|
From: | Richard Henderson |
Subject: | Re: [Qemu-devel] Qemu deadlocks in tb_lock when using SVM+SoftMMU |
Date: | Tue, 7 Mar 2017 06:21:34 +1100 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.7.0 |
On 03/07/2017 03:58 AM, Paolo Bonzini wrote:
On 06/03/2017 02:34, Richard Henderson wrote:My guess is that everything from cpu_svm_check_intercept_param on should be done from do_interrupt instead of during raise_interrupt.From cpu_svm_check_intercept_param, or from cpu_vmexit? The former seemsto be too early because it will usually not even do anything, but treating cpu_vmexit like an exception is a very good idea indeed. This is my uncompiled take.
I hadn't considered that approach. But it looks very plausible. r~
[Prev in Thread] | Current Thread | [Next in Thread] |