[Qemu-devel] [PATCH 0/2] ppc: Allow "-prom-env" parameter for the pseries machine, too

[Qemu-devel] [PATCH 0/2] ppc: Allow "-prom-env" parameter for the pseries machine, too

[Thomas Huth]

Here's a slightly updated version of the patch that enables the "-prom-env"
parameter for the pseries machine. The code is now only executed if
nb_prom_envs > 0. That means if the user did _not_ specify a "-prom-env"
parameter, there is no change in behavior (i.e. SLOF will initialize the
NVRAM instead). This fixes the problem with grub2 not working because it does
not find the "/options/real-mode?" property in the device tree.

If you currently want to use the "-prom-env" parameter and grub2 together,
you simply have to specify -prom-env "real-mode?=true" to get it working.
This will be fixed by a later update to SLOF, so that you then do not have to
specify this explicitely anymore.

The second patch adds a test case for the -prom-env parameter on the
pseries machine.

The patches have to be applied on top of David's ppc-for-2.8 branch since
they rely on the chrp_nvram rework patches that are queued there.

Thomas Huth (2):
  spapr_nvram: Pre-initialize the NVRAM to support the -prom-env
    parameter
  tests: Add pseries machine to the prom-env-test, too

 hw/nvram/spapr_nvram.c |  6 ++++++
 tests/prom-env-test.c  | 23 ++++++++++++++---------
 2 files changed, 20 insertions(+), 9 deletions(-)

-- 
1.8.3.1

[Qemu-devel] [PATCH 1/2] spapr_nvram: Pre-initialize the NVRAM to support the -prom-env parameter

[Thomas Huth]

In case we do not load the NVRAM contents from a file and the user
specified the "-prom-env" parameter, use the new CHRP NVRAM helper
functions to pre-initialize the NVRAM partitions, so that the SLOF
firmware now can pick up the environment variables from the -prom-env
parameter, too.

Signed-off-by: Thomas Huth <thuth@redhat.com>
---
 hw/nvram/spapr_nvram.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/hw/nvram/spapr_nvram.c b/hw/nvram/spapr_nvram.c
index 4de5f70..eb42ea3 100644
--- a/hw/nvram/spapr_nvram.c
+++ b/hw/nvram/spapr_nvram.c
@@ -31,6 +31,7 @@
 #include "sysemu/block-backend.h"
 #include "sysemu/device_tree.h"
 #include "hw/sysbus.h"
+#include "hw/nvram/chrp_nvram.h"
 #include "hw/ppc/spapr.h"
 #include "hw/ppc/spapr_vio.h"
 
@@ -162,6 +163,11 @@ static void spapr_nvram_realize(VIOsPAPRDevice *dev, Error **errp)
             error_setg(errp, "can't read spapr-nvram contents");
             return;
         }
+    } else if (nb_prom_envs > 0) {
+        /* Create a system partition to pass the -prom-env variables */
+        chrp_nvram_create_system_partition(nvram->buf, MIN_NVRAM_SIZE / 4);
+        chrp_nvram_create_free_partition(&nvram->buf[MIN_NVRAM_SIZE / 4],
+                                         nvram->size - MIN_NVRAM_SIZE / 4);
     }
 
     spapr_rtas_register(RTAS_NVRAM_FETCH, "nvram-fetch", rtas_nvram_fetch);
-- 
1.8.3.1

[Qemu-devel] [PATCH 2/2] tests: Add pseries machine to the prom-env-test, too

[Thomas Huth]

Now that we also support the "-prom-env" parameter for the pseries
machine, we can enable this test for this machine, too. Since booting
with TCG is rather slow with the pseries machine, we also enable
the "-nodefaults" parameter for this test now, so that SLOF does not
have to check that much devices during boot and thus runs a little
bit faster.

