[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH 3/4] Revert part of "x86/mwait-idle: disable IBRS during long idle"
- To: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
- From: Jan Beulich <jbeulich@xxxxxxxx>
- Date: Wed, 25 Jun 2025 11:58:02 +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: Roger Pau Monné <roger.pau@xxxxxxxxxx>, Xen-devel <xen-devel@xxxxxxxxxxxxxxxxxxxx>
- Delivery-date: Wed, 25 Jun 2025 09:58:18 +0000
- List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
On 24.06.2025 18:39, Andrew Cooper wrote:
> Most of the patch (handling of CPUIDLE_FLAG_IBRS) is fine, but the
> adjustements to mwait_idle() are not.
>
> spec_ctrl_{enter,exit}_idle() do more than just alter MSR_SPEC_CTRL.IBRS. The
> VERW and RSB stuff are **unsafe** to omit.
>
> The only reason this doesn't need an XSA is because no changes were made to
> the lower level mwait_idle_with_hints(), and thus it remained properly
> protected.
>
> I.e. This change only served to double the expensive operations in the case it
> was trying to optimise.
>
> I have an idea of how to plumb this more nicely, but it requires larger
> changes to legacy IBRS handling to not make spec_ctrl_enter_idle() vulnerable
> in other ways.
What are the concerns here? As it looks skipping the MSR write would look
to require checking some (per-CPU) conditional. Conditional branches can't
really be of concern, or the "if (cx->ibrs_disable)" that you're now
removing again would have been of concern, too. Hence simply a new SCF_
flag would look to be sufficient, for mwait_idle() to convey the necessary
info to spec_ctrl_enter_idle()?
> In the short term, simply take out the perf hit.
>
> Fixes: 08acdf9a2615 ("x86/mwait-idle: disable IBRS during long idle")
> Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx>
|
