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

Re: [XEN PATCH 9/9] x86/smp: start APs in parallel during boot


  • To: Krystian Hebel <krystian.hebel@xxxxxxxxx>
  • From: Jan Beulich <jbeulich@xxxxxxxx>
  • Date: Wed, 13 Mar 2024 14:30:33 +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: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>, Roger Pau Monné <roger.pau@xxxxxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxxx, Wei Liu <wl@xxxxxxx>
  • Delivery-date: Wed, 13 Mar 2024 13:30:44 +0000
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>

On 12.03.2024 18:13, Krystian Hebel wrote:
> 
> On 8.02.2024 13:37, Jan Beulich wrote:
>> On 14.11.2023 18:50, Krystian Hebel wrote:
>>> Multiple delays are required when sending IPIs and waiting for
>>> responses. During boot, 4 such IPIs were sent per each AP. With this
>>> change, only one set of broadcast IPIs is sent. This reduces boot time,
>>> especially for platforms with large number of cores.
>> Yet APs do their startup work in parallel only for a brief period of
>> time, if I'm not mistaken. Othwerwise I can't see why you'd still have
>> cpu_up() in __start_xen().
> cpu_up() is left because multiple notifiers aren't easy to convert to work
> in parallel. In terms of lines of code it looks like a brief period, but all
> the delays along the way were taking much more time than the actual
> work. As the gain was already more than what I hoped for, I decided
> against spending too much time trying to fix the notifiers' code for
> minimal profit.

Which is all fine. Just that by title of this patch and the cover letter
I expected more. Adding "partly" or some such in both places may help.

>>> --- a/xen/arch/x86/smpboot.c
>>> +++ b/xen/arch/x86/smpboot.c
>>> @@ -425,7 +425,7 @@ void start_secondary(unsigned int cpu)
>>>   
>>>   static int wakeup_secondary_cpu(int phys_apicid, unsigned long start_eip)
>>>   {
>>> -    unsigned long send_status = 0, accept_status = 0;
>>> +    unsigned long send_status = 0, accept_status = 0, sh = 0;
>> sh doesn't need to be 64 bits wide, does it?
> No, will change.
>>
>>>       int maxlvt, timeout, i;
>>>   
>>>       /*
>>> @@ -445,6 +445,12 @@ static int wakeup_secondary_cpu(int phys_apicid, 
>>> unsigned long start_eip)
>>>       if ( tboot_in_measured_env() && !tboot_wake_ap(phys_apicid, 
>>> start_eip) )
>>>           return 0;
>>>   
>>> +    /*
>>> +     * Use destination shorthand for broadcasting IPIs during boot.
>>> +     */
>> Nit (style): This is a single line comment.
> Ack
>>
>>> +    if ( phys_apicid == BAD_APICID )
>>> +        sh = APIC_DEST_ALLBUT;
>> I think the latest for this the function parameter wants changing to
>> unsigned int (in another prereq patch).
> What do you mean, phys_apicid in wakeup_secondary_cpu()? It is passed
> as signed int since __cpu_up(), should I change all of those to unsigned?

That would be best, yes. BAD_APICID, after all, is an unsigned constant
(no matter that its definition involves a unary minus operator).

Jan



 


Rackspace

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