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

Re: [PATCH v2 1/7] x86/smp: do not use shorthand IPI destinations in CPU hot{,un}plug contexts


  • To: Roger Pau Monne <roger.pau@xxxxxxxxxx>
  • From: Jan Beulich <jbeulich@xxxxxxxx>
  • Date: Tue, 11 Jun 2024 09:42:39 +0200
  • 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>, George Dunlap <george.dunlap@xxxxxxxxxx>, Julien Grall <julien@xxxxxxx>, Stefano Stabellini <sstabellini@xxxxxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxxx
  • Delivery-date: Tue, 11 Jun 2024 07:42:49 +0000
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>

On 10.06.2024 16:20, Roger Pau Monne wrote:
> Due to the current rwlock logic, if the CPU calling get_cpu_maps() does so 
> from
> a cpu_hotplug_{begin,done}() region the function will still return success,
> because a CPU taking the rwlock in read mode after having taken it in write
> mode is allowed.  Such behavior however defeats the purpose of get_cpu_maps(),

I'm not happy to see you still have this claim here. The behavior may (appear
to) defeat the purpose here, but as expressed previously I don't view that as
being a general pattern.

> as it should always return false when called with a CPU hot{,un}plug operation
> is in progress.  Otherwise the logic in send_IPI_mask() is wrong, as it could
> decide to use the shorthand even when a CPU operation is in progress.
> 
> Introduce a new helper to detect whether the current caller is between a
> cpu_hotplug_{begin,done}() region and use it in send_IPI_mask() to restrict
> shorthand usage.
> 
> Fixes: 5500d265a2a8 ('x86/smp: use APIC ALLBUT destination shorthand when 
> possible')
> Signed-off-by: Roger Pau Monné <roger.pau@xxxxxxxxxx>
> ---
> Changes since v1:
>  - Modify send_IPI_mask() to detect CPU hotplug context.
> ---
>  xen/arch/x86/smp.c       |  2 +-
>  xen/common/cpu.c         |  5 +++++
>  xen/include/xen/cpu.h    | 10 ++++++++++
>  xen/include/xen/rwlock.h |  2 ++
>  4 files changed, 18 insertions(+), 1 deletion(-)
> 
> diff --git a/xen/arch/x86/smp.c b/xen/arch/x86/smp.c
> index 7443ad20335e..04c6a0572319 100644
> --- a/xen/arch/x86/smp.c
> +++ b/xen/arch/x86/smp.c
> @@ -88,7 +88,7 @@ void send_IPI_mask(const cpumask_t *mask, int vector)
>       * the system have been accounted for.
>       */
>      if ( system_state > SYS_STATE_smp_boot &&
> -         !unaccounted_cpus && !disabled_cpus &&
> +         !unaccounted_cpus && !disabled_cpus && !cpu_in_hotplug_context() &&
>           /* NB: get_cpu_maps lock requires enabled interrupts. */
>           local_irq_is_enabled() && (cpus_locked = get_cpu_maps()) &&
>           (park_offline_cpus ||

Along with changing the condition you also want to update the comment to
reflect the code adjustment.

If we can agree on respective wording in both places, I'd be happy to make
respective adjustments while committing; the code changes themselves are
Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx>

Jan



 


Rackspace

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