drm/linux-core/drm_agpsupport.c

709 lines
18 KiB
C
Raw Normal View History

2003-05-26 18:37:33 -06:00
/**
* \file drm_agpsupport.c
2003-05-26 18:37:33 -06:00
* DRM support for AGP/GART backend
*
2003-05-26 18:37:33 -06:00
* \author Rickard E. (Rik) Faith <faith@valinux.com>
* \author Gareth Hughes <gareth@valinux.com>
*/
/*
2000-02-22 08:43:59 -07:00
* Copyright 1999 Precision Insight, Inc., Cedar Park, Texas.
2000-06-08 08:38:22 -06:00
* Copyright 2000 VA Linux Systems, Inc., Sunnyvale, California.
2000-02-22 08:43:59 -07:00
* All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
2001-02-15 01:12:14 -07:00
*
2000-02-22 08:43:59 -07:00
* The above copyright notice and this permission notice (including the next
* paragraph) shall be included in all copies or substantial portions of the
* Software.
2001-02-15 01:12:14 -07:00
*
2000-02-22 08:43:59 -07:00
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
2001-02-15 01:12:14 -07:00
* VA LINUX SYSTEMS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
2000-02-22 08:43:59 -07:00
* OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
2001-02-15 01:12:14 -07:00
* ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
* OTHER DEALINGS IN THE SOFTWARE.
2000-02-22 08:43:59 -07:00
*/
#include "drmP.h"
2000-08-04 04:04:55 -06:00
#include <linux/module.h>
2001-02-15 01:12:14 -07:00
2004-08-24 05:15:53 -06:00
#if __OS_HAS_AGP
2003-05-26 18:37:33 -06:00
/**
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
* Get AGP information.
2003-05-26 18:37:33 -06:00
*
* \param inode device inode.
* \param filp file pointer.
* \param cmd command.
* \param arg pointer to a (output) drm_agp_info structure.
* \return zero on success or a negative number on failure.
*
* Verifies the AGP device has been initialized and acquired and fills in the
* drm_agp_info structure with the information in drm_agp_head::agp_info.
*/
2007-07-15 20:32:51 -06:00
int drm_agp_info(struct drm_device * dev, struct drm_agp_info *info)
{
DRM_AGP_KERN *kern;
2000-02-22 08:43:59 -07:00
if (!dev->agp || !dev->agp->acquired)
return -EINVAL;
2000-02-22 08:43:59 -07:00
kern = &dev->agp->agp_info;
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
info->agp_version_major = kern->version.major;
info->agp_version_minor = kern->version.minor;
info->mode = kern->mode;
info->aperture_base = kern->aper_base;
info->aperture_size = kern->aper_size * 1024 * 1024;
info->memory_allowed = kern->max_memory << PAGE_SHIFT;
info->memory_used = kern->current_memory << PAGE_SHIFT;
info->id_vendor = kern->device->vendor;
info->id_device = kern->device->device;
return 0;
}
EXPORT_SYMBOL(drm_agp_info);
int drm_agp_info_ioctl(struct inode *inode, struct file *filp,
unsigned int cmd, unsigned long arg)
{
2007-07-15 20:32:51 -06:00
struct drm_file *priv = filp->private_data;
struct drm_device *dev = priv->head->dev;
struct drm_agp_info info;
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
int err;
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
err = drm_agp_info(dev, &info);
if (err)
return err;
if (copy_to_user((struct drm_agp_info __user *) arg, &info, sizeof(info)))
return -EFAULT;
return 0;
}
2000-02-22 08:43:59 -07:00
2003-05-26 18:37:33 -06:00
/**
* Acquire the AGP device.
2003-05-26 18:37:33 -06:00
*
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
* \param dev DRM device that is to acquire AGP.
* \return zero on success or a negative number on failure.
2003-05-26 18:37:33 -06:00
*
* Verifies the AGP device hasn't been acquired before and calls
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
* \c agp_backend_acquire.
