linux-arch.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Dave Martin <Dave.Martin@arm.com>
To: linux-arm-kernel@lists.infradead.org
Cc: linux-arch@vger.kernel.org, libc-alpha@sourceware.org,
	Ard Biesheuvel <ard.biesheuvel@linaro.org>,
	Szabolcs Nagy <szabolcs.nagy@arm.com>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will.deacon@arm.com>,
	Richard Sandiford <richard.sandiford@arm.com>,
	kvmarm@lists.cs.columbia.edu
Subject: [PATCH 12/27] arm64/sve: Support vector length resetting for new processes
Date: Wed,  9 Aug 2017 13:05:18 +0100	[thread overview]
Message-ID: <1502280338-23002-13-git-send-email-Dave.Martin@arm.com> (raw)
In-Reply-To: <1502280338-23002-1-git-send-email-Dave.Martin@arm.com>

It's desirable to be able to reset the vector length to some sane
default for new processes, since the new binary and its libraries
processes may or may not be SVE-aware.

This patch tracks the desired post-exec vector length (if any) in a
new thread member sve_vl_onexec, and adds a new thread flag
TIF_SVE_VL_INHERIT to control whether to inherit or reset the
vector length.  Currently these are inactive.  Subsequent patches
will provide the capability to configure them.

Signed-off-by: Dave Martin <Dave.Martin@arm.com>
---
 arch/arm64/include/asm/processor.h   |  1 +
 arch/arm64/include/asm/thread_info.h |  1 +
 arch/arm64/kernel/fpsimd.c           | 13 +++++++++++++
 3 files changed, 15 insertions(+)

diff --git a/arch/arm64/include/asm/processor.h b/arch/arm64/include/asm/processor.h
index 969feed..da8802a 100644
--- a/arch/arm64/include/asm/processor.h
+++ b/arch/arm64/include/asm/processor.h
@@ -87,6 +87,7 @@ struct thread_struct {
 	struct fpsimd_state	fpsimd_state;
 	void			*sve_state;	/* SVE registers, if any */
 	u16			sve_vl;		/* SVE vector length */
+	u16			sve_vl_onexec;	/* SVE vl after next exec */
 	unsigned long		fault_address;	/* fault info */
 	unsigned long		fault_code;	/* ESR_EL1 value */
 	struct debug_info	debug;		/* debugging */
diff --git a/arch/arm64/include/asm/thread_info.h b/arch/arm64/include/asm/thread_info.h
index 1a4b30b..bf9c552 100644
--- a/arch/arm64/include/asm/thread_info.h
+++ b/arch/arm64/include/asm/thread_info.h
@@ -97,6 +97,7 @@ struct thread_info {
 #define TIF_SINGLESTEP		21
 #define TIF_32BIT		22	/* 32bit process */
 #define TIF_SVE			23	/* Scalable Vector Extension in use */
+#define TIF_SVE_VL_INHERIT	24	/* Inherit sve_vl_onexec across exec */
 
 #define _TIF_SIGPENDING		(1 << TIF_SIGPENDING)
 #define _TIF_NEED_RESCHED	(1 << TIF_NEED_RESCHED)
diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
index 37dd1b2..80ecb2d 100644
--- a/arch/arm64/kernel/fpsimd.c
+++ b/arch/arm64/kernel/fpsimd.c
@@ -108,6 +108,9 @@ static void task_fpsimd_save(void);
  */
 static DEFINE_PER_CPU(struct fpsimd_state *, fpsimd_last_state);
 
+/* Default VL for tasks that don't set it explicitly: */
+static int sve_default_vl = -1;
+
 static void sve_free(struct task_struct *task)
 {
 	kfree(task->thread.sve_state);
@@ -392,6 +395,9 @@ void fpsimd_flush_thread(void)
 		clear_thread_flag(TIF_SVE);
 		sve_free(current);
 
+		current->thread.sve_vl = current->thread.sve_vl_onexec ?
+			current->thread.sve_vl_onexec : sve_default_vl;
+
 		/*
 		 * User tasks must have a valid vector length set, but tasks
 		 * forked early (e.g., init) may not initially have one.
@@ -401,6 +407,13 @@ void fpsimd_flush_thread(void)
 		 * If not, something went badly wrong.
 		 */
 		BUG_ON(!sve_vl_valid(current->thread.sve_vl));
+
+		/*
+		 * If the task is not set to inherit, ensure that the vector
+		 * length will be reset by a subsequent exec:
+		 */
+		if (!test_thread_flag(TIF_SVE_VL_INHERIT))
+			current->thread.sve_vl_onexec = 0;
 	}
 
