From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 0939AC433EF for ; Wed, 23 Feb 2022 05:27:22 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S238338AbiBWF1n (ORCPT ); Wed, 23 Feb 2022 00:27:43 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:58106 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S238342AbiBWF0t (ORCPT ); Wed, 23 Feb 2022 00:26:49 -0500 Received: from mail-yb1-xb49.google.com (mail-yb1-xb49.google.com [IPv6:2607:f8b0:4864:20::b49]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 7706F6D18A for ; Tue, 22 Feb 2022 21:25:09 -0800 (PST) Received: by mail-yb1-xb49.google.com with SMTP id s11-20020a255e0b000000b0062277953037so26418223ybb.21 for ; Tue, 22 Feb 2022 21:25:09 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20210112; h=date:in-reply-to:message-id:mime-version:references:subject:from:to :cc; bh=8mE7Rl7hvX81PshmRchAKXUw+5mI1RoqZfuGnjfnK/Y=; b=nQn+PHDiNT2rMQDvqUgnDcQMr757vUw1xTfeWksvd/uBs+Y7jycIPiRL7S70urPZ7R 8fkcrVfE/buziQatz1VzPrbZiKkVL88UDR47kWH9EmA5cOthED7jMA11HUYJo4xm65uX LbfLzZbh1vmkmQVGPzCZHhrvQkG2yMMs6jlVlPGs+f1MYI9eWUOQGRHF/yHWkKAQuG9+ 90/ZWnlFIBlEl9DgOrfwf+XgiBD2V3yGyekMz3rvqGQAKX6yRkMkTlYi4dCG11i6F1bh Phi+HrSt7E6OdAajjE7T2WGE+PvAmrOCFwLD3vGl5nSuUdeg1Ly1y8XU57nLBdCyfJFb Mrkw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:date:in-reply-to:message-id:mime-version :references:subject:from:to:cc; bh=8mE7Rl7hvX81PshmRchAKXUw+5mI1RoqZfuGnjfnK/Y=; b=RpA4rP1G45TeX61rDnip0doVlA9/iwyLora50rY0MHWedi8v4vTBpvlKTomKUeM6pt 2U6YCKg1pSqTwyl0VJiZ4Nb81EnlqafTvHv9m+JNT8mS0t/WeQZRjFt4iM9jV7Wb6aGz zHllINcGOdwxh3W2lBw3q3eeom4AGjbTPq/+6MqJ3nKSOmCqSdFw0ABLzCPioKhD5o7P n2TkOYEN/Rw7Gkk5Ye8y/0GD5S0EM1W3zK3VhGwF0Wg9TOlvN7ZDBdtdMt9VgOp3hRPE 3WKVXR3wh/mnzfxhhRgDD2U1ZHvCi4Pg1YZzhFWAoZ84L/yewpaPwsn2c2TDzl+M3JzU anvQ== X-Gm-Message-State: AOAM530o8Xf+1sVuYv5DlqLLS/Ob9il8k+Z/MOq0Yi7Bk+e2HddclIsp 7eMhQqzemzUoG7pZCUY23rZYbgf55xz9bI5SO04K2Xgz7f1pv8S5x8qjFfuvCT4v8wGFvAOCxNk wqoMLaAa3iH93ZRf/6BgK1vngmiV15UZwBrBt2WEB+HG83VjDjKOIoQ7fJpj+fbHnRBjNDtCB X-Google-Smtp-Source: ABdhPJw219sWzgLSxca1tmB8U/7JNof+qQBV1MFnda0VeJBWRGjfZ3IqT80R/0u8HdfUK/arZaU6tbXLeppv X-Received: from js-desktop.svl.corp.google.com ([2620:15c:2cd:202:ccbe:5d15:e2e6:322]) (user=junaids job=sendgmr) by 2002:a0d:df4e:0:b0:2d0:ab1e:6055 with SMTP id i75-20020a0ddf4e000000b002d0ab1e6055mr27301388ywe.333.1645593896772; Tue, 22 Feb 2022 21:24:56 -0800 (PST) Date: Tue, 22 Feb 2022 21:22:08 -0800 In-Reply-To: <20220223052223.1202152-1-junaids@google.com> Message-Id: <20220223052223.1202152-33-junaids@google.com> Mime-Version: 1.0 References: <20220223052223.1202152-1-junaids@google.com> X-Mailer: git-send-email 2.35.1.473.g83b2b277ed-goog Subject: [RFC PATCH 32/47] x86: asi: Allocate FPU state separately when ASI is enabled. From: Junaid Shahid To: linux-kernel@vger.kernel.org Cc: kvm@vger.kernel.org, pbonzini@redhat.com, jmattson@google.com, pjt@google.com, oweisse@google.com, alexandre.chartre@oracle.com, rppt@linux.ibm.com, dave.hansen@linux.intel.com, peterz@infradead.org, tglx@linutronix.de, luto@kernel.org, linux-mm@kvack.org Content-Type: text/plain; charset="UTF-8" Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org We are going to be mapping the task_struct in the restricted ASI address space. However, the task_struct