[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <lsq.1465842997.389522944@decadent.org.uk>
Date: Mon, 13 Jun 2016 19:36:37 +0100
From: Ben Hutchings <ben@...adent.org.uk>
To: linux-kernel@...r.kernel.org, stable@...r.kernel.org
CC: akpm@...ux-foundation.org,
"Chris Wilson" <chris@...is-wilson.co.uk>,
"Tvrtko Ursulin" <tvrtko.ursulin@...el.com>,
"Jani Nikula" <jani.nikula@...el.com>,
"Michał Winiarski" <michal.winiarski@...el.com>
Subject: [PATCH 3.16 032/114] drm/i915/userptr: Hold mmref whilst calling
get-user-pages
3.16.36-rc1 review patch. If anyone has any objections, please let me know.
------------------
From: Chris Wilson <chris@...is-wilson.co.uk>
commit db9f9203e27495b2d151b695504e286eec026e9b upstream.
Holding a reference to the containing task_struct is not sufficient to
prevent the mm_struct from being reaped under memory pressure. If this
happens whilst we are calling get_user_pages(), explosions erupt -
sometimes an immediate GPF, sometimes page flag corruption. To prevent
the target mm from being reaped as we are reading from it, acquire a
reference before we begin.
Testcase: igt/gem_shrink/*userptr
Signed-off-by: Chris Wilson <chris@...is-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@...el.com>
Cc: Michał Winiarski <michal.winiarski@...el.com>
Reviewed-by: Michał Winiarski <michal.winiarski@...el.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1459864801-28606-2-git-send-email-chris@chris-wilson.co.uk
(cherry picked from commit 40313f0cd0b711a7a5905e5182422799e157d8aa)
Signed-off-by: Jani Nikula <jani.nikula@...el.com>
[bwh: Backported to 3.16:
- s/get_user_pages_remote/get_user_pages/
- s/npages/num_pages/]
Signed-off-by: Ben Hutchings <ben@...adent.org.uk>
---
drivers/gpu/drm/i915/i915_gem_userptr.c | 29 +++++++++++++++++------------
1 file changed, 17 insertions(+), 12 deletions(-)
--- a/drivers/gpu/drm/i915/i915_gem_userptr.c
+++ b/drivers/gpu/drm/i915/i915_gem_userptr.c
@@ -411,19 +411,24 @@ __i915_gem_userptr_get_pages_worker(stru
if (pvec != NULL) {
struct mm_struct *mm = obj->userptr.mm;
- down_read(&mm->mmap_sem);
- while (pinned < num_pages) {
- ret = get_user_pages(work->task, mm,
- obj->userptr.ptr + pinned * PAGE_SIZE,
- num_pages - pinned,
- !obj->userptr.read_only, 0,
- pvec + pinned, NULL);
- if (ret < 0)
- break;
+ ret = -EFAULT;
+ if (atomic_inc_not_zero(&mm->mm_users)) {
+ down_read(&mm->mmap_sem);
+ while (pinned < num_pages) {
+ ret = get_user_pages
+ (work->task, mm,
+ obj->userptr.ptr + pinned * PAGE_SIZE,
+ num_pages - pinned,
+ !obj->userptr.read_only, 0,
+ pvec + pinned, NULL);
+ if (ret < 0)
+ break;
- pinned += ret;
+ pinned += ret;
+ }
+ up_read(&mm->mmap_sem);
+ mmput(mm);
}
- up_read(&mm->mmap_sem);
}
mutex_lock(&dev->struct_mutex);
Powered by blists - more mailing lists