Re: [libvirt] [PATCHv3 10/12] qemu: conf: Add support for memory device cold(un)plug
On 03/18/2015 01:51 AM, Zhu Guihua wrote:
On 03/17/2015 10:19 PM, Peter Krempa wrote:
> Add a few helpers that allow to operate with memory device definitions
> on the domain config and use them to implement memory device coldplug in
> the qemu driver.
> ---
>
> Notes:
> Version 2:
> - no changes
>
> src/conf/domain_conf.c | 100
> +++++++++++++++++++++++++++++++++++++++++++++++
> src/conf/domain_conf.h | 10 +++++
> src/libvirt_private.syms | 4 ++
> src/qemu/qemu_driver.c | 15 ++++++-
> 4 files changed, 127 insertions(+), 2 deletions(-)
>
> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
> index 8c2234f..1a02e46 100644
> --- a/src/conf/domain_conf.c
> +++ b/src/conf/domain_conf.c
> @@ -12811,6 +12811,106 @@ virDomainRNGRemove(virDomainDefPtr def,
> }
>
>
> +static int
> +virDomainMemoryFindByDefInternal(virDomainDefPtr def,
> + virDomainMemoryDefPtr mem,
> + bool allowAddressFallback)
> +{
> + size_t i;
> +
> + for (i = 0; i < def->nmems; i++) {
> + virDomainMemoryDefPtr tmp = def->mems[i];
> +
> + /* address, if present */
> + if (allowAddressFallback) {
> + if (tmp->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)
> + continue;
> + } else {
> + if (mem->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE &&
> + !virDomainDeviceInfoAddressIsEqual(&tmp->info,
> &mem->info))
> + continue;
> + }
> +
> + /* alias, if present */
> + if (mem->info.alias &&
> + STRNEQ_NULLABLE(tmp->info.alias, mem->info.alias))
> + continue;
> +
> + /* target info -> always present */
> + if (tmp->model != mem->model ||
> + tmp->targetNode != mem->targetNode ||
> + tmp->size != mem->size)
I have tested your series with our qemu memory hot remove patch series,
here would be a possible error.
When hotplug a memory device, its size has been aligned. So the compare for
size here would fail possiblely.
hmm.. Not sure that's necessary - although Peter can make the final
determination... Commit id '57b215a' doesn't modify each def->mems[i]
entry in qemuDomainAlignMemorySizes, rather it gets a value from
virDomainDefSetMemoryInitial and then does the rounding.
If the stored def->mems[i]->size value is/was modified, then I'd agree,
but it doesn't appear to be that way.
If there is a rounding of the value - then please just point it out
John
Thanks,
Zhu
> + continue;
> +
> + /* source stuff -> match with device */
> + if (tmp->pagesize != mem->pagesize)
> + continue;
> +
> + if (!virBitmapEqual(tmp->sourceNodes, mem->sourceNodes))
> + continue;
> +
> + break;
> + }
> +
> + if (i == def->nmems)
> + return -1;
> +
> + return i;
> +}
> +
> +
> +int
> +virDomainMemoryFindByDef(virDomainDefPtr def,
> + virDomainMemoryDefPtr mem)
> +{
> + return virDomainMemoryFindByDefInternal(def, mem, false);
> +}
> +
[...]
--
libvir-list mailing list
libvir-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list