2003-05-26 18:37:33 -06:00
*/
2007-07-15 20:32:51 -06:00
int drm_agp_acquire(struct drm_device * dev)
2000-02-22 08:43:59 -07:00
{
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,11)
int retcode;
#endif
2000-02-22 08:43:59 -07:00
if (!dev->agp)
return -ENODEV;
if (dev->agp->acquired)
return -EBUSY;
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,11)
if ((retcode = agp_backend_acquire()))
2003-08-15 04:31:54 -06:00
return retcode;
#else
if (!(dev->agp->bridge = agp_backend_acquire(dev->pdev)))
return -ENODEV;
#endif
2000-02-22 08:43:59 -07:00
dev->agp->acquired = 1;
return 0;
}
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
EXPORT_SYMBOL(drm_agp_acquire);
2000-02-22 08:43:59 -07:00
2003-05-26 18:37:33 -06:00
/**
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
* Acquire the AGP device (ioctl).
2003-05-26 18:37:33 -06:00
*
* \param inode device inode.
* \param filp file pointer.
* \param cmd command.
* \param arg user argument.
* \return zero on success or a negative number on failure.
*
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
* Verifies the AGP device hasn't been acquired before and calls
* \c agp_backend_acquire.
2003-05-26 18:37:33 -06:00
*/
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
int drm_agp_acquire_ioctl(struct inode *inode, struct file *filp,
unsigned int cmd, unsigned long arg)
2000-02-22 08:43:59 -07:00
{
2007-07-15 20:32:51 -06:00
struct drm_file *priv = filp->private_data;
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
2007-07-15 20:32:51 -06:00
return drm_agp_acquire( (struct drm_device *) priv->head->dev );
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
}
2000-02-22 08:43:59 -07:00
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
/**
* Release the AGP device.
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
*
* \param dev DRM device that is to release AGP.
* \return zero on success or a negative number on failure.
*
* Verifies the AGP device has been acquired and calls \c agp_backend_release.
*/
2007-07-15 20:32:51 -06:00
int drm_agp_release(struct drm_device *dev)
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
{
if (!dev->agp || !dev->agp->acquired)
2001-05-01 11:07:59 -06:00
return -EINVAL;
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,11)
agp_backend_release();
#else
agp_backend_release(dev->agp->bridge);
#endif
2000-02-22 08:43:59 -07:00
dev->agp->acquired = 0;
return 0;
2001-02-15 01:12:14 -07:00
2000-02-22 08:43:59 -07:00
}
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
EXPORT_SYMBOL(drm_agp_release);
2000-02-22 08:43:59 -07:00
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
int drm_agp_release_ioctl(struct inode *inode, struct file *filp,
unsigned int cmd, unsigned long arg)
{
2007-07-15 20:32:51 -06:00
struct drm_file *priv = filp->private_data;
struct drm_device *dev = priv->head->dev;
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
return drm_agp_release(dev);
}
2003-05-26 18:37:33 -06:00
/**
* Enable the AGP bus.
*
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
* \param dev DRM device that has previously acquired AGP.
* \param mode Requested AGP mode.
2003-05-26 18:37:33 -06:00
* \return zero on success or a negative number on failure.
*
* Verifies the AGP device has been acquired but not enabled, and calls
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
* \c agp_enable.