Signed-off-by: Thomas Huth <thuth@redhat.com>
---
 tests/prom-env-test.c | 23 ++++++++++++++---------
 1 file changed, 14 insertions(+), 9 deletions(-)

diff --git a/tests/prom-env-test.c b/tests/prom-env-test.c
index 7a62857..6f70fe2 100644
--- a/tests/prom-env-test.c
+++ b/tests/prom-env-test.c
@@ -9,11 +9,12 @@
  * This work is licensed under the terms of the GNU GPL, version 2
  * or later. See the COPYING file in the top-level directory.
  *
- * This test is used to check that some OpenBIOS machines can be started
- * successfully in TCG mode. To do this, we first put some Forth code into
- * the "boot-command" Open Firmware environment variable. This Forth code
- * writes a well-known magic value to a known location in memory. Then we
- * start the guest so that OpenBIOS can boot and finally run the Forth code.
+ * This test is used to check that some Open Firmware based machines (i.e.
+ * OpenBIOS or SLOF) can be started successfully in TCG mode. To do this, we
+ * first put some Forth code into the "boot-command" Open Firmware environment
+ * variable. This Forth code writes a well-known magic value to a known location
+ * in memory. Then we start the guest so that the firmware can boot and finally
+ * run the Forth code.
  * The testing code here then can finally check whether the value has been
  * successfully written into the guest memory.
  */
@@ -45,7 +46,8 @@ static void test_machine(const void *machine)
 {
     char *args;
 
-    args = g_strdup_printf("-M %s,accel=tcg -prom-env 'boot-command=%x %x l!'",
+    args = g_strdup_printf("-M %s,accel=tcg -nodefaults "
+                           "-prom-env 'boot-command=%x %x l!'",
                            (const char *)machine, MAGIC, ADDRESS);
 
     qtest_start(args);
@@ -71,13 +73,16 @@ int main(int argc, char *argv[])
 {
     const char *sparc_machines[] = { "SPARCbook", "Voyager", "SS-20", NULL };
     const char *sparc64_machines[] = { "sun4u", "sun4v", NULL };
-    const char *mac_machines[] = { "mac99", "g3beige", NULL };
+    const char *ppc_machines[] = { "mac99", "g3beige", NULL };
+    const char *ppc64_machines[] = { "mac99", "g3beige", "pseries", NULL };
     const char *arch = qtest_get_arch();
 
     g_test_init(&argc, &argv, NULL);
 
-    if (!strcmp(arch, "ppc") || !strcmp(arch, "ppc64")) {
-        add_tests(mac_machines);
+    if (!strcmp(arch, "ppc")) {
+        add_tests(ppc_machines);
+    } else if (!strcmp(arch, "ppc64")) {
+        add_tests(ppc64_machines);
     } else if (!strcmp(arch, "sparc")) {
         add_tests(sparc_machines);
     } else if (!strcmp(arch, "sparc64")) {
-- 
1.8.3.1

Re: [Qemu-devel] [PATCH 0/2] ppc: Allow "-prom-env" parameter for the pseries machine, too

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1612 bytes --]

On Wed, Oct 26, 2016 at 09:35:39AM +0200, Thomas Huth wrote:
> Here's a slightly updated version of the patch that enables the "-prom-env"
> parameter for the pseries machine. The code is now only executed if
> nb_prom_envs > 0. That means if the user did _not_ specify a "-prom-env"
> parameter, there is no change in behavior (i.e. SLOF will initialize the
> NVRAM instead). This fixes the problem with grub2 not working because it does
> not find the "/options/real-mode?" property in the device tree.
> 
> If you currently want to use the "-prom-env" parameter and grub2 together,
> you simply have to specify -prom-env "real-mode?=true" to get it working.
> This will be fixed by a later update to SLOF, so that you then do not have to
> specify this explicitely anymore.
> 
> The second patch adds a test case for the -prom-env parameter on the
> pseries machine.
> 
> The patches have to be applied on top of David's ppc-for-2.8 branch since
> they rely on the chrp_nvram rework patches that are queued there.

Applied to ppc-for-2.8, replacing earlier versions.

> 
> Thomas Huth (2):
>   spapr_nvram: Pre-initialize the NVRAM to support the -prom-env
>     parameter
>   tests: Add pseries machine to the prom-env-test, too
> 
>  hw/nvram/spapr_nvram.c |  6 ++++++
>  tests/prom-env-test.c  | 23 ++++++++++++++---------
>  2 files changed, 20 insertions(+), 9 deletions(-)
> 

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]