linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Dave Hansen <dave.hansen@linux.intel.com>
To: linux-mm@kvack.org
Cc: linux-kernel@vger.kernel.org,
	Dave Hansen <dave.hansen@linux.intel.com>,
	tglx@linutronix.de, mingo@redhat.com, bp@alien8.de,
	x86@kernel.org, luto@kernel.org
Subject: [RFC][PATCH 1/8] x86/pkeys: add PKRU storage outside of task XSAVE buffer
Date: Tue, 22 Jun 2021 15:24:57 -0700	[thread overview]
Message-ID: <20210622222457.C16E9CB5@viggo.jf.intel.com> (raw)
In-Reply-To: <20210622222455.E901B5AC@viggo.jf.intel.com>


From: Dave Hansen <dave.hansen@linux.intel.com>

PKRU has space in the task XSAVE buffer, but is not context-switched by
XSAVE/XRSTOR.  It is switched more eagerly than other FPU state because
PKRU affects things like copy_to/from_user().  This is because PKRU
affects user *PERMISSION* accesses, not just accesses made from user
*MODE* itself.

Prepare to move PKRU away from being XSAVE-managed.  Allocate space in
the thread_struct for it and save/restore it in the context-switch path
separately from the XSAVE-managed features.

Leave the XSAVE storage in place for now to ensure bisectability.

Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: x86@kernel.org
Cc: Andy Lutomirski <luto@kernel.org>
---

 b/arch/x86/include/asm/pkru.h  |    5 +++++
 b/arch/x86/kernel/cpu/common.c |    3 +++
 b/arch/x86/kernel/process_64.c |    9 ++++-----
 b/arch/x86/mm/pkeys.c          |    2 ++
 4 files changed, 14 insertions(+), 5 deletions(-)

diff -puN arch/x86/include/asm/pkru.h~pkru-stash-thread-value arch/x86/include/asm/pkru.h
--- a/arch/x86/include/asm/pkru.h~pkru-stash-thread-value	2021-06-22 14:49:06.594051763 -0700
+++ b/arch/x86/include/asm/pkru.h	2021-06-22 14:49:06.607051763 -0700
@@ -44,11 +44,16 @@ static inline void write_pkru(u32 pkru)
 	if (!cpu_feature_enabled(X86_FEATURE_OSPKE))
 		return;
 	/*
+	 * Update the actual register.
+	 *
 	 * WRPKRU is relatively expensive compared to RDPKRU.
 	 * Avoid WRPKRU when it would not change the value.
 	 */
 	if (pkru != rdpkru())
 		wrpkru(pkru);
+
+	/* Update the thread-local, context-switched value: */
+	current->thread.pkru = pkru;
 }
 
 static inline void pkru_write_default(void)
diff -puN arch/x86/kernel/cpu/common.c~pkru-stash-thread-value arch/x86/kernel/cpu/common.c
--- a/arch/x86/kernel/cpu/common.c~pkru-stash-thread-value	2021-06-22 14:49:06.596051763 -0700
+++ b/arch/x86/kernel/cpu/common.c	2021-06-22 14:49:06.608051763 -0700
@@ -482,6 +482,9 @@ static __always_inline void setup_pku(st
 	cr4_set_bits(X86_CR4_PKE);
 	/* Load the default PKRU value */
 	pkru_write_default();
+
+	/* Establish the default value for future tasks: */
+	init_task.thread.pkru = init_pkru_value;
 }
 
 #ifdef CONFIG_X86_INTEL_MEMORY_PROTECTION_KEYS
diff -puN arch/x86/kernel/process_64.c~pkru-stash-thread-value arch/x86/kernel/process_64.c
--- a/arch/x86/kernel/process_64.c~pkru-stash-thread-value	2021-06-22 14:49:06.599051763 -0700
+++ b/arch/x86/kernel/process_64.c	2021-06-22 14:49:06.608051763 -0700
@@ -349,15 +349,14 @@ static __always_inline void load_seg_leg
 static __always_inline void x86_pkru_load(struct thread_struct *prev,
 					  struct thread_struct *next)
 {
-	if (!cpu_feature_enabled(X86_FEATURE_OSPKE))
-		return;
+	u32 pkru = read_pkru();
 
 	/* Stash the prev task's value: */
-	prev->pkru = rdpkru();
+	prev->pkru = pkru;
 
 	/*
-	 * PKRU writes are slightly expensive.  Avoid them when not
-	 * strictly necessary:
+	 * PKRU writes are slightly expensive.  Avoid
+	 * them when not strictly necessary:
 	 */
 	if (prev->pkru != next->pkru)
 		wrpkru(next->pkru);
diff -puN arch/x86/mm/pkeys.c~pkru-stash-thread-value arch/x86/mm/pkeys.c
--- a/arch/x86/mm/pkeys.c~pkru-stash-thread-value	2021-06-22 14:49:06.604051763 -0700
+++ b/arch/x86/mm/pkeys.c	2021-06-22 14:49:06.609051763 -0700
@@ -159,6 +159,8 @@ static ssize_t init_pkru_write_file(stru
 		return -EINVAL;
 
 	WRITE_ONCE(init_pkru_value, new_init_pkru);
+	WRITE_ONCE(init_task.thread.pkru, new_init_pkru);
+
 	return count;
 }
 
_

  reply	other threads:[~2021-06-22 22:25 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-22 22:24 [RFC][PATCH 0/8] x86/pkeys: remove PKRU from kernel XSAVE buffer Dave Hansen
2021-06-22 22:24 ` Dave Hansen [this message]
2021-06-22 22:24 ` [RFC][PATCH 2/8] x86/fpu: hook up PKRU into signal user ABIs Dave Hansen
2021-06-22 22:25 ` [RFC][PATCH 3/8] x86/fpu: separate the setup of xfeatures not in fpstate Dave Hansen
2021-06-22 22:25 ` [RFC][PATCH 4/8] x86/fpu: remove PKRU from FPU user state clearing Dave Hansen
2021-06-22 22:25 ` [RFC][PATCH 5/8] x86/fpu: XSAVE buffer access routine rename Dave Hansen
2021-06-22 22:25 ` [RFC][PATCH 6/8] x86/fpu: update xstate size calculations for non-XSAVE-managed features Dave Hansen
2021-06-22 22:25 ` [RFC][PATCH 7/8] x86/fpu: actually stop using XSAVE on PKRU Dave Hansen
2021-06-22 22:25 ` [RFC][PATCH 8/8] x86/pkeys: remove init_pkru_value variable Dave Hansen

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210622222457.C16E9CB5@viggo.jf.intel.com \
    --to=dave.hansen@linux.intel.com \
    --cc=bp@alien8.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=luto@kernel.org \
    --cc=mingo@redhat.com \
    --cc=tglx@linutronix.de \
    --cc=x86@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).