2003-05-26 18:37:33 -06:00
*/
2007-07-15 20:32:51 -06:00
int drm_agp_enable(struct drm_device *dev, struct drm_agp_mode mode)
2000-02-22 08:43:59 -07:00
{
if (!dev->agp || !dev->agp->acquired)
2001-05-01 11:07:59 -06:00
return -EINVAL;
2000-02-22 08:43:59 -07:00
dev->agp->mode = mode.mode;
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,11)
agp_enable(mode.mode);
#else
agp_enable(dev->agp->bridge, mode.mode);
#endif
dev->agp->base = dev->agp->agp_info.aper_base;
2000-02-22 08:43:59 -07:00
dev->agp->enabled = 1;
return 0;
}
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
EXPORT_SYMBOL(drm_agp_enable);
int drm_agp_enable_ioctl(struct inode *inode, struct file *filp,
unsigned int cmd, unsigned long arg)
{
2007-07-15 20:32:51 -06:00
struct drm_file *priv = filp->private_data;
struct drm_device *dev = priv->head->dev;
struct drm_agp_mode mode;
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
if (copy_from_user(&mode, (struct drm_agp_mode __user *) arg, sizeof(mode)))
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
return -EFAULT;
return drm_agp_enable(dev, mode);
}
2000-02-22 08:43:59 -07:00
2003-05-26 18:37:33 -06:00
/**
* Allocate AGP memory.
*
* \param inode device inode.
* \param filp file pointer.
* \param cmd command.
* \param arg pointer to a drm_agp_buffer structure.
* \return zero on success or a negative number on failure.
*
2003-05-26 18:37:33 -06:00
* Verifies the AGP device is present and has been acquired, allocates the
* memory via alloc_agp() and creates a drm_agp_mem entry for it.
2003-05-26 18:37:33 -06:00
*/
2007-07-15 20:32:51 -06:00
int drm_agp_alloc(struct drm_device *dev, struct drm_agp_buffer *request)
2000-02-22 08:43:59 -07:00
{
2007-07-15 20:32:51 -06:00
struct drm_agp_mem *entry;
DRM_AGP_MEM *memory;
unsigned long pages;
u32 type;
2001-02-15 01:12:14 -07:00
2003-08-15 04:31:54 -06:00
if (!dev->agp || !dev->agp->acquired)
return -EINVAL;
if (!(entry = drm_alloc(sizeof(*entry), DRM_MEM_AGPLISTS)))
2000-02-22 08:43:59 -07:00
return -ENOMEM;
2001-02-15 01:12:14 -07:00
memset(entry, 0, sizeof(*entry));
2000-02-22 08:43:59 -07:00
pages = (request->size + PAGE_SIZE - 1) / PAGE_SIZE;
type = (u32) request->type;
if (!(memory = drm_alloc_agp(dev, pages, type))) {
drm_free(entry, sizeof(*entry), DRM_MEM_AGPLISTS);
return -ENOMEM;
}
2001-02-15 01:12:14 -07:00
entry->handle = (unsigned long)memory->key + 1;
entry->memory = memory;
entry->bound = 0;
entry->pages = pages;
list_add(&entry->head, &dev->agp->memory);
2000-02-22 08:43:59 -07:00
request->handle = entry->handle;
request->physical = memory->physical;
return 0;
}
EXPORT_SYMBOL(drm_agp_alloc);
int drm_agp_alloc_ioctl(struct inode *inode, struct file *filp,
unsigned int cmd, unsigned long arg)
{
2007-07-15 20:32:51 -06:00
struct drm_file *priv = filp->private_data;
struct drm_device *dev = priv->head->dev;
struct drm_agp_buffer request;
struct drm_agp_buffer __user *argp = (void __user *)arg;
int err;
if (copy_from_user(&request, argp, sizeof(request)))
return -EFAULT;
err = drm_agp_alloc(dev, &request);
if (err)
return err;
2000-02-22 08:43:59 -07:00
if (copy_to_user(argp, &request, sizeof(request))) {
2007-07-15 20:32:51 -06:00
struct drm_agp_mem *entry;
list_for_each_entry(entry, &dev->agp->memory, head) {
if (entry->handle == request.handle)
break;
}
list_del(&entry->head);
drm_free_agp(entry->memory, entry->pages);
drm_free(entry, sizeof(*entry), DRM_MEM_AGPLISTS);
2000-02-22 08:43:59 -07:00
return -EFAULT;
}
2000-02-22 08:43:59 -07:00
return 0;
}
2003-05-26 18:37:33 -06:00
/**
* Search for the AGP memory entry associated with a handle.
*
* \param dev DRM device structure.