also contains the FPU register state embedded inside it, which can contain sensitive information. So when ASI is enabled, always allocate the FPU state from a separate slab cache to keep it out of task_struct. Signed-off-by: Junaid Shahid --- arch/x86/include/asm/fpu/api.h | 1 + arch/x86/kernel/fpu/core.c | 45 ++++++++++++++++++++++++++++++++-- arch/x86/kernel/fpu/init.c | 7 ++++-- arch/x86/kernel/fpu/internal.h | 1 + arch/x86/kernel/fpu/xstate.c | 21 +++++++++++++--- arch/x86/kernel/process.c | 7 +++++- 6 files changed, 74 insertions(+), 8 deletions(-) diff --git a/arch/x86/include/asm/fpu/api.h b/arch/x86/include/asm/fpu/api.h index c2767a6a387e..6f5ca3c2ef4a 100644 --- a/arch/x86/include/asm/fpu/api.h +++ b/arch/x86/include/asm/fpu/api.h @@ -112,6 +112,7 @@ extern void fpu__init_cpu(void); extern void fpu__init_system(struct cpuinfo_x86 *c); extern void fpu__init_check_bugs(void); extern void fpu__resume_cpu(void); +extern void fpstate_cache_init(void); #ifdef CONFIG_MATH_EMULATION extern void fpstate_init_soft(struct swregs_state *soft); diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index 8ea306b1bf8e..d7859573973d 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -59,6 +59,8 @@ static DEFINE_PER_CPU(bool, in_kernel_fpu); */ DEFINE_PER_CPU(struct fpu *, fpu_fpregs_owner_ctx); +struct kmem_cache *fpstate_cachep; + static bool kernel_fpu_disabled(void) { return this_cpu_read(in_kernel_fpu); @@ -443,7 +445,9 @@ static void __fpstate_reset(struct fpstate *fpstate) void fpstate_reset(struct fpu *fpu) { /* Set the fpstate pointer to the default fpstate */ - fpu->fpstate = &fpu->__fpstate; + if (!cpu_feature_enabled(X86_FEATURE_ASI)) + fpu->fpstate = &fpu->__fpstate; + __fpstate_reset(fpu->fpstate); /* Initialize the permission related info in fpu */ @@ -464,6 +468,26 @@ static inline void fpu_inherit_perms(struct fpu *dst_fpu) } } +void fpstate_cache_init(void) +{ + if (cpu_feature_enabled(X86_FEATURE_ASI)) { + size_t fpstate_size; + + /* TODO: Is the ALIGN-64 really needed? */ + fpstate_size = fpu_kernel_cfg.default_size + + ALIGN(offsetof(struct fpstate, regs), 64); + + fpstate_cachep = kmem_cache_create_usercopy( + "fpstate", + fpstate_size, + __alignof__(struct fpstate), + SLAB_PANIC | SLAB_ACCOUNT, + offsetof(struct fpstate, regs), + fpu_kernel_cfg.default_size, + NULL); + } +} + /* Clone current's FPU state on fork */ int fpu_clone(struct task_struct *dst, unsigned long clone_flags) { @@ -473,6 +497,22 @@ int fpu_clone(struct task_struct *dst, unsigned long clone_flags) /* The new task's FPU state cannot be valid in the hardware. */ dst_fpu->last_cpu = -1; + if (cpu_feature_enabled(X86_FEATURE_ASI)) { + dst_fpu->fpstate = kmem_cache_alloc_node( + fpstate_cachep, GFP_KERNEL, + page_to_nid(virt_to_page(dst))); + if (!dst_fpu->fpstate) + return -ENOMEM; + + /* + * TODO: We may be able to skip the copy since the registers are + * restored below anyway. + */ + memcpy(dst_fpu->fpstate, src_fpu->fpstate, + fpu_kernel_cfg.default_size + + offsetof(struct fpstate, regs)); + } + fpstate_reset(dst_fpu); if (!cpu_feature_enabled(X86_FEATURE_FPU)) @@ -531,7 +571,8 @@ int fpu_clone(struct task_struct *dst, unsigned long clone_flags) void fpu_thread_struct_whitelist(unsigned long *offset, unsigned long *size) { *offset = offsetof(struct thread_struct, fpu.