linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: "Thomas Hellström (VMware)" <thomas_os@shipmail.org>
To: linux-mm@kvack.org, dri-devel@lists.freedesktop.org,
	linux-kernel@vger.kernel.org
Cc: pv-drivers@vmware.com, linux-graphics-maintainer@vmware.com,
	"Thomas Hellstrom" <thellstrom@vmware.com>,
	"Andrew Morton" <akpm@linux-foundation.org>,
	"Michal Hocko" <mhocko@suse.com>,
	"Matthew Wilcox (Oracle)" <willy@infradead.org>,
	"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>,
	"Ralph Campbell" <rcampbell@nvidia.com>,
	"Jérôme Glisse" <jglisse@redhat.com>,
	"Christian König" <christian.koenig@amd.com>,
	"Dan Williams" <dan.j.williams@intel.com>,
	"Roland Scheidegger" <sroland@vmware.com>
Subject: [PATCH v5 7/9] drm: Add a drm_get_unmapped_area() helper
Date: Tue,  3 Mar 2020 11:22:45 +0100	[thread overview]
Message-ID: <20200303102247.4635-8-thomas_os@shipmail.org> (raw)
In-Reply-To: <20200303102247.4635-1-thomas_os@shipmail.org>

From: Thomas Hellstrom <thellstrom@vmware.com>

Unaligned virtual addresses makes it unlikely that huge page-table entries
can be used.
So align virtual buffer object address huge page boundaries to the
underlying physical address huge page boundaries taking buffer object
sizes into account to determine when it might be possible to use huge
page-table entries.

Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/drm_file.c | 136 +++++++++++++++++++++++++++++++++++++
 include/drm/drm_file.h     |   5 ++
 2 files changed, 141 insertions(+)

diff --git a/drivers/gpu/drm/drm_file.c b/drivers/gpu/drm/drm_file.c
index 92d16724f949..40fae356d202 100644
--- a/drivers/gpu/drm/drm_file.c
+++ b/drivers/gpu/drm/drm_file.c
@@ -39,10 +39,13 @@
 #include <linux/poll.h>
 #include <linux/slab.h>
 
+#include <uapi/asm/mman.h>
+
 #include <drm/drm_client.h>
 #include <drm/drm_drv.h>
 #include <drm/drm_file.h>
 #include <drm/drm_print.h>
+#include <drm/drm_vma_manager.h>
 
