[libvirt] [PATCH] qemu: Fix compilation error when enum variable size differs from 'int'
Since commit bcd9a564b631aa virDomainNumatuneGetMode returns the value via a pointer rather than in the return value. The change triggered problems with platforms where the compiler decides to use a data type of size different than integer at the point where we typecast it. Work around the issue by using an intermediate variable of the correct type that gets casted back by the default typecasting rules. --- src/qemu/qemu_driver.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index aa0acde..2b9f125 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -10566,14 +10566,16 @@ qemuDomainGetNumaParameters(virDomainPtr dom, virMemoryParameterPtr param = ¶ms[i]; switch (i) { - case 0: /* fill numa mode here */ + case 0: { /* fill numa mode here */ + virDomainNumatuneMemMode tmp; + virDomainNumatuneGetMode(def->numa, -1, &tmp); + if (virTypedParameterAssign(param, VIR_DOMAIN_NUMA_MODE, - VIR_TYPED_PARAM_INT, 0) < 0) + VIR_TYPED_PARAM_INT, tmp) < 0) goto cleanup; - virDomainNumatuneGetMode(def->numa, -1, - (virDomainNumatuneMemMode *) ¶m->value.i); break; + } case 1: /* fill numa nodeset here */ nodeset = virDomainNumatuneFormatNodeset(def->numa, NULL, -1); -- 2.4.1
On 25.05.2015 17:18, Peter Krempa wrote:
Since commit bcd9a564b631aa virDomainNumatuneGetMode returns the value via a pointer rather than in the return value. The change triggered problems with platforms where the compiler decides to use a data type of size different than integer at the point where we typecast it.
Work around the issue by using an intermediate variable of the correct type that gets casted back by the default typecasting rules. --- src/qemu/qemu_driver.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index aa0acde..2b9f125 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -10566,14 +10566,16 @@ qemuDomainGetNumaParameters(virDomainPtr dom, virMemoryParameterPtr param = ¶ms[i];
switch (i) { - case 0: /* fill numa mode here */ + case 0: { /* fill numa mode here */ + virDomainNumatuneMemMode tmp; + virDomainNumatuneGetMode(def->numa, -1, &tmp); + if (virTypedParameterAssign(param, VIR_DOMAIN_NUMA_MODE, - VIR_TYPED_PARAM_INT, 0) < 0) + VIR_TYPED_PARAM_INT, tmp) < 0) goto cleanup;
- virDomainNumatuneGetMode(def->numa, -1, - (virDomainNumatuneMemMode *) ¶m->value.i); break; + }
case 1: /* fill numa nodeset here */ nodeset = virDomainNumatuneFormatNodeset(def->numa, NULL, -1);
I guess we saw the same coverity report since you're fixing the code I'm just looking at. But I think what is coverity trying to tell us is that in all other places virDomainNumatuneGetMode() is checked for return value, here it's not. Or did you really see problem you're describing in the commit message? Michal
On Mon, May 25, 2015 at 17:35:19 +0200, Michal Privoznik wrote:
On 25.05.2015 17:18, Peter Krempa wrote:
Since commit bcd9a564b631aa virDomainNumatuneGetMode returns the value via a pointer rather than in the return value. The change triggered problems with platforms where the compiler decides to use a data type of size different than integer at the point where we typecast it.
Work around the issue by using an intermediate variable of the correct type that gets casted back by the default typecasting rules. --- src/qemu/qemu_driver.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index aa0acde..2b9f125 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -10566,14 +10566,16 @@ qemuDomainGetNumaParameters(virDomainPtr dom, virMemoryParameterPtr param = ¶ms[i];
switch (i) { - case 0: /* fill numa mode here */ + case 0: { /* fill numa mode here */ + virDomainNumatuneMemMode tmp; + virDomainNumatuneGetMode(def->numa, -1, &tmp); + if (virTypedParameterAssign(param, VIR_DOMAIN_NUMA_MODE, - VIR_TYPED_PARAM_INT, 0) < 0) + VIR_TYPED_PARAM_INT, tmp) < 0) goto cleanup;
- virDomainNumatuneGetMode(def->numa, -1, - (virDomainNumatuneMemMode *) ¶m->value.i); break; + }
case 1: /* fill numa nodeset here */ nodeset = virDomainNumatuneFormatNodeset(def->numa, NULL, -1);
I guess we saw the same coverity report since you're fixing the code I'm just looking at. But I think what is coverity trying to tell us is that in all other places virDomainNumatuneGetMode() is checked for return value, here it's not. Or did you really see problem you're describing in the commit message?
The problem I'm trying to solve is failure to compile libvirt on CentOS 5. The compilation error can be seen at: https://ci.centos.org/view/libvirt-project/job/libvirt-daemon-build/systems=... if you scroll down enough. Peter
On 25.05.2015 17:47, Peter Krempa wrote:
On Mon, May 25, 2015 at 17:35:19 +0200, Michal Privoznik wrote:
On 25.05.2015 17:18, Peter Krempa wrote:
Since commit bcd9a564b631aa virDomainNumatuneGetMode returns the value via a pointer rather than in the return value. The change triggered problems with platforms where the compiler decides to use a data type of size different than integer at the point where we typecast it.
