Subject: Re: [PATCH] xen/balloon: Mark unallocated host memory as UNUSABLE
To: Boris Ostrovsky <boris.ostrovsky@oracle.com>, xen-devel@lists.xen.org,
        linux-kernel@vger.kernel.org
Cc: helgaas@kernel.org, christian.koenig@amd.com
References: <1513119066-10748-1-git-send-email-boris.ostrovsky@oracle.com>
 <14852f9f-f234-3226-f32b-3080ca114332@suse.com>
 <a8495745-b56a-83ff-993f-a745c8edf2c0@oracle.com>
 <1385e130-7300-66a4-3995-838244600746@suse.com>
 <9f7c7d7b-e8a7-d851-472d-972a815bf8e2@oracle.com>
From: Juergen Gross <jgross@suse.com>
Message-ID: <4cd0eb4c-9dd3-37db-9d34-42ffb93952fd@suse.com>
Date: Fri, 15 Dec 2017 16:33:40 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.5.0
MIME-Version: 1.0
In-Reply-To: <9f7c7d7b-e8a7-d851-472d-972a815bf8e2@oracle.com>
Content-Type: text/plain; charset=utf-8
Content-Language: de-DE
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1551
Lines: 40

On 15/12/17 15:58, Boris Ostrovsky wrote:
> On 12/15/2017 09:47 AM, Juergen Gross wrote:
>> On 15/12/17 15:24, Boris Ostrovsky wrote:
>>>>> +
>>>>> +	hostmem_resource->start = max_addr;
>>>>> +	hostmem_resource->end = entry->addr + entry->size;
>>>>> +	for (; i < memmap.nr_entries; i++) {
>>>>> +		entry = &xen_e820_table->entries[i];
>>>>> +		if (entry->type == E820_TYPE_RAM)
>>>> Shouldn't that be != ?
>>> No, the idea here is to populate hostmem_resource with ranges already
>>> taken by things other than RAM, leaving memory regions as available for
>>> the balloon hotplug. This will allow us to use allocate_resource(),
>>> which searches for a free range, in the balloon driver.
>> But why says the comment "Host memory not allocated to dom0" then?
> 
> hostmem_resource is created starting from the end of dom0 RAM and ending
> at the end of host RAM. If these two are the same (or if the former is
> larger, which I don't think is possible) then there is nothing to do, as
> the hostmem_resource will be empty. That's what the comment is referring
> to.
> 
>> And why are you trying to allocate from this resource in case of
>> hotplugging memory (and fall back to iomem_resource in case of
>> failure)?
> 
> Because that area (end of dom0 RAM through end of host RAM) is not going
> to be used by anyone else and thus is available. That was the idea
> behind f5775e0b6116.

Aah, now I've got it.

Could you please add a comment like:

/* Mark non-RAM regions as not available. */

above the test? That would have helped me. :-)


Juergen