On 11/02/2011 12:42 PM, Stefan Berger wrote:
On 11/02/2011 02:00 PM, Eric Blake wrote:
libvirt.c guarantees that nparams is non-zero for scheduler parameters.
* src/test/test_driver.c (testDomainGetSchedulerParamsFlags): Drop
redundant check.
---
src/test/test_driver.c | 4 ----
1 files changed, 0 insertions(+), 4 deletions(-)
diff --git a/src/test/test_driver.c b/src/test/test_driver.c
index 4e1973a..de0f716 100644
--- a/src/test/test_driver.c
+++ b/src/test/test_driver.c
@@ -2709,10 +2709,6 @@ testDomainGetSchedulerParamsFlags(virDomainPtr
domain,
goto cleanup;
}
- if (*nparams< 1) {
- testError(VIR_ERR_INVALID_ARG, "%s", _("Invalid parameter count"));
- goto cleanup;
- }
Following the note on the top of this mail, this seems not even to be
possible (and hopefully never will be).
My recent commit 5465bc0 included explicit documentation in libvirt.c
that nparams can't be zero on virDomainGetSchedulerParams, fairly close
to the actual code in that file that guarantees that condition before
calling any driver callback.
strcpy(params[0].field, VIR_DOMAIN_SCHEDULER_WEIGHT);
Someone is still using strcpy!
At least this use was safe. But I went ahead and converted that to
virStrcpyStatic.
params[0].type = VIR_TYPED_PARAM_UINT;
/* XXX */
ACK
I've gone ahead and pushed this series.
--
Eric Blake eblake@xxxxxxxxxx +1-801-349-2682
Libvirt virtualization library http://libvirt.org
--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list