* \param handle AGP memory handle.
* \return pointer to the drm_agp_mem structure associated with \p handle.
*
2003-05-26 18:37:33 -06:00
* Walks through drm_agp_head::memory until finding a matching handle.
*/
2007-07-15 20:32:51 -06:00
static struct drm_agp_mem *drm_agp_lookup_entry(struct drm_device * dev,
unsigned long handle)
2000-02-22 08:43:59 -07:00
{
2007-07-15 20:32:51 -06:00
struct drm_agp_mem *entry;
2000-02-22 08:43:59 -07:00
list_for_each_entry(entry, &dev->agp->memory, head) {
2003-08-15 04:31:54 -06:00
if (entry->handle == handle)
return entry;
2000-02-22 08:43:59 -07:00
}
return NULL;
}
2003-05-26 18:37:33 -06:00
/**
* Unbind AGP memory from the GATT (ioctl).
2003-05-26 18:37:33 -06:00
*
* \param inode device inode.
* \param filp file pointer.
* \param cmd command.
* \param arg pointer to a drm_agp_binding structure.
* \return zero on success or a negative number on failure.
*
* Verifies the AGP device is present and acquired, looks-up the AGP memory
* entry and passes it to the unbind_agp() function.
2003-05-26 18:37:33 -06:00
*/
2007-07-15 20:32:51 -06:00
int drm_agp_unbind(struct drm_device *dev, struct drm_agp_binding *request)
2000-02-22 08:43:59 -07:00
{
2007-07-15 20:32:51 -06:00
struct drm_agp_mem *entry;
int ret;
2000-02-22 08:43:59 -07:00
2003-08-15 04:31:54 -06:00
if (!dev->agp || !dev->agp->acquired)
return -EINVAL;
if (!(entry = drm_agp_lookup_entry(dev, request->handle)))
2000-02-22 08:43:59 -07:00
return -EINVAL;
2003-08-15 04:31:54 -06:00
if (!entry->bound)
return -EINVAL;
ret = drm_unbind_agp(entry->memory);
if (ret == 0)
entry->bound = 0;
return ret;
2000-02-22 08:43:59 -07:00
}
EXPORT_SYMBOL(drm_agp_unbind);
int drm_agp_unbind_ioctl(struct inode *inode, struct file *filp,
unsigned int cmd, unsigned long arg)
{
2007-07-15 20:32:51 -06:00
struct drm_file *priv = filp->private_data;
struct drm_device *dev = priv->head->dev;
struct drm_agp_binding request;
if (copy_from_user
(&request, (struct drm_agp_binding __user *) arg, sizeof(request)))
return -EFAULT;
return drm_agp_unbind(dev, &request);
}
2000-02-22 08:43:59 -07:00
2003-05-26 18:37:33 -06:00
/**
* Bind AGP memory into the GATT (ioctl)
2003-05-26 18:37:33 -06:00
*
* \param inode device inode.
* \param filp file pointer.
* \param cmd command.
* \param arg pointer to a drm_agp_binding structure.
* \return zero on success or a negative number on failure.
*
* Verifies the AGP device is present and has been acquired and that no memory
* is currently bound into the GATT. Looks-up the AGP memory entry and passes
* it to bind_agp() function.