__fpstate.regs); - *size = fpu_kernel_cfg.default_size; + *size = cpu_feature_enabled(X86_FEATURE_ASI) + ? 0 : fpu_kernel_cfg.default_size; } /* diff --git a/arch/x86/kernel/fpu/init.c b/arch/x86/kernel/fpu/init.c index 621f4b6cac4a..8b722bf98135 100644 --- a/arch/x86/kernel/fpu/init.c +++ b/arch/x86/kernel/fpu/init.c @@ -161,9 +161,11 @@ static void __init fpu__init_task_struct_size(void) /* * Add back the dynamically-calculated register state - * size. + * size, except when ASI is enabled, since in that case + * the FPU state is always allocated dynamically. */ - task_size += fpu_kernel_cfg.default_size; + if (!cpu_feature_enabled(X86_FEATURE_ASI)) + task_size += fpu_kernel_cfg.default_size; /* * We dynamically size 'struct fpu', so we require that @@ -223,6 +225,7 @@ static void __init fpu__init_init_fpstate(void) */ void __init fpu__init_system(struct cpuinfo_x86 *c) { + current->thread.fpu.fpstate = ¤t->thread.fpu.__fpstate; fpstate_reset(¤t->thread.fpu); fpu__init_system_early_generic(c); diff --git a/arch/x86/kernel/fpu/internal.h b/arch/x86/kernel/fpu/internal.h index dbdb31f55fc7..30acc7d0cb1a 100644 --- a/arch/x86/kernel/fpu/internal.h +++ b/arch/x86/kernel/fpu/internal.h @@ -3,6 +3,7 @@ #define __X86_KERNEL_FPU_INTERNAL_H extern struct fpstate init_fpstate; +extern struct kmem_cache *fpstate_cachep; /* CPU feature check wrappers */ static __always_inline __pure bool use_xsave(void) diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c index d28829403ed0..96d12f351f19 100644 --- a/arch/x86/kernel/fpu/xstate.c +++ b/arch/x86/kernel/fpu/xstate.c @@ -13,6 +13,7 @@ #include #include #include +#include #include #include @@ -1495,8 +1496,15 @@ arch_initcall(xfd_update_static_branch) void fpstate_free(struct fpu *fpu) { - if (fpu->fpstate && fpu->fpstate != &fpu->__fpstate) - vfree(fpu->fpstate); + WARN_ON_ONCE(cpu_feature_enabled(X86_FEATURE_ASI) && + fpu->fpstate == &fpu->__fpstate); + + if (fpu->fpstate && fpu->fpstate != &fpu->__fpstate) { + if (fpu->fpstate->is_valloc) + vfree(fpu->fpstate); + else + kmem_cache_free(fpstate_cachep, fpu->fpstate); + } } /** @@ -1574,7 +1582,14 @@ static int fpstate_realloc(u64 xfeatures, unsigned int ksize, fpregs_unlock(); - vfree(curfps); + WARN_ON_ONCE(cpu_feature_enabled(X86_FEATURE_ASI) && !curfps); + if (curfps) { + if (curfps->is_valloc) + vfree(curfps); + else + kmem_cache_free(fpstate_cachep, curfps); + } + return 0; } diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index c8d4a00a4de7..f9bd1c3415d4 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -80,6 +80,11 @@ EXPORT_PER_CPU_SYMBOL(cpu_tss_rw); DEFINE_PER_CPU(bool, __tss_limit_invalid); EXPORT_PER_CPU_SYMBOL_GPL(__tss_limit_invalid); +void __init arch_task_cache_init(void) +{ + fpstate_cache_init(); +} + /* * this gets called so that we can store lazy state into memory and copy the * current task into the new thread. @@ -101,7 +106,7 @@ int arch_dup_task_struct(struct task_struct *dst, struct task_struct *src) #ifdef CONFIG_X86_64 void arch_release_task_struct(struct task_struct *tsk) { - if (fpu_state_size_dynamic()) + if (fpu_state_size_dynamic() || cpu_feature_enabled(X86_FEATURE_ASI)) fpstate_free(&tsk->thread.fpu); } #endif -- 2.35.1.473.g83b2b277ed-goog