Re: [PATCH 6/9] eal: initialize vdevs right next to pci devices

On Wed, Feb 10, 2016 at 12:05 PM, Jan Viktorin [off-list ref] wrote:

On Fri, 29 Jan 2016 15:08:33 +0100
David Marchand [off-list ref] wrote:

quoted

--- a/lib/librte_eal/common/include/rte_dev.h
+++ b/lib/librte_eal/common/include/rte_dev.h

@@ -156,7 +156,7 @@ void rte_eal_driver_register(struct rte_driver *driver);
 void rte_eal_driver_unregister(struct rte_driver *driver);

 /**
- * Initalize all the registered drivers in this process
+ * Scan all devargs and attach to drivers if available
  */
 int rte_eal_dev_init(void);

Move this to a separate commit? Is it just a forgotten doc comment?

Should be in previous commit, yes.

quoted

diff --git a/lib/librte_eal/linuxapp/eal/eal.c b/lib/librte_eal/linuxapp/eal/eal.c
index 62241ee..95313af 100644
--- a/lib/librte_eal/linuxapp/eal/eal.c
+++ b/lib/librte_eal/linuxapp/eal/eal.c

@@ -834,8 +834,10 @@ rte_eal_init(int argc, char **argv)
              rte_config.master_lcore, (int)thread_id, cpuset,
              ret == 0 ? "" : "...");

+#ifndef RTE_NEXT_ABI
      if (rte_eal_dev_init() < 0)
              rte_panic("Cannot init pmd devices\n");
+#endif

      RTE_LCORE_FOREACH_SLAVE(i) {

@@ -873,6 +875,11 @@ rte_eal_init(int argc, char **argv)
      rte_eal_mp_remote_launch(sync_func, NULL, SKIP_MASTER);
      rte_eal_mp_wait_lcore();

+#ifdef RTE_NEXT_ABI
+     if (rte_eal_dev_init() < 0)
+             rte_panic("Cannot probe vdev devices\n");
+#endif
+
      /* Probe & Initialize PCI devices */
      if (rte_eal_pci_probe())
              rte_panic("Cannot probe PCI\n");

I cannot see the point why it is enclosed in the RTE_NEXT_ABI. Is it
such a serious breakage?

No.
We can make this move unconditional.

Thanks.

-- 
David Marchand