 	set_thread_flag(TIF_FOREIGN_FPSTATE);
-- 
2.1.4

WARNING: multiple messages have this Message-ID (diff)
From: Dave Martin <Dave.Martin@arm.com>
To: linux-arm-kernel@lists.infradead.org
Cc: Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will.deacon@arm.com>,
	Ard Biesheuvel <ard.biesheuvel@linaro.org>,
	Szabolcs Nagy <szabolcs.nagy@arm.com>,
	Richard Sandiford <richard.sandiford@arm.com>,
	kvmarm@lists.cs.columbia.edu, libc-alpha@sourceware.org,
	linux-arch@vger.kernel.org
Subject: [PATCH 12/27] arm64/sve: Support vector length resetting for new processes
Date: Wed,  9 Aug 2017 13:05:18 +0100	[thread overview]
Message-ID: <1502280338-23002-13-git-send-email-Dave.Martin@arm.com> (raw)
Message-ID: <20170809120518.vPVJlab6goDUXxRoE0QsC7powbWqDb0veX8JogQZaEA@z> (raw)
In-Reply-To: <1502280338-23002-1-git-send-email-Dave.Martin@arm.com>

It's desirable to be able to reset the vector length to some sane
default for new processes, since the new binary and its libraries
processes may or may not be SVE-aware.

This patch tracks the desired post-exec vector length (if any) in a
new thread member sve_vl_onexec, and adds a new thread flag
TIF_SVE_VL_INHERIT to control whether to inherit or reset the
vector length.  Currently these are inactive.  Subsequent patches
will provide the capability to configure them.

Signed-off-by: Dave Martin <Dave.Martin@arm.com>
---
 arch/arm64/include/asm/processor.h   |  1 +
 arch/arm64/include/asm/thread_info.h |  1 +
 arch/arm64/kernel/fpsimd.c           | 13 +++++++++++++
 3 files changed, 15 insertions(+)

diff --git a/arch/arm64/include/asm/processor.h b/arch/arm64/include/asm/processor.h
index 969feed..da8802a 100644
--- a/arch/arm64/include/asm/processor.h
+++ b/arch/arm64/include/asm/processor.h
@@ -87,6 +87,7 @@ struct thread_struct {
 	struct fpsimd_state	fpsimd_state;
 	void			*sve_state;	/* SVE registers, if any */
 	u16			sve_vl;		/* SVE vector length */
+	u16			sve_vl_onexec;	/* SVE vl after next exec */
 	unsigned long		fault_address;	/* fault info */
 	unsigned long		fault_code;	/* ESR_EL1 value */
 	struct debug_info	debug;		/* debugging */
diff --git a/arch/arm64/include/asm/thread_info.h b/arch/arm64/include/asm/thread_info.h
index 1a4b30b..bf9c552 100644
--- a/arch/arm64/include/asm/thread_info.h
+++ b/arch/arm64/include/asm/thread_info.h
@@ -97,6 +97,7 @@ struct thread_info {
 #define TIF_SINGLESTEP		21
 #define TIF_32BIT		22	/* 32bit process */
 #define TIF_SVE			23	/* Scalable Vector Extension in use */
+#define TIF_SVE_VL_INHERIT	24	/* Inherit sve_vl_onexec across exec */
 
 #define _TIF_SIGPENDING		(1 << TIF_SIGPENDING)
 #define _TIF_NEED_RESCHED	(1 << TIF_NEED_RESCHED)
diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
index 37dd1b2..80ecb2d 100644
--- a/arch/arm64/kernel/fpsimd.c
+++ b/arch/arm64/kernel/fpsimd.c
@@ -108,6 +108,9 @@ static void task_fpsimd_save(void);
  */
 static DEFINE_PER_CPU(struct fpsimd_state *, fpsimd_last_state);
 
+/* Default VL for tasks that don't set it explicitly: */
+static int sve_default_vl = -1;
+
 static void sve_free(struct task_struct *task)
 {
 	kfree(task->thread.sve_state);
@@ -392,6 +395,9 @@ void fpsimd_flush_thread(void)
 		clear_thread_flag(TIF_SVE);
 		sve_free(current);
 
+		current->thread.sve_vl = current->thread.sve_vl_onexec ?
+			current->thread.sve_vl_onexec : sve_default_vl;
+
 		/*
 		 * User tasks must have a valid vector length set, but tasks
 		 * forked early (e.g., init) may not initially have one.
@@ -401,6 +407,13 @@ void fpsimd_flush_thread(void)
 		 * If not, something went badly wrong.
 		 */
 		BUG_ON(!sve_vl_valid(current->thread.sve_vl));
+
+		/*
+		 * If the task is not set to inherit, ensure that the vector
+		 * length will be reset by a subsequent exec:
+		 */
+		if (!test_thread_flag(TIF_SVE_VL_INHERIT))
+			current->thread.sve_vl_onexec = 0;
 	}
 
 	set_thread_flag(TIF_FOREIGN_FPSTATE);
-- 
2.1.4

  parent reply	other threads:[~2017-08-09 12:05 UTC|newest]

