[Qemu-devel] [PATCH] hw/ppc/spapr_pci.c: Avoid functions not in glib 2.12 (g_hash_table_iter_*)

[Qemu-devel] [PATCH] hw/ppc/spapr_pci.c: Avoid functions not in glib 2.12 (g_hash_table_iter_*)

[Peter Maydell]

The g_hash_table_iter_* functions for iterating through a hash table
are not present in glib 2.12, which is our current minimum requirement.
Rewrite the code to use g_hash_table_foreach() instead.

Cc: qemu-stable@nongnu.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
I have tested that this builds with a glib 2.12, and also that it
passes 'make check', but no further testing beyond that. Somebody
with an spapr migration test should check it doesn't break things...

The observant will note that since this is fixing breakage
introduced in commit 9a321e92343 (merged in late June) we
obviously released 2.1 in a "doesn't build all targets on
glib 2.12" state, and nobody actually complained... So there's
maybe scope for debate about moving the minimum version up,
but I think for 2.2 we should stick with the current definition.

The cc:stable is in the interests of fixing that build breakage
in 2.1.x.

 hw/ppc/spapr_pci.c | 28 +++++++++++++++++-----------
 1 file changed, 17 insertions(+), 11 deletions(-)

diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c
index ad0da7f..21b95b3 100644
--- a/hw/ppc/spapr_pci.c
+++ b/hw/ppc/spapr_pci.c
@@ -704,28 +704,34 @@ static const VMStateDescription vmstate_spapr_pci_msi = {
     },
 };
 
+static void spapr_pci_fill_msi_devs(gpointer key, gpointer value,
+                                    gpointer opaque)
+{
+    sPAPRPHBState *sphb = opaque;
+
+    sphb->msi_devs[sphb->msi_devs_num].key = *(uint32_t *)key;
+    sphb->msi_devs[sphb->msi_devs_num].value = *(spapr_pci_msi *)value;
+    sphb->msi_devs_num++;
+}
+
 static void spapr_pci_pre_save(void *opaque)
 {
     sPAPRPHBState *sphb = opaque;
-    GHashTableIter iter;
-    gpointer key, value;
-    int i;
+    int msi_devs_num;
 
     if (sphb->msi_devs) {
         g_free(sphb->msi_devs);
         sphb->msi_devs = NULL;
     }
-    sphb->msi_devs_num = g_hash_table_size(sphb->msi);
-    if (!sphb->msi_devs_num) {
+    sphb->msi_devs_num = 0;
+    msi_devs_num = g_hash_table_size(sphb->msi);
+    if (!msi_devs_num) {
         return;
     }
-    sphb->msi_devs = g_malloc(sphb->msi_devs_num * sizeof(spapr_pci_msi_mig));
+    sphb->msi_devs = g_malloc(msi_devs_num * sizeof(spapr_pci_msi_mig));
 
-    g_hash_table_iter_init(&iter, sphb->msi);
-    for (i = 0; g_hash_table_iter_next(&iter, &key, &value); ++i) {
-        sphb->msi_devs[i].key = *(uint32_t *) key;
-        sphb->msi_devs[i].value = *(spapr_pci_msi *) value;
-    }
+    g_hash_table_foreach(sphb->msi, spapr_pci_fill_msi_devs, sphb);
+    assert(sphb->msi_devs_num == msi_devs_num);
 }
 
 static int spapr_pci_post_load(void *opaque, int version_id)
-- 
1.9.1

Re: [Qemu-devel] [PATCH] hw/ppc/spapr_pci.c: Avoid functions not in glib 2.12 (g_hash_table_iter_*)

[Peter Maydell]

Ping! It would be nice to be able to get glib2.12 builds fixed...