2003-05-26 18:37:33 -06:00
*/
2007-07-15 20:32:51 -06:00
int drm_agp_bind(struct drm_device *dev, struct drm_agp_binding *request)
2000-02-22 08:43:59 -07:00
{
2007-07-15 20:32:51 -06:00
struct drm_agp_mem *entry;
int retcode;
int page;
2001-02-15 01:12:14 -07:00
if (!dev->agp || !dev->agp->acquired)
2001-05-01 11:07:59 -06:00
return -EINVAL;
if (!(entry = drm_agp_lookup_entry(dev, request->handle)))
2000-02-22 08:43:59 -07:00
return -EINVAL;
2003-08-15 04:31:54 -06:00
if (entry->bound)
return -EINVAL;
page = (request->offset + PAGE_SIZE - 1) / PAGE_SIZE;
if ((retcode = drm_bind_agp(entry->memory, page)))
2003-08-15 04:31:54 -06:00
return retcode;
2000-02-22 08:43:59 -07:00
entry->bound = dev->agp->base + (page << PAGE_SHIFT);
2001-02-15 01:12:14 -07:00
DRM_DEBUG("base = 0x%lx entry->bound = 0x%lx\n",
2000-05-18 00:14:27 -06:00
dev->agp->base, entry->bound);
2000-02-22 08:43:59 -07:00
return 0;
}
EXPORT_SYMBOL(drm_agp_bind);
int drm_agp_bind_ioctl(struct inode *inode, struct file *filp,
unsigned int cmd, unsigned long arg)
{
2007-07-15 20:32:51 -06:00
struct drm_file *priv = filp->private_data;
struct drm_device *dev = priv->head->dev;
struct drm_agp_binding request;
if (copy_from_user
(&request, (struct drm_agp_binding __user *) arg, sizeof(request)))
return -EFAULT;
return drm_agp_bind(dev, &request);
}
2000-02-22 08:43:59 -07:00
2003-05-26 18:37:33 -06:00
/**
* Free AGP memory (ioctl).
2003-05-26 18:37:33 -06:00
*
* \param inode device inode.
* \param filp file pointer.
* \param cmd command.
* \param arg pointer to a drm_agp_buffer structure.
* \return zero on success or a negative number on failure.
*
* Verifies the AGP device is present and has been acquired and looks up the
* AGP memory entry. If the memory it's currently bound, unbind it via
* unbind_agp(). Frees it via free_agp() as well as the entry itself
2003-05-26 18:37:33 -06:00
* and unlinks from the doubly linked list it's inserted in.
*/
2007-07-15 20:32:51 -06:00
int drm_agp_free(struct drm_device *dev, struct drm_agp_buffer *request)
2000-02-22 08:43:59 -07:00
{
2007-07-15 20:32:51 -06:00
struct drm_agp_mem *entry;
2001-02-15 01:12:14 -07:00
2003-08-15 04:31:54 -06:00
if (!dev->agp || !dev->agp->acquired)
return -EINVAL;
if (!(entry = drm_agp_lookup_entry(dev, request->handle)))
2000-02-22 08:43:59 -07:00
return -EINVAL;
2003-08-15 04:31:54 -06:00
if (entry->bound)
drm_unbind_agp(entry->memory);
2003-08-15 04:31:54 -06:00
list_del(&entry->head);
2001-02-15 01:12:14 -07:00
drm_free_agp(entry->memory, entry->pages);
drm_free(entry, sizeof(*entry), DRM_MEM_AGPLISTS);
2000-02-22 08:43:59 -07:00
return 0;
}
EXPORT_SYMBOL(drm_agp_free);
int drm_agp_free_ioctl(struct inode *inode, struct file *filp,
unsigned int cmd, unsigned long arg)
{
2007-07-15 20:32:51 -06:00
struct drm_file *priv = filp->private_data;
struct drm_device *dev = priv->head->dev;
struct drm_agp_buffer request;
if (copy_from_user
(&request, (struct drm_agp_buffer __user *) arg, sizeof(request)))
return -EFAULT;
return drm_agp_free(dev, &request);
}
2000-02-22 08:43:59 -07:00
2003-05-26 18:37:33 -06:00
/**
* Initialize the AGP resources.
2003-05-26 18:37:33 -06:00
*
* \return pointer to a drm_agp_head structure.
*
* Gets the drm_agp_t structure which is made available by the agpgart module
* via the inter_module_* functions. Creates and initializes a drm_agp_head
* structure.