Thread overview: 124+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-08-09 12:05 [PATCH 00/27] ARM Scalable Vector Extension (SVE) Dave Martin
2017-08-09 12:05 ` [PATCH 01/27] regset: Add support for dynamically sized regsets Dave Martin
2017-08-09 12:05   ` Dave Martin
2017-08-18 11:52   ` Alex Bennée
2017-08-18 11:52     ` Alex Bennée
2017-08-09 12:05 ` [PATCH 02/27] arm64: KVM: Hide unsupported AArch64 CPU features from guests Dave Martin
2017-08-09 12:05   ` Dave Martin
2017-08-16 11:10   ` Marc Zyngier
2017-08-16 20:32     ` Dave Martin
2017-08-17  8:45       ` Marc Zyngier
2017-08-17  9:57         ` Dave Martin
2017-08-17  9:57           ` Dave Martin
2017-08-09 12:05 ` [PATCH 03/27] arm64: efi: Add missing Kconfig dependency on KERNEL_MODE_NEON Dave Martin
2017-08-18 12:02   ` Alex Bennée
2017-08-18 12:02     ` Alex Bennée
2017-08-09 12:05 ` [PATCH 04/27] arm64: Port deprecated instruction emulation to new sysctl interface Dave Martin
2017-08-18 12:09   ` Alex Bennée
2017-08-18 12:09     ` Alex Bennée
2017-08-09 12:05 ` [PATCH 05/27] arm64: fpsimd: Simplify uses of {set,clear}_ti_thread_flag() Dave Martin
2017-08-15 17:11   ` Ard Biesheuvel
2017-08-18 16:36   ` [PATCH 05/27] arm64: fpsimd: Simplify uses of {set, clear}_ti_thread_flag() Alex Bennée
2017-08-18 16:36     ` Alex Bennée
2017-08-09 12:05 ` [PATCH 06/27] arm64/sve: System register and exception syndrome definitions Dave Martin
2017-08-21  9:33   ` Alex Bennée
2017-08-21  9:33     ` Alex Bennée
2017-08-21 12:34     ` Alex Bennée
2017-08-21 12:34       ` Alex Bennée
2017-08-21 14:26       ` Dave Martin
2017-08-21 14:50         ` Alex Bennée
2017-08-21 14:50           ` Alex Bennée
2017-08-21 15:19           ` Dave Martin
2017-08-21 15:34             ` Alex Bennée
2017-08-21 15:34               ` Alex Bennée
2017-08-21 13:56     ` Dave Martin
2017-08-21 13:56       ` Dave Martin
2017-08-21 14:36       ` Alex Bennée
2017-08-21 14:36         ` Alex Bennée
2017-08-09 12:05 ` [PATCH 07/27] arm64/sve: Low-level SVE architectural state manipulation functions Dave Martin
2017-08-21 10:11   ` Alex Bennée
2017-08-21 10:11     ` Alex Bennée
2017-08-21 14:38     ` Dave Martin
2017-08-21 14:38       ` Dave Martin
2017-08-09 12:05 ` [PATCH 08/27] arm64/sve: Kconfig update and conditional compilation support Dave Martin
2017-08-21 10:12   ` Alex Bennée
2017-08-21 10:12     ` Alex Bennée
2017-08-09 12:05 ` [PATCH 09/27] arm64/sve: Signal frame and context structure definition Dave Martin
2017-08-09 12:05   ` Dave Martin
2017-08-22 10:22   ` Alex Bennée
2017-08-22 10:22     ` Alex Bennée
2017-08-22 11:17     ` Dave Martin
2017-08-22 13:53       ` Alex Bennée
2017-08-22 13:53         ` Alex Bennée
2017-08-22 14:21         ` Dave Martin
2017-08-22 14:21           ` Dave Martin
2017-08-22 15:03           ` Alex Bennée
2017-08-22 15:03             ` Alex Bennée
2017-08-22 15:41             ` Dave Martin
2017-08-09 12:05 ` [PATCH 10/27] arm64/sve: Low-level CPU setup Dave Martin
2017-08-09 12:05   ` Dave Martin
2017-08-22 15:04   ` Alex Bennée
2017-08-22 15:04     ` Alex Bennée
2017-08-22 15:33     ` Dave Martin
2017-08-09 12:05 ` [PATCH 11/27] arm64/sve: Core task context handling Dave Martin
2017-08-15 17:31   ` Ard Biesheuvel
2017-08-16 10:40     ` Dave Martin
2017-08-17 16:42     ` Dave Martin
2017-08-17 16:46       ` Ard Biesheuvel
2017-08-22 16:21   ` Alex Bennée
2017-08-22 16:21     ` Alex Bennée
2017-08-22 17:19     ` Dave Martin
2017-08-22 18:39       ` Alex Bennée
