020
- cpusets and SMP are currently not supported in libbsd for RTEMS
- Disable the ifc_cpus context variable and replace its usage,
essentially hard-coding for cpu 0
---
freebsd/sys/net/iflib.c | 24 ++++++++++++++++++++++++
1 file changed, 24 insertions(+)
diff --git a/freebsd/sys/net/iflib.c b/freebsd/sys/net/iflib.c
index 89bc5e79..73521478 100644
--- a/freebsd/sys/net/iflib.c
+++ b/freebsd/sys/net/iflib.c
@@ -161,7 +161,9 @@ struct iflib_ctx {
device_t ifc_dev;
if_t ifc_ifp;
+#ifndef __rtems__
cpuset_t ifc_cpus;
+#endif /* __rtems__ */
if_shared_ctx_t ifc_sctx;
struct if_softc_ctx ifc_softc_ctx;
@@ -1279,6 +1281,7 @@ prefetch2cachelines(void *x)
#define prefetch2cachelines(x)
#endif
+#ifndef __rtems__
static void
iflib_gen_mac(if_ctx_t ctx)
{
@@ -1310,6 +1313,7 @@ iflib_gen_mac(if_ctx_t ctx)
mac[4] = digest[1];
mac[5] = digest[2];
}
+#endif /* __rtems__ */
static void
iru_init(if_rxd_update_t iru, iflib_rxq_t rxq, uint8_t flid)
@@ -4448,6 +4452,7 @@ get_ctx_core_offset(if_ctx_t ctx)
qc = max(scctx->isc_ntxqsets, scctx->isc_nrxqsets);
mtx_lock(&cpu_offset_mtx);
+#ifndef __rtems__
SLIST_FOREACH(op, &cpu_offsets, entries) {
if (CPU_CMP(&ctx->ifc_cpus, &op->set) == 0) {
ret = op->offset;
@@ -4457,6 +4462,7 @@ get_ctx_core_offset(if_ctx_t ctx)
break;
}
}
+#endif /* __rtems__ */
if (ret == CORE_OFFSET_UNSPECIFIED) {
ret = 0;
op = malloc(sizeof(struct cpu_offset), M_IFLIB,
@@ -4467,7 +4473,9 @@ get_ctx_core_offset(if_ctx_t ctx)
} else {
op->offset = qc;
op->refcount = 1;
+#ifndef __rtems__
CPU_COPY(&ctx->ifc_cpus, &op->set);
+#endif /* __rtems__ */
SLIST_INSERT_HEAD(&cpu_offsets, op, entries);
}
}
@@ -4483,7 +4491,11 @@ unref_ctx_core_offset(if_ctx_t ctx)
mtx_lock(&cpu_offset_mtx);
SLIST_FOREACH_SAFE(op, &cpu_offsets, entries, top) {
+#ifndef __rtems__
if (CPU_CMP(&ctx->ifc_cpus, &op->set) == 0) {
+#else /* __rtems__ */
+ {
+#endif /* __rtems__ */
MPASS(op->refcount > 0);
op->refcount--;
if (op->refcount == 0) {
@@ -4594,12 +4606,14 @@ iflib_device_register(device_t dev, void *sc, if_shared_ctx_t sctx, if_ctx_t *ct
taskqgroup_attach(qgroup_if_config_tqg, &ctx->ifc_admin_task, ctx,
-1, "admin");
+#ifndef __rtems__
/* Set up cpu set. If it fails, use the set of all CPUs. */
if (bus_get_cpus(dev, INTR_CPUS, sizeof(ctx->ifc_cpus), &ctx->ifc_cpus) != 0) {
device_printf(dev, "Unable to fetch CPU list\n");
CPU_COPY(&all_cpus, &ctx->ifc_cpus);
}
MPASS(CPU_COUNT(&ctx->ifc_cpus) > 0);
+#endif /* __rtems__ */
/*
** Now set up MSI or MSI-X, should return us the number of supported
@@ -5427,7 +5441,9 @@ iflib_queues_alloc(if_ctx_t ctx)
txq->ift_br_offset = 0;
}
/* XXX fix this */
+#ifndef __rtems__
txq->ift_timer.c_cpu = cpu;
+#endif /* __rtems__ */
if (iflib_txsd_alloc(txq)) {
device_printf(dev, "Critical Failure setting up TX buffers\n");
@@ -6233,11 +6249,19 @@ iflib_msix_init(if_ctx_t ctx)
#else
queues = queuemsgs;
#endif
+#ifndef __rtems__
queues = imin(CPU_COUNT(&ctx->ifc_cpus), queues);
+#else /* __rtems__ */
+ queues = imin(1, queues);
+#endif /* __rtems__ */
if (bootverbose)
device_printf(dev,
"intr CPUs: %d queue msgs: %d admincnt: %d\n",
+#ifndef __rtems__
CPU_COUNT(&ctx->ifc_cpus), queuemsgs, admincnt);
+#else /* __rtems__ */
+ 1, queuemsgs, admincnt);
+#endif /* __rtems__ */
#ifdef RSS
/* If we're doing RSS, clamp at the number of RSS buckets */
if (queues > rss_getnumbuckets())
--
2.12.3
More information about the devel
mailing list