*/
2007-07-15 20:32:51 -06:00
struct drm_agp_head *drm_agp_init(struct drm_device *dev)
{
2007-07-15 20:32:51 -06:00
struct drm_agp_head *head = NULL;
if (!(head = drm_alloc(sizeof(*head), DRM_MEM_AGPLISTS)))
return NULL;
memset((void *)head, 0, sizeof(*head));
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,11)
agp_copy_info(&head->agp_info);
#else
head->bridge = agp_find_bridge(dev->pdev);
if (!head->bridge) {
if (!(head->bridge = agp_backend_acquire(dev->pdev))) {
drm_free(head, sizeof(*head), DRM_MEM_AGPLISTS);
return NULL;
}
agp_copy_info(head->bridge, &head->agp_info);
agp_backend_release(head->bridge);
} else {
agp_copy_info(head->bridge, &head->agp_info);
}
#endif
if (head->agp_info.chipset == NOT_SUPPORTED) {
drm_free(head, sizeof(*head), DRM_MEM_AGPLISTS);
return NULL;
}
INIT_LIST_HEAD(&head->memory);
head->cant_use_aperture = head->agp_info.cant_use_aperture;
head->page_mask = head->agp_info.page_mask;
return head;
}
/** Calls agp_allocate_memory() */
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,11)
DRM_AGP_MEM *drm_agp_allocate_memory(size_t pages, u32 type)
{
return agp_allocate_memory(pages, type);
}
#else
DRM_AGP_MEM *drm_agp_allocate_memory(struct agp_bridge_data *bridge,
size_t pages, u32 type)
{
return agp_allocate_memory(bridge, pages, type);
}
#endif
/** Calls agp_free_memory() */
int drm_agp_free_memory(DRM_AGP_MEM * handle)
{
if (!handle)
2003-08-15 04:31:54 -06:00
return 0;
agp_free_memory(handle);
return 1;
}
2001-02-15 01:12:14 -07:00
/** Calls agp_bind_memory() */
int drm_agp_bind_memory(DRM_AGP_MEM * handle, off_t start)
{
if (!handle)
2003-08-15 04:31:54 -06:00
return -EINVAL;
return agp_bind_memory(handle, start);
}
Adds support for PCI cards to MGA DRM This patch adds serveral new ioctls and a new query to get_param query to support PCI MGA cards. Two ioctls were added to implement interrupt based waiting. With this change, the client-side driver no longer needs to map the primary DMA region or the MMIO region. Previously, end-of-frame waiting was done by busy waiting in the client-side driver until one of the MMIO registers (the current DMA pointer) matched a pointer to the end of primary DMA space. By using interrupts, the busy waiting and the extra mappings are removed. A third ioctl was added to bootstrap DMA. This ioctl, which is used by the X-server, moves a *LOT* of code from the X-server into the kernel. This allows the kernel to do whatever needs to be done to setup DMA buffers. The entire process and the locations of the buffers are hidden from user-mode. Additionally, a get_param query was added to differentiate between G4x0 cards and G550 cards. A gap was left in the numbering sequence so that, if needed, G450 cards could be distinguished from G400 cards. According to Ville Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering differently. This seems the most compatible way to let the client-side driver know which card it's own. Doing this very small change now eliminates the need to bump the DRM minor version twice. http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2 A number of ioctl handlers in linux-core were also modified so that they could be called in-kernel. In these cases, the in-kernel callable version kept the existing name (e.g., drm_agp_acquire) and the ioctl handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl). This patch also replaces the drm_agp_do_release function with drm_agp_release. drm_agp_release (drm_core_agp_release in the previous patch) is very similar to drm_agp_do_release, and I saw no reason to have both. This commit *breaks the build* on BSD. Eric said that he would make the required updates to the BSD side soon. Xorg bug: 3259 Reviewed by: Eric Anholt
2005-06-14 16:34:11 -06:00
EXPORT_SYMBOL(drm_agp_bind_memory);
/** Calls agp_unbind_memory() */
int drm_agp_unbind_memory(DRM_AGP_MEM * handle)
{
if (!handle)
2003-08-15 04:31:54 -06:00
return -EINVAL;
return agp_unbind_memory(handle);
}
2006-08-22 02:09:57 -06:00
/*
* AGP ttm backend interface.