2017-08-22 18:39         ` Alex Bennée
2017-08-09 12:05 ` Dave Martin [this message]
2017-08-09 12:05   ` [PATCH 12/27] arm64/sve: Support vector length resetting for new processes Dave Martin
2017-08-22 16:22   ` Alex Bennée
2017-08-22 16:22     ` Alex Bennée
2017-08-22 17:22     ` Dave Martin
2017-08-22 17:22       ` Dave Martin
2017-08-09 12:05 ` [PATCH 13/27] arm64/sve: Signal handling support Dave Martin
2017-08-23  9:38   ` Alex Bennée
2017-08-23  9:38     ` Alex Bennée
2017-08-23 11:30     ` Dave Martin
2017-08-09 12:05 ` [PATCH 14/27] arm64/sve: Backend logic for setting the vector length Dave Martin
2017-08-23 15:33   ` Alex Bennée
2017-08-23 15:33     ` Alex Bennée
2017-08-23 17:29     ` Dave Martin
2017-08-09 12:05 ` [PATCH 15/27] arm64/sve: Probe SVE capabilities and usable vector lengths Dave Martin
2017-08-16 17:48   ` Suzuki K Poulose
2017-08-17 10:04     ` Dave Martin
2017-08-17 10:04       ` Dave Martin
2017-08-17 10:46       ` Suzuki K Poulose
2017-08-17 10:46         ` Suzuki K Poulose
2017-08-09 12:05 ` [PATCH 16/27] arm64/sve: Preserve SVE registers around kernel-mode NEON use Dave Martin
2017-08-09 12:05   ` Dave Martin
2017-08-15 17:37   ` Ard Biesheuvel
2017-08-15 17:37     ` Ard Biesheuvel
2017-08-09 12:05 ` [PATCH 17/27] arm64/sve: Preserve SVE registers around EFI runtime service calls Dave Martin
2017-08-15 17:44   ` Ard Biesheuvel
2017-08-16  9:13     ` Dave Martin
2017-08-09 12:05 ` [PATCH 18/27] arm64/sve: ptrace and ELF coredump support Dave Martin
2017-08-09 12:05 ` [PATCH 19/27] arm64/sve: Add prctl controls for userspace vector length management Dave Martin
2017-08-09 12:05 ` [PATCH 20/27] arm64/sve: Add sysctl to set the default vector length for new processes Dave Martin
2017-08-09 12:05 ` [PATCH 21/27] arm64/sve: KVM: Prevent guests from using SVE Dave Martin
2017-08-09 12:05   ` Dave Martin
2017-08-15 16:33   ` Marc Zyngier
2017-08-15 16:33     ` Marc Zyngier
2017-08-16 10:50     ` Dave Martin
2017-08-16 11:20       ` Marc Zyngier
2017-08-16 11:22         ` Marc Zyngier
2017-08-16 11:35         ` Dave Martin
2017-08-09 12:05 ` [PATCH 22/27] arm64/sve: KVM: Treat guest SVE use as undefined instruction execution Dave Martin
2017-08-09 12:05   ` Dave Martin
2017-08-09 12:05 ` [PATCH 23/27] arm64/sve: KVM: Hide SVE from CPU features exposed to guests Dave Martin
2017-08-15 16:37   ` Marc Zyngier
2017-08-16 10:54     ` Dave Martin
2017-08-16 11:10       ` Marc Zyngier
2017-08-16 11:22         ` Dave Martin
2017-08-09 12:05 ` [PATCH 24/27] arm64/sve: Detect SVE and activate runtime support Dave Martin
2017-08-16 17:53   ` Suzuki K Poulose
2017-08-17 10:00     ` Dave Martin
2017-08-17 10:00       ` Dave Martin
2017-08-09 12:05 ` [PATCH 25/27] arm64/sve: Add documentation Dave Martin
2017-08-09 12:05 ` [RFC PATCH 26/27] arm64: signal: Report signal frame size to userspace via auxv Dave Martin
2017-08-09 12:05 ` [RFC PATCH 27/27] arm64/sve: signal: Include SVE when computing AT_MINSIGSTKSZ Dave Martin

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=1502280338-23002-13-git-send-email-Dave.Martin@arm.com \
    --to=dave.martin@arm.com \
    --cc=ard.biesheuvel@linaro.org \
    --cc=catalin.marinas@arm.com \
    --cc=kvmarm@lists.cs.columbia.edu \
    --cc=libc-alpha@sourceware.org \
    --cc=linux-arch@vger.kernel.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=richard.sandiford@arm.com \
    --cc=szabolcs.nagy@arm.com \
    --cc=will.deacon@arm.com \
    /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).