thanks
-- PMM

On 22 October 2014 18:41, Peter Maydell <peter.maydell@linaro.org> wrote:
> The g_hash_table_iter_* functions for iterating through a hash table
> are not present in glib 2.12, which is our current minimum requirement.
> Rewrite the code to use g_hash_table_foreach() instead.
>
> Cc: qemu-stable@nongnu.org
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> ---
> I have tested that this builds with a glib 2.12, and also that it
> passes 'make check', but no further testing beyond that. Somebody
> with an spapr migration test should check it doesn't break things...
>
> The observant will note that since this is fixing breakage
> introduced in commit 9a321e92343 (merged in late June) we
> obviously released 2.1 in a "doesn't build all targets on
> glib 2.12" state, and nobody actually complained... So there's
> maybe scope for debate about moving the minimum version up,
> but I think for 2.2 we should stick with the current definition.
>
> The cc:stable is in the interests of fixing that build breakage
> in 2.1.x.
>
>  hw/ppc/spapr_pci.c | 28 +++++++++++++++++-----------
>  1 file changed, 17 insertions(+), 11 deletions(-)
>
> diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c
> index ad0da7f..21b95b3 100644
> --- a/hw/ppc/spapr_pci.c
> +++ b/hw/ppc/spapr_pci.c
> @@ -704,28 +704,34 @@ static const VMStateDescription vmstate_spapr_pci_msi = {
>      },
>  };
>
> +static void spapr_pci_fill_msi_devs(gpointer key, gpointer value,
> +                                    gpointer opaque)
> +{
> +    sPAPRPHBState *sphb = opaque;
> +
> +    sphb->msi_devs[sphb->msi_devs_num].key = *(uint32_t *)key;
> +    sphb->msi_devs[sphb->msi_devs_num].value = *(spapr_pci_msi *)value;
> +    sphb->msi_devs_num++;
> +}
> +
>  static void spapr_pci_pre_save(void *opaque)
>  {
>      sPAPRPHBState *sphb = opaque;
> -    GHashTableIter iter;
> -    gpointer key, value;
> -    int i;
> +    int msi_devs_num;
>
>      if (sphb->msi_devs) {
>          g_free(sphb->msi_devs);
>          sphb->msi_devs = NULL;
>      }
> -    sphb->msi_devs_num = g_hash_table_size(sphb->msi);
> -    if (!sphb->msi_devs_num) {
> +    sphb->msi_devs_num = 0;
> +    msi_devs_num = g_hash_table_size(sphb->msi);
> +    if (!msi_devs_num) {
>          return;
>      }
> -    sphb->msi_devs = g_malloc(sphb->msi_devs_num * sizeof(spapr_pci_msi_mig));
> +    sphb->msi_devs = g_malloc(msi_devs_num * sizeof(spapr_pci_msi_mig));
>
> -    g_hash_table_iter_init(&iter, sphb->msi);
> -    for (i = 0; g_hash_table_iter_next(&iter, &key, &value); ++i) {
> -        sphb->msi_devs[i].key = *(uint32_t *) key;
> -        sphb->msi_devs[i].value = *(spapr_pci_msi *) value;
> -    }
> +    g_hash_table_foreach(sphb->msi, spapr_pci_fill_msi_devs, sphb);
> +    assert(sphb->msi_devs_num == msi_devs_num);
>  }
>
>  static int spapr_pci_post_load(void *opaque, int version_id)
> --
> 1.9.1

Re: [Qemu-devel] [PATCH] hw/ppc/spapr_pci.c: Avoid functions not in glib 2.12 (g_hash_table_iter_*)

[Alexander Graf]

On 30.10.14 13:33, Peter Maydell wrote:
> Ping! It would be nice to be able to get glib2.12 builds fixed...

Thanks, applied to ppc-next.


Alex

Re: [Qemu-devel] [PATCH] hw/ppc/spapr_pci.c: Avoid functions not in glib 2.12 (g_hash_table_iter_*)

[Alexey Kardashevskiy]

On 11/03/2014 03:31 PM, Alexander Graf wrote:
> 
> 
> On 30.10.14 13:33, Peter Maydell wrote:
>> Ping! It would be nice to be able to get glib2.12 builds fixed...
> 
> Thanks, applied to ppc-next.

Uff. I misread the original mail and decided I have to do the patch :)
I checked ppc-next, works fine, thanks!


-- 
Alexey