*/
#ifndef AGP_USER_TYPES
#define AGP_USER_TYPES (1 << 16)
#define AGP_USER_MEMORY (AGP_USER_TYPES)
#define AGP_USER_CACHED_MEMORY (AGP_USER_TYPES + 1)
#endif
#define AGP_REQUIRED_MAJOR 0
#define AGP_REQUIRED_MINOR 102
2007-07-15 21:37:02 -06:00
static int drm_agp_needs_unbind_cache_adjust(struct drm_ttm_backend *backend) {
2006-10-12 08:10:47 -06:00
return ((backend->flags & DRM_BE_FLAG_BOUND_CACHED) ? 0 : 1);
2006-08-22 02:09:57 -06:00
}
2007-07-15 20:32:51 -06:00
static int drm_agp_populate(struct drm_ttm_backend *backend, unsigned long num_pages,
2006-08-22 02:09:57 -06:00
struct page **pages) {
2007-07-15 20:32:51 -06:00
struct drm_agp_ttm_backend *agp_be =
container_of(backend, struct drm_agp_ttm_backend, backend);
2006-08-22 02:09:57 -06:00
struct page **cur_page, **last_page = pages + num_pages;
DRM_AGP_MEM *mem;
if (drm_alloc_memctl(num_pages * sizeof(void *)))
return -1;
2006-08-22 02:09:57 -06:00
DRM_DEBUG("drm_agp_populate_ttm\n");
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,11)
mem = drm_agp_allocate_memory(num_pages, AGP_USER_MEMORY);
2006-08-22 02:09:57 -06:00
#else
mem = drm_agp_allocate_memory(agp_be->bridge, num_pages, AGP_USER_MEMORY);
2006-08-22 02:09:57 -06:00
#endif
if (!mem) {
drm_free_memctl(num_pages *sizeof(void *));
2006-08-22 02:09:57 -06:00
return -1;
}
2006-08-22 02:09:57 -06:00
DRM_DEBUG("Current page count is %ld\n", (long) mem->page_count);
mem->page_count = 0;
for (cur_page = pages; cur_page < last_page; ++cur_page) {
mem->memory[mem->page_count++] = phys_to_gart(page_to_phys(*cur_page));
2006-08-22 02:09:57 -06:00
}
agp_be->mem = mem;
2006-08-22 02:09:57 -06:00
return 0;
}
2007-07-15 21:37:02 -06:00
static int drm_agp_bind_ttm(struct drm_ttm_backend *backend,
unsigned long offset,
int cached)
{
2007-07-15 20:32:51 -06:00
struct drm_agp_ttm_backend *agp_be =
container_of(backend, struct drm_agp_ttm_backend, backend);
DRM_AGP_MEM *mem = agp_be->mem;
2006-08-22 02:09:57 -06:00
int ret;
DRM_DEBUG("drm_agp_bind_ttm\n");
mem->is_flushed = TRUE;
mem->type = (cached) ? AGP_USER_CACHED_MEMORY :
AGP_USER_MEMORY;
2006-08-22 02:09:57 -06:00
ret = drm_agp_bind_memory(mem, offset);
if (ret) {
DRM_ERROR("AGP Bind memory failed\n");
}
DRM_FLAG_MASKED(backend->flags, (cached) ? DRM_BE_FLAG_BOUND_CACHED : 0,
DRM_BE_FLAG_BOUND_CACHED);
2006-08-22 02:09:57 -06:00
return ret;
}
2007-07-15 21:37:02 -06:00
static int drm_agp_unbind_ttm(struct drm_ttm_backend *backend) {
2006-08-22 02:09:57 -06:00
2007-07-15 20:32:51 -06:00
struct drm_agp_ttm_backend *agp_be =
container_of(backend, struct drm_agp_ttm_backend, backend);
2006-08-22 02:09:57 -06:00
