[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH v2] xen/page_alloc: Simplify domain_adjust_tot_pages


  • To: Alejandro Vallejo <alejandro.vallejo@xxxxxxxxx>
  • From: Jan Beulich <jbeulich@xxxxxxxx>
  • Date: Wed, 12 Mar 2025 09:21:07 +0100
  • Autocrypt: addr=jbeulich@xxxxxxxx; keydata= xsDiBFk3nEQRBADAEaSw6zC/EJkiwGPXbWtPxl2xCdSoeepS07jW8UgcHNurfHvUzogEq5xk hu507c3BarVjyWCJOylMNR98Yd8VqD9UfmX0Hb8/BrA+Hl6/DB/eqGptrf4BSRwcZQM32aZK 7Pj2XbGWIUrZrd70x1eAP9QE3P79Y2oLrsCgbZJfEwCgvz9JjGmQqQkRiTVzlZVCJYcyGGsD /0tbFCzD2h20ahe8rC1gbb3K3qk+LpBtvjBu1RY9drYk0NymiGbJWZgab6t1jM7sk2vuf0Py O9Hf9XBmK0uE9IgMaiCpc32XV9oASz6UJebwkX+zF2jG5I1BfnO9g7KlotcA/v5ClMjgo6Gl MDY4HxoSRu3i1cqqSDtVlt+AOVBJBACrZcnHAUSuCXBPy0jOlBhxPqRWv6ND4c9PH1xjQ3NP nxJuMBS8rnNg22uyfAgmBKNLpLgAGVRMZGaGoJObGf72s6TeIqKJo/LtggAS9qAUiuKVnygo 3wjfkS9A3DRO+SpU7JqWdsveeIQyeyEJ/8PTowmSQLakF+3fote9ybzd880fSmFuIEJldWxp Y2ggPGpiZXVsaWNoQHN1c2UuY29tPsJgBBMRAgAgBQJZN5xEAhsDBgsJCAcDAgQVAggDBBYC AwECHgECF4AACgkQoDSui/t3IH4J+wCfQ5jHdEjCRHj23O/5ttg9r9OIruwAn3103WUITZee e7Sbg12UgcQ5lv7SzsFNBFk3nEQQCACCuTjCjFOUdi5Nm244F+78kLghRcin/awv+IrTcIWF hUpSs1Y91iQQ7KItirz5uwCPlwejSJDQJLIS+QtJHaXDXeV6NI0Uef1hP20+y8qydDiVkv6l IreXjTb7DvksRgJNvCkWtYnlS3mYvQ9NzS9PhyALWbXnH6sIJd2O9lKS1Mrfq+y0IXCP10eS FFGg+Av3IQeFatkJAyju0PPthyTqxSI4lZYuJVPknzgaeuJv/2NccrPvmeDg6Coe7ZIeQ8Yj t0ARxu2xytAkkLCel1Lz1WLmwLstV30g80nkgZf/wr+/BXJW/oIvRlonUkxv+IbBM3dX2OV8 AmRv1ySWPTP7AAMFB/9PQK/VtlNUJvg8GXj9ootzrteGfVZVVT4XBJkfwBcpC/XcPzldjv+3 HYudvpdNK3lLujXeA5fLOH+Z/G9WBc5pFVSMocI71I8bT8lIAzreg0WvkWg5V2WZsUMlnDL9 mpwIGFhlbM3gfDMs7MPMu8YQRFVdUvtSpaAs8OFfGQ0ia3LGZcjA6Ik2+xcqscEJzNH+qh8V m5jjp28yZgaqTaRbg3M/+MTbMpicpZuqF4rnB0AQD12/3BNWDR6bmh+EkYSMcEIpQmBM51qM EKYTQGybRCjpnKHGOxG0rfFY1085mBDZCH5Kx0cl0HVJuQKC+dV2ZY5AqjcKwAxpE75MLFkr wkkEGBECAAkFAlk3nEQCGwwACgkQoDSui/t3IH7nnwCfcJWUDUFKdCsBH/E5d+0ZnMQi+G0A nAuWpQkjM1ASeQwSHEeAWPgskBQL
  • Cc: xen-devel@xxxxxxxxxxxxxxxxxxxx, Andrew Cooper <andrew.cooper3@xxxxxxxxxx>, Anthony PERARD <anthony.perard@xxxxxxxxxx>, Michal Orzel <michal.orzel@xxxxxxx>, Julien Grall <julien@xxxxxxx>, Stefano Stabellini <sstabellini@xxxxxxxxxx>, Roger Pau Monné <roger.pau@xxxxxxxxxx>
  • Delivery-date: Wed, 12 Mar 2025 08:21:29 +0000
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>

On 11.03.2025 19:35, Alejandro Vallejo wrote:
> On Tue Mar 11, 2025 at 3:45 PM GMT, Jan Beulich wrote:
>> On 11.03.2025 16:42, Roger Pau Monné wrote:
>>> On Tue, Mar 11, 2025 at 02:53:04PM +0000, Alejandro Vallejo wrote:
>>>> On Tue Mar 11, 2025 at 12:46 PM GMT, Roger Pau Monné wrote:
>>>>> On Tue, Mar 04, 2025 at 11:10:00AM +0000, Alejandro Vallejo wrote:
>>>>>> The logic has too many levels of indirection and it's very hard to
>>>>>> understand it its current form. Split it between the corner case where
>>>>>> the adjustment is bigger than the current claim and the rest to avoid 5
>>>>>> auxiliary variables.
>>>>>>
>>>>>> Add a functional change to prevent negative adjustments from
>>>>>> re-increasing the claim. This has the nice side effect of avoiding
>>>>>> taking the heap lock here on every free.
>>>>>>
>>>>>> While at it, fix incorrect field name in nearby comment.
>>>>>>
>>>>>> Signed-off-by: Alejandro Vallejo <alejandro.vallejo@xxxxxxxxx>
>>>>>
>>>>> Acked-by: Roger Pau Monné <roger.pau@xxxxxxxxxx>
>>>>
>>>> Thanks.
>>>>
>>>>> I think it would be nice to also ensure that once the domain is
>>>>> finished building (maybe when it's unpaused for the first
>>>>> time?) d->outstanding_pages is set to 0.  IMO the claim system was
>>>>> designed to avoid races during domain building, and shouldn't be used
>>>>> once the domain is already running.
>>>>>
>>>>> Thanks, Roger.
>>>>
>>>> As a matter of implementation that's already the case by toolstack being 
>>>> "nice"
>>>> and unconditionally clearing claims after populating the physmap.
>>>
>>> I see.  Another option would be to refuse the unpause a domain if it
>>> still has pending claims.  However I don't know how that will work out
>>> with all possible toolstacks.
>>>
>>>> However, I agree the hypervisor should do it on its own. I didn't find a
>>>> suitable place for it. 
>>>
>>> You could do it in arch_domain_creation_finished().
>>
>> Except that better wouldn't be arch-specific.
> 
> Why would it have to be arch-specific though? As far as the hypervisor is
> concerned, it doesn't seem to be.

Together with Roger's earlier clarification on his original remark, I fear
I don't understand the question: I asked that it not be arch-specific. And
Roger clarified that he also didn't mean it to be.

Jan



 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.