 #include "drm_crtc_internal.h"
 #include "drm_internal.h"
@@ -796,3 +799,136 @@ struct file *mock_drm_getfile(struct drm_minor *minor, unsigned int flags)
 	return file;
 }
 EXPORT_SYMBOL_FOR_TESTS_ONLY(mock_drm_getfile);
+
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+/*
+ * drm_addr_inflate() attempts to construct an aligned area by inflating
+ * the area size and skipping the unaligned start of the area.
+ * adapted from shmem_get_unmapped_area()
+ */
+static unsigned long drm_addr_inflate(unsigned long addr,
+				      unsigned long len,
+				      unsigned long pgoff,
+				      unsigned long flags,
+				      unsigned long huge_size)
+{
+	unsigned long offset, inflated_len;
+	unsigned long inflated_addr;
+	unsigned long inflated_offset;
+
+	offset = (pgoff << PAGE_SHIFT) & (huge_size - 1);
+	if (offset && offset + len < 2 * huge_size)
+		return addr;
+	if ((addr & (huge_size - 1)) == offset)
+		return addr;
+
+	inflated_len = len + huge_size - PAGE_SIZE;
+	if (inflated_len > TASK_SIZE)
+		return addr;
+	if (inflated_len < len)
+		return addr;
+
+	inflated_addr = current->mm->get_unmapped_area(NULL, 0, inflated_len,
+						       0, flags);
+	if (IS_ERR_VALUE(inflated_addr))
+		return addr;
+	if (inflated_addr & ~PAGE_MASK)
+		return addr;
+
+	inflated_offset = inflated_addr & (huge_size - 1);
+	inflated_addr += offset - inflated_offset;
+	if (inflated_offset > offset)
+		inflated_addr += huge_size;
+
+	if (inflated_addr > TASK_SIZE - len)
+		return addr;
+
+	return inflated_addr;
+}
+
+/**
+ * drm_get_unmapped_area() - Get an unused user-space virtual memory area
+ * suitable for huge page table entries.
+ * @file: The struct file representing the address space being mmap()'d.
+ * @uaddr: Start address suggested by user-space.
+ * @len: Length of the area.
+ * @pgoff: The page offset into the address space.
+ * @flags: mmap flags
+ * @mgr: The address space manager used by the drm driver. This argument can
+ * probably be removed at some point when all drivers use the same
+ * address space manager.
+ *
+ * This function attempts to find an unused user-space virtual memory area
+ * that can accommodate the size we want to map, and that is properly
+ * aligned to facilitate huge page table entries matching actual
+ * huge pages or huge page aligned memory in buffer objects. Buffer objects
+ * are assumed to start at huge page boundary pfns (io memory) or be
+ * populated by huge pages aligned to the start of the buffer object
+ * (system- or coherent memory). Adapted from shmem_get_unmapped_area.
+ *
+ * Return: aligned user-space address.
+ */
+unsigned long drm_get_unmapped_area(struct file *file,
+				    unsigned long uaddr, unsigned long len,
+				    unsigned long pgoff, unsigned long flags,
+				    struct drm_vma_offset_manager *mgr)
+{
+	unsigned long addr;
+	unsigned long inflated_addr;
+	struct drm_vma_offset_node *node;
+
+	if (len > TASK_SIZE)
+		return -ENOMEM;
+
+	/*
+	 * @pgoff is the file page-offset the huge page boundaries of
+	 * which typically aligns to physical address huge page boundaries.
+	 * That's not true for DRM, however, where physical address huge
+	 * page boundaries instead are aligned with the offset from
+	 * buffer object start. So adjust @pgoff to be the offset from
+	 * buffer object start.
+	 */
+	drm_vma_offset_lock_lookup(mgr);
+	node = drm_vma_offset_lookup_locked(mgr, pgoff, 1);
+	if (node)
+		pgoff -= node->vm_node.start;
+	drm_vma_offset_unlock_lookup(mgr);
+
+	addr = current->mm->get_unmapped_area(file, uaddr, len, pgoff, flags);
+	if (IS_ERR_VALUE(addr))
+		return addr;
+	if (addr & ~PAGE_MASK)
+		return addr;
+	if (addr > TASK_SIZE - len)
+		return addr;
+
+	if (len < HPAGE_PMD_SIZE)
+		return addr;
+	if (flags & MAP_FIXED)
+		return addr;
+	/*
+	 * Our priority is to support MAP_SHARED mapped hugely;
+	 * and support MAP_PRIVATE mapped hugely too, until it is COWed.
+	 * But if caller specified an address hint, respect that as before.
+	 */
+	if (uaddr)
+		return addr;
+
+	inflated_addr = drm_addr_inflate(addr, len, pgoff, flags,
+					 HPAGE_PMD_SIZE);
+
+	if (IS_ENABLED(CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD) &&
+	    len >= HPAGE_PUD_SIZE)
+		inflated_addr = drm_addr_inflate(inflated_addr, len, pgoff,
+						 flags, HPAGE_PUD_SIZE);
+	return inflated_addr;
+}
+#else /* CONFIG_TRANSPARENT_HUGEPAGE */
+unsigned long drm_get_unmapped_area(struct file *file,
+				    unsigned long uaddr, unsigned long len,
+				    unsigned long pgoff, unsigned long flags,
+				    struct drm_vma_offset_manager *mgr)
+{
+	return current->mm->get_unmapped_area(file, uaddr, len, pgoff, flags);
+}
+#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
diff --git a/include/drm/drm_file.h b/include/drm/drm_file.h
index 8b099b347817..06fe3da09b27 100644
--- a/include/drm/drm_file.h
+++ b/include/drm/drm_file.h
@@ -387,6 +387,11 @@ void drm_event_cancel_free(struct drm_device *dev,
 			   struct drm_pending_event *p);
 void drm_send_event_locked(struct drm_device *dev, struct drm_pending_event *e);
 void drm_send_event(struct drm_device *dev, struct drm_pending_event *e);
+struct drm_vma_offset_manager;
+unsigned long drm_get_unmapped_area(struct file *file,
+				    unsigned long uaddr, unsigned long len,
+				    unsigned long pgoff, unsigned long flags,
+				    struct drm_vma_offset_manager *mgr);
 
 struct file *mock_drm_getfile(struct drm_minor *minor, unsigned int flags);
 
-- 
2.21.1



  parent reply	other threads:[~2020-03-03 10:23 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-03 10:22 [PATCH v5 0/9] Huge page-table entries for TTM Thomas Hellström (VMware)
2020-03-03 10:22 ` [PATCH v5 1/9] fs: Constify vma argument to vma_is_dax Thomas Hellström (VMware)
2020-03-03 10:22 ` [PATCH v5 2/9] mm: Introduce vma_is_special_huge Thomas Hellström (VMware)
2020-03-03 10:22 ` [PATCH v5 3/9] mm: Split huge pages on write-notify or COW Thomas Hellström (VMware)
2020-03-03 10:22 ` [PATCH v5 4/9] mm: Add vmf_insert_pfn_xxx_prot() for huge page-table entries Thomas Hellström (VMware)
2020-03-03 10:22 ` [PATCH v5 5/9] drm/ttm, drm/vmwgfx: Support huge TTM pagefaults Thomas Hellström (VMware)
2020-03-03 10:22 ` [PATCH v5 6/9] drm/vmwgfx: Support huge page faults Thomas Hellström (VMware)
2020-03-03 10:22 ` Thomas Hellström (VMware) [this message]
2020-03-03 21:06   ` [PATCH v5 7/9] drm: Add a drm_get_unmapped_area() helper kbuild test robot
2020-03-03 10:22 ` [PATCH v5 8/9] drm/vmwgfx: Introduce a huge page aligning TTM range manager Thomas Hellström (VMware)
2020-03-03 10:22 ` [PATCH v5 9/9] drm/vmwgfx: Hook up the helpers to align buffer objects Thomas Hellström (VMware)

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=20200303102247.4635-8-thomas_os@shipmail.org \
    --to=thomas_os@shipmail.org \
    --cc=akpm@linux-foundation.org \
    --cc=christian.koenig@amd.com \
    --cc=dan.j.williams@intel.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=jglisse@redhat.com \
    --cc=kirill.shutemov@linux.intel.com \
    --cc=linux-graphics-maintainer@vmware.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@suse.com \
    --cc=pv-drivers@vmware.com \
    --cc=rcampbell@nvidia.com \
    --cc=sroland@vmware.com \
    --cc=thellstrom@vmware.com \
    --cc=willy@infradead.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).