DRM_DEBUG("drm_agp_unbind_ttm\n");
if (agp_be->mem->is_bound)
return drm_agp_unbind_memory(agp_be->mem);
2006-08-22 02:09:57 -06:00
else
return 0;
}
2007-07-15 21:37:02 -06:00
static void drm_agp_clear_ttm(struct drm_ttm_backend *backend) {
2006-08-22 02:09:57 -06:00
2007-07-15 20:32:51 -06:00
struct drm_agp_ttm_backend *agp_be =
container_of(backend, struct drm_agp_ttm_backend, backend);
DRM_AGP_MEM *mem = agp_be->mem;
2006-08-22 02:09:57 -06:00
DRM_DEBUG("drm_agp_clear_ttm\n");
if (mem) {
unsigned long num_pages = mem->page_count;
backend->func->unbind(backend);
2006-08-22 02:09:57 -06:00
agp_free_memory(mem);
drm_free_memctl(num_pages *sizeof(void *));
2006-08-22 02:09:57 -06:00
}
agp_be->mem = NULL;
2006-08-22 02:09:57 -06:00
}
2007-07-15 21:37:02 -06:00
static void drm_agp_destroy_ttm(struct drm_ttm_backend *backend) {
2006-08-22 02:09:57 -06:00
2007-07-15 20:32:51 -06:00
struct drm_agp_ttm_backend *agp_be;
2006-08-22 02:09:57 -06:00
if (backend) {
DRM_DEBUG("drm_agp_destroy_ttm\n");
2007-07-15 20:32:51 -06:00
agp_be = container_of(backend, struct drm_agp_ttm_backend, backend);
if (agp_be) {
if (agp_be->mem) {
backend->func->clear(backend);
2006-08-22 02:09:57 -06:00
}
drm_ctl_free(agp_be, sizeof(*agp_be), DRM_MEM_TTM);
2006-08-22 02:09:57 -06:00
}
}
}
2007-07-15 21:37:02 -06:00
static struct drm_ttm_backend_func agp_ttm_backend =
{
.needs_ub_cache_adjust = drm_agp_needs_unbind_cache_adjust,
.populate = drm_agp_populate,
.clear = drm_agp_clear_ttm,
.bind = drm_agp_bind_ttm,
.unbind = drm_agp_unbind_ttm,
.destroy = drm_agp_destroy_ttm,
};
2007-07-15 20:32:51 -06:00
struct drm_ttm_backend *drm_agp_init_ttm(struct drm_device *dev)
{
2006-08-22 02:09:57 -06:00
2007-07-15 20:32:51 -06:00
struct drm_agp_ttm_backend *agp_be;
struct agp_kern_info *info;
if (!dev->agp) {
DRM_ERROR("AGP is not initialized.\n");
return NULL;
}
info = &dev->agp->agp_info;
if (info->version.major != AGP_REQUIRED_MAJOR ||
info->version.minor < AGP_REQUIRED_MINOR) {
DRM_ERROR("Wrong agpgart version %d.%d\n"
"\tYou need at least version %d.%d.\n",
info->version.major,
info->version.minor,
AGP_REQUIRED_MAJOR,
AGP_REQUIRED_MINOR);
return NULL;
}
2006-08-22 02:09:57 -06:00
agp_be = drm_ctl_calloc(1, sizeof(*agp_be), DRM_MEM_TTM);
2006-08-22 02:09:57 -06:00
if (!agp_be)
return NULL;
agp_be->mem = NULL;
agp_be->bridge = dev->agp->bridge;
agp_be->populated = FALSE;
agp_be->backend.func = &agp_ttm_backend;
agp_be->backend.mem_type = DRM_BO_MEM_TT;
return &agp_be->backend;
2006-08-22 02:09:57 -06:00
}
EXPORT_SYMBOL(drm_agp_init_ttm);
2006-08-22 02:09:57 -06:00
#endif /* __OS_HAS_AGP */