Re: [Xen-devel] xenstored memory leak

[David Vrabel <david.vrabel@xxxxxxxxxx>]

On 13/07/16 14:32, Juergen Gross wrote:
> On 13/07/16 15:17, David Vrabel wrote:
>> On 13/07/16 14:07, Wei Liu wrote:
>>>
>>> My gut feeling is that xenstored shouldn't have the knowledge to
>>> associate a watch with a "process". The concept of a process is only
>>> meaningful to OS, which wouldn't work on cross-domain xenstored setup.
>>> Maybe the OS xenbus driver should reap all watches on behalf the dead
>>> process. This would also avoid a crashed QEMU leaking resources.
>>
>> The Linux driver already cleans up open transactions and removes watches
>> when the file handle is released.
> 
> Hmm, does this work reliably? I observed a memory leak of about 5kB per
> create/destroy domain pair with xenstored _and_ with xenstore domain.

Don't know.

I only took a quick look at the xenbus_file_release() function and it
looked like it ought to be doing the right thing...

David

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel