From patchwork Fri Jan 15 22:07:44 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Chris Wilson X-Patchwork-Id: 12024233 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 89DB4C433E0 for ; Fri, 15 Jan 2021 22:08:03 +0000 (UTC) Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id D5961239EE for ; Fri, 15 Jan 2021 22:08:02 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org D5961239EE Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=chris-wilson.co.uk Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=intel-gfx-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 5605D6E4A7; Fri, 15 Jan 2021 22:08:02 +0000 (UTC) Received: from fireflyinternet.com (unknown [77.68.26.236]) by gabe.freedesktop.org (Postfix) with ESMTPS id 0A60F6E4A5 for ; Fri, 15 Jan 2021 22:08:00 +0000 (UTC) X-Default-Received-SPF: pass (skip=forwardok (res=PASS)) x-ip-name=78.156.65.138; Received: from haswell.alporthouse.com (unverified [78.156.65.138]) by fireflyinternet.com (Firefly Internet (M1)) with ESMTP id 23611775-1500050 for multiple; Fri, 15 Jan 2021 22:07:46 +0000 From: Chris Wilson To: intel-gfx@lists.freedesktop.org Date: Fri, 15 Jan 2021 22:07:44 +0000 Message-Id: <20210115220744.1402642-1-chris@chris-wilson.co.uk> X-Mailer: git-send-email 2.30.0 MIME-Version: 1.0 Subject: [Intel-gfx] [PATCH xf86-video-intel] sna: Probe userptr storage on construction X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Chris Wilson Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" Not all user address ranges are valid for userptr, as we can only utilize struct pages (and not indirect pfnmaps). Currently we use a set-domain call after userptr to validate the range by populating the userptr, but this has the drawback of forcing the population ever for userptr that are never used. I915_USERPTR_PROBE introduces a method for doing a simple vma lookup to check that the userptr is valid at the time of construction, leaving the more costly page lookup for lazy execution. See https://patchwork.freedesktop.org/series/33449/ Signed-off-by: Chris Wilson --- src/sna/kgem.c | 49 ++++++++++++++++++++++++++++++++++++++++++++----- src/sna/kgem.h | 1 + 2 files changed, 45 insertions(+), 5 deletions(-) diff --git a/src/sna/kgem.c b/src/sna/kgem.c index 7b645da8d..7b143dd34 100644 --- a/src/sna/kgem.c +++ b/src/sna/kgem.c @@ -170,6 +170,8 @@ struct local_i915_gem_userptr { uint64_t user_size; uint32_t flags; #define I915_USERPTR_READ_ONLY 0x1 +#define I915_USERPTR_PROBE 0x2 +#define I915_USERPTR_POPULATE 0x4 #define I915_USERPTR_UNSYNCHRONIZED 0x80000000 uint32_t handle; }; @@ -520,16 +522,14 @@ static bool gem_set_caching(int fd, uint32_t handle, int caching) return do_ioctl(fd, LOCAL_IOCTL_I915_GEM_SET_CACHING, &arg) == 0; } -static uint32_t gem_userptr(int fd, void *ptr, size_t size, int read_only) +static uint32_t gem_userptr(int fd, void *ptr, size_t size, unsigned int flags) { struct local_i915_gem_userptr arg; VG_CLEAR(arg); arg.user_ptr = (uintptr_t)ptr; arg.user_size = size; - arg.flags = I915_USERPTR_UNSYNCHRONIZED; - if (read_only) - arg.flags |= I915_USERPTR_READ_ONLY; + arg.flags = flags | I915_USERPTR_UNSYNCHRONIZED; if (DBG_NO_UNSYNCHRONIZED_USERPTR || do_ioctl(fd, LOCAL_IOCTL_I915_GEM_USERPTR, &arg)) { @@ -1390,6 +1390,29 @@ static bool test_has_userptr(struct kgem *kgem) return arg.handle != 0; } +static bool test_has_userptr_probe(struct kgem *kgem) +{ + struct local_i915_gem_userptr arg; + void *ptr; + + if (DBG_NO_USERPTR) + return false; + + if (posix_memalign(&ptr, PAGE_SIZE, PAGE_SIZE)) + return false; + + VG_CLEAR(arg); + arg.user_ptr = (uintptr_t)ptr; + arg.user_size = PAGE_SIZE; + arg.flags = I915_USERPTR_PROBE; + + if (do_ioctl(kgem->fd, LOCAL_IOCTL_I915_GEM_USERPTR, &arg) == 0) + gem_close(kgem->fd, arg.handle); + + free(ptr); + return arg.handle != 0; +} + static bool test_has_create2(struct kgem *kgem) { #if defined(USE_CREATE2) @@ -2034,6 +2057,12 @@ void kgem_init(struct kgem *kgem, int fd, struct pci_device *dev, unsigned gen) DBG(("%s: has userptr? %d\n", __FUNCTION__, kgem->has_userptr)); + kgem->has_userptr_probe = false; + if (kgem->has_userptr) + kgem->has_userptr_probe = test_has_userptr_probe(kgem); + DBG(("%s: has userptr probe? %d\n", __FUNCTION__, + kgem->has_userptr_probe)); + kgem->has_create2 = test_has_create2(kgem); DBG(("%s: has create2? %d\n", __FUNCTION__, kgem->has_create2)); @@ -7038,15 +7067,25 @@ static bool probe(struct kgem *kgem, uint32_t handle) .read_domains = I915_GEM_DOMAIN_CPU, }; + if (kgem->has_userptr_probe) + return true; + return do_ioctl(kgem->fd, DRM_IOCTL_I915_GEM_SET_DOMAIN, &arg) == 0; } static uint32_t probe_userptr(struct kgem *kgem, void *ptr, size_t size, int read_only) { + unsigned int flags; uint32_t handle; - handle = gem_userptr(kgem->fd, ptr, size, read_only); + flags = 0; + if (read_only) + flags |= I915_USERPTR_READ_ONLY; + if (kgem->has_userptr_probe) + flags |= I915_USERPTR_PROBE; + + handle = gem_userptr(kgem->fd, ptr, size, flags); if (handle && !probe(kgem, handle)) { gem_close(kgem->fd, handle); handle = 0; diff --git a/src/sna/kgem.h b/src/sna/kgem.h index 6a087a574..b62f956c6 100644 --- a/src/sna/kgem.h +++ b/src/sna/kgem.h @@ -183,6 +183,7 @@ struct kgem { uint32_t has_create2 :1; uint32_t has_userptr :1; + uint32_t has_userptr_probe :1; uint32_t has_blt :1; uint32_t has_relaxed_fencing :1; uint32_t has_relaxed_delta :1;