Work around the issue by using an intermediate variable of the correct type that gets casted back by the default typecasting rules. --- src/qemu/qemu_driver.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index aa0acde..2b9f125 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -10566,14 +10566,16 @@ qemuDomainGetNumaParameters(virDomainPtr dom, virMemoryParameterPtr param = ¶ms[i];
switch (i) { - case 0: /* fill numa mode here */ + case 0: { /* fill numa mode here */ + virDomainNumatuneMemMode tmp; + virDomainNumatuneGetMode(def->numa, -1, &tmp); + if (virTypedParameterAssign(param, VIR_DOMAIN_NUMA_MODE, - VIR_TYPED_PARAM_INT, 0) < 0) + VIR_TYPED_PARAM_INT, tmp) < 0) goto cleanup;
- virDomainNumatuneGetMode(def->numa, -1, - (virDomainNumatuneMemMode *) ¶m->value.i); break; + }
case 1: /* fill numa nodeset here */ nodeset = virDomainNumatuneFormatNodeset(def->numa, NULL, -1);
I guess we saw the same coverity report since you're fixing the code I'm just looking at. But I think what is coverity trying to tell us is that in all other places virDomainNumatuneGetMode() is checked for return value, here it's not. Or did you really see problem you're describing in the commit message?
The problem I'm trying to solve is failure to compile libvirt on CentOS 5. The compilation error can be seen at:
https://ci.centos.org/view/libvirt-project/job/libvirt-daemon-build/systems=...
if you scroll down enough.
Ah, okay; So ACK. Lets see what will Coverity think after you push this. Michal
On Mon, May 25, 2015 at 18:28:42 +0200, Michal Privoznik wrote:
On 25.05.2015 17:47, Peter Krempa wrote:
On Mon, May 25, 2015 at 17:35:19 +0200, Michal Privoznik wrote:
On 25.05.2015 17:18, Peter Krempa wrote:
Since commit bcd9a564b631aa virDomainNumatuneGetMode returns the value via a pointer rather than in the return value. The change triggered problems with platforms where the compiler decides to use a data type of size different than integer at the point where we typecast it.
Work around the issue by using an intermediate variable of the correct type that gets casted back by the default typecasting rules. --- src/qemu/qemu_driver.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index aa0acde..2b9f125 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -10566,14 +10566,16 @@ qemuDomainGetNumaParameters(virDomainPtr dom, virMemoryParameterPtr param = ¶ms[i];
switch (i) { - case 0: /* fill numa mode here */ + case 0: { /* fill numa mode here */ + virDomainNumatuneMemMode tmp; + virDomainNumatuneGetMode(def->numa, -1, &tmp); + if (virTypedParameterAssign(param, VIR_DOMAIN_NUMA_MODE, - VIR_TYPED_PARAM_INT, 0) < 0) + VIR_TYPED_PARAM_INT, tmp) < 0) goto cleanup;
- virDomainNumatuneGetMode(def->numa, -1, - (virDomainNumatuneMemMode *) ¶m->value.i); break; + }
case 1: /* fill numa nodeset here */ nodeset = virDomainNumatuneFormatNodeset(def->numa, NULL, -1);
I guess we saw the same coverity report since you're fixing the code I'm just looking at. But I think what is coverity trying to tell us is that in all other places virDomainNumatuneGetMode() is checked for return value, here it's not. Or did you really see problem you're describing in the commit message?
The problem I'm trying to solve is failure to compile libvirt on CentOS 5. The compilation error can be seen at:
https://ci.centos.org/view/libvirt-project/job/libvirt-daemon-build/systems=...
if you scroll down enough.
Ah, okay; So ACK. Lets see what will Coverity think after you push this.
Ah, I see what coverity might complain about. The value will not be initialized if neither the specific nor the default mode is specified in the XML. I'll send a v2 fixing both the coverity issue and the compilation problem since this one would not fix coverity as it would be possible to have @tmp uninitialized. Peter
On Mon, May 25, 2015 at 05:35:19PM +0200, Michal Privoznik wrote:
On 25.05.2015 17:18, Peter Krempa wrote:
Since commit bcd9a564b631aa virDomainNumatuneGetMode returns the value via a pointer rather than in the return value. The change triggered problems with platforms where the compiler decides to use a data type of size different than integer at the point where we typecast it.
Work around the issue by using an intermediate variable of the correct type that gets casted back by the default typecasting rules. --- src/qemu/qemu_driver.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index aa0acde..2b9f125 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -10566,14 +10566,16 @@ qemuDomainGetNumaParameters(virDomainPtr dom, virMemoryParameterPtr param = ¶ms[i];
switch (i) { - case 0: /* fill numa mode here */ + case 0: { /* fill numa mode here */ + virDomainNumatuneMemMode tmp; + virDomainNumatuneGetMode(def->numa, -1, &tmp); + if (virTypedParameterAssign(param, VIR_DOMAIN_NUMA_MODE, - VIR_TYPED_PARAM_INT, 0) < 0) + VIR_TYPED_PARAM_INT, tmp) < 0) goto cleanup;
- virDomainNumatuneGetMode(def->numa, -1, - (virDomainNumatuneMemMode *) ¶m->value.i); break; + }
case 1: /* fill numa nodeset here */ nodeset = virDomainNumatuneFormatNodeset(def->numa, NULL, -1);
I guess we saw the same coverity report since you're fixing the code I'm just looking at. But I think what is coverity trying to tell us is that in all other places virDomainNumatuneGetMode() is checked for return value, here it's not. Or did you really see problem you're describing in the commit message?
I think Peter is trying to fix the problem of compilation on centos 5: https://ci.centos.org/view/libvirt-project/job/libvirt-daemon-build/systems=... And I already had similar patch prepared, I only declared the "tmp" as "memMode" for the whole function because switch cases with curly brackets -- Yuck! I'd say ACK, but I'm not sure whether Michal was trying to point out something more that I misunderstood, so I leave it up to him.
participants (3)
-
Martin Kletzander -
Michal Privoznik -
Peter Krempa