This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
[RFC PATCH 06/10] arm64/sve: Disallow VL setting for individual threads by default
- From: Dave Martin <Dave dot Martin at arm dot com>
- To: linux-arm-kernel at lists dot infradead dot org
- Cc: Ard Biesheuvel <ard dot biesheuvel at linaro dot org>, Marc Zyngier <Marc dot Zyngier at arm dot com>, Alan Hayward <alan dot hayward at arm dot com>, Christoffer Dall <christoffer dot dall at linaro dot org>, linux-arch at vger dot kernel dot org, libc-alpha at sourceware dot org, Florian Weimer <fweimer at redhat dot com>, Joseph Myers <joseph at codesourcery dot com>, Szabolcs Nagy <szabolcs dot nagy at arm dot com>, Torvald Riegel <triegel at redhat dot com>, gdb at sourceware dot org, Yao Qi <qiyaoltc at gmail dot com>
- Date: Thu, 12 Jan 2017 11:26:05 +0000
- Subject: [RFC PATCH 06/10] arm64/sve: Disallow VL setting for individual threads by default
- Authentication-results: sourceware.org; auth=none
- References: <1484220369-23970-1-git-send-email-Dave.Martin@arm.com>
General-purpose code in userspace is not expected to work correctly
if multiple threads are allowed to run concurrently with different
vector lengths in a single process.
This patch adds an explicit flag PR_SVE_SET_VL_THREAD to request
this behaviour. Without the flag, vector length setting is
permitted only for a single-threaded process (which matches the
expected usage model of setting the vector length at process
startup).
Signed-off-by: Dave Martin <Dave.Martin@arm.com>
---
arch/arm64/kernel/fpsimd.c | 12 +++++++++++-
include/uapi/linux/prctl.h | 1 +
2 files changed, 12 insertions(+), 1 deletion(-)
diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
index 5f2c24a..32caca3 100644
--- a/arch/arm64/kernel/fpsimd.c
+++ b/arch/arm64/kernel/fpsimd.c
@@ -21,6 +21,7 @@
#include <linux/cpu_pm.h>
#include <linux/kernel.h>
#include <linux/init.h>
+#include <linux/prctl.h>
#include <linux/sched.h>
#include <linux/signal.h>
#include <linux/hardirq.h>
@@ -498,8 +499,17 @@ int sve_set_vector_length(struct task_struct *task,
{
BUG_ON(task == current && preemptible());
+ /*
+ * To avoid accidents, forbid setting for individual threads of a
+ * multithreaded process. User code that knows what it's doing can
+ * pass PR_SVE_SET_VL_THREAD to override this restriction:
+ */
+ if (!(flags & PR_SVE_SET_VL_THREAD) && get_nr_threads(task) != 1)
+ return -EINVAL;
+
+ flags &= ~(unsigned long)PR_SVE_SET_VL_THREAD;
if (flags)
- return -EINVAL; /* No flags defined yet */
+ return -EINVAL; /* No other flags defined yet */
if (!sve_vl_valid(vl))
return -EINVAL;
diff --git a/include/uapi/linux/prctl.h b/include/uapi/linux/prctl.h
index e32e2da..c55530b 100644
--- a/include/uapi/linux/prctl.h
+++ b/include/uapi/linux/prctl.h
@@ -199,6 +199,7 @@ struct prctl_mm_map {
/* arm64 Scalable Vector Extension controls */
#define PR_SVE_SET_VL 48 /* set task vector length */
+# define PR_SVE_SET_VL_THREAD (1 << 1) /* set just this thread */
#define PR_SVE_GET_VL 49 /* get task vector length */
#endif /* _LINUX_PRCTL_H */
--
2.1.4