linux-kernel - Re: [PATCH] drm/vmwgfx: switch from ioremap

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Mon, 19 Oct 2015 21:34:11 +0000
From:	"Williams, Dan J" <dan.j.williams@...el.com>
To:	"thellstrom@...are.com" <thellstrom@...are.com>
CC:	"dri-devel@...ts.freedesktop.org" <dri-devel@...ts.freedesktop.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"syeh@...are.com" <syeh@...are.com>,
	"airlied@...ux.ie" <airlied@...ux.ie>
Subject: Re: [PATCH] drm/vmwgfx: switch from ioremap_cache to memremap

On Tue, 2015-10-13 at 20:52 +0200, Thomas Hellstrom wrote:
> > Ok, I'll make local read_fifo() and write_fifo() macros to make this
> > explicit.  Are these names ok with you?
> 
> Sure.
> 

So I ended up just leaving the __iomem annotation on mmio_virt for now
until the implementation can be converted to use explicit barriers where
necessary.

8<-----
Subject: drm/vmwgfx: switch from ioremap_cache to memremap

From: Dan Williams <dan.j.williams@...el.com>

Per commit 2e586a7e017a "drm/vmwgfx: Map the fifo as cached" the driver
expects the fifo registers to be cacheable.  In preparation for
deprecating ioremap_cache() convert its usage in vmwgfx to memremap().

Cc: David Airlie <airlied@...ux.ie>
Cc: Thomas Hellstrom <thellstrom@...are.com>
Cc: Sinclair Yeh <syeh@...are.com>
Cc: dri-devel@...ts.freedesktop.org
Signed-off-by: Dan Williams <dan.j.williams@...el.com>
---
 drivers/gpu/drm/vmwgfx/vmwgfx_drv.c |   14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 2c7a25c71af2..33e9eda77bad 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -752,8 +752,14 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 	ttm_lock_set_kill(&dev_priv->fbdev_master.lock, false, SIGTERM);
 	dev_priv->active_master = &dev_priv->fbdev_master;
 
-	dev_priv->mmio_virt = ioremap_cache(dev_priv->mmio_start,
-					    dev_priv->mmio_size);
+	/*
+	 * Force __iomem for this mapping until the implied compiler
+	 * barriers and {READ|WRITE}_ONCE semantics from the
+	 * io{read|write}32() accessors can be replaced with explicit
+	 * barriers.
+	 */
+	dev_priv->mmio_virt = (void __iomem *) memremap(dev_priv->mmio_start,
+					    dev_priv->mmio_size, MEMREMAP_WB);
 
 	if (unlikely(dev_priv->mmio_virt == NULL)) {
 		ret = -ENOMEM;
@@ -907,7 +913,7 @@ out_no_irq:
 out_no_device:
 	ttm_object_device_release(&dev_priv->tdev);
 out_err4:
-	iounmap(dev_priv->mmio_virt);
+	memunmap((void __force *) dev_priv->mmio_virt);
 out_err3:
 	vmw_ttm_global_release(dev_priv);
 out_err0:
@@ -958,7 +964,7 @@ static int vmw_driver_unload(struct drm_device *dev)
 		pci_release_regions(dev->pdev);
 
 	ttm_object_device_release(&dev_priv->tdev);
-	iounmap(dev_priv->mmio_virt);
+	memunmap((void __force *) dev_priv->mmio_virt);
 	if (dev_priv->ctx.staged_bindings)
 		vmw_binding_state_free(dev_priv->ctx.staged_bindings);
 	vmw_ttm_global_release(dev_priv);