core needs three new files
authorJon Smirl <jonsmirl@yahoo.com>
Mon, 27 Sep 2004 19:53:51 +0000 (19:53 +0000)
committerJon Smirl <jonsmirl@yahoo.com>
Mon, 27 Sep 2004 19:53:51 +0000 (19:53 +0000)
linux-core/drm_memory.c [new file with mode: 0644]
linux-core/drm_memory_debug.c [new file with mode: 0644]
linux-core/drm_module.h [new file with mode: 0644]

diff --git a/linux-core/drm_memory.c b/linux-core/drm_memory.c
new file mode 100644 (file)
index 0000000..81f109b
--- /dev/null
@@ -0,0 +1,181 @@
+/** 
+ * \file drm_memory.h 
+ * Memory management wrappers for DRM
+ *
+ * \author Rickard E. (Rik) Faith <faith@valinux.com>
+ * \author Gareth Hughes <gareth@valinux.com>
+ */
+
+/* 
+ * Created: Thu Feb  4 14:00:34 1999 by faith@valinux.com
+ *
+ * Copyright 1999 Precision Insight, Inc., Cedar Park, Texas.
+ * Copyright 2000 VA Linux Systems, Inc., Sunnyvale, California.
+ * 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:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * 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
+ * VA LINUX SYSTEMS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/config.h>
+#include <linux/highmem.h>
+#include "drmP.h"
+
+#ifdef DEBUG_MEMORY
+#include "drm_memory_debug.h"
+#else
+
+/** No-op. */
+void drm_mem_init(void)
+{
+}
+
+/**
+ * Called when "/proc/dri/%dev%/mem" is read.
+ * 
+ * \param buf output buffer.
+ * \param start start of output data.
+ * \param offset requested start offset.
+ * \param len requested number of bytes.
+ * \param eof whether there is no more data to return.
+ * \param data private data.
+ * \return number of written bytes.
+ *
+ * No-op. 
+ */
+int drm_mem_info(char *buf, char **start, off_t offset,
+                 int len, int *eof, void *data)
+{
+       return 0;
+}
+
+/** Wrapper around kmalloc() */
+void *drm_calloc(size_t nmemb, size_t size, int area)
+{
+       void *addr;
+
+       addr = kmalloc(size * nmemb, GFP_KERNEL);
+       if (addr != NULL)
+               memset((void *)addr, 0, size * nmemb);
+
+       return addr;
+}
+EXPORT_SYMBOL(drm_calloc);
+
+/** Wrapper around kmalloc() and kfree() */
+void *drm_realloc(void *oldpt, size_t oldsize, size_t size, int area)
+{
+       void *pt;
+
+       if (!(pt = kmalloc(size, GFP_KERNEL))) return NULL;
+       if (oldpt && oldsize) {
+               memcpy(pt, oldpt, oldsize);
+               kfree(oldpt);
+       }
+       return pt;
+}
+
+/**
+ * Allocate pages.
+ *
+ * \param order size order.
+ * \param area memory area. (Not used.)
+ * \return page address on success, or zero on failure.
+ *
+ * Allocate and reserve free pages.
+ */
+unsigned long drm_alloc_pages(int order, int area)
+{
+       unsigned long address;
+       unsigned long bytes       = PAGE_SIZE << order;
+       unsigned long addr;
+       unsigned int  sz;
+
+       address = __get_free_pages(GFP_KERNEL, order);
+       if (!address) 
+               return 0;
+
+                               /* Zero */
+       memset((void *)address, 0, bytes);
+
+                               /* Reserve */
+       for (addr = address, sz = bytes;
+            sz > 0;
+            addr += PAGE_SIZE, sz -= PAGE_SIZE) {
+               SetPageReserved(virt_to_page(addr));
+       }
+
+       return address;
+}
+
+/**
+ * Free pages.
+ * 
+ * \param address address of the pages to free.
+ * \param order size order.
+ * \param area memory area. (Not used.)
+ *
+ * Unreserve and free pages allocated by alloc_pages().
+ */
+void drm_free_pages(unsigned long address, int order, int area)
+{
+       unsigned long bytes = PAGE_SIZE << order;
+       unsigned long addr;
+       unsigned int  sz;
+
+       if (!address) 
+               return;
+
+       /* Unreserve */
+       for (addr = address, sz = bytes;
+            sz > 0;
+            addr += PAGE_SIZE, sz -= PAGE_SIZE) {
+               ClearPageReserved(virt_to_page(addr));
+       }
+
+       free_pages(address, order);
+}
+
+
+#if __OS_HAS_AGP
+/** Wrapper around agp_allocate_memory() */
+DRM_AGP_MEM *drm_alloc_agp(int pages, u32 type)
+{
+       return drm_agp_allocate_memory(pages, type);
+}
+
+/** Wrapper around agp_free_memory() */
+int drm_free_agp(DRM_AGP_MEM *handle, int pages)
+{
+       return drm_agp_free_memory(handle) ? 0 : -EINVAL;
+}
+
+/** Wrapper around agp_bind_memory() */
+int drm_bind_agp(DRM_AGP_MEM *handle, unsigned int start)
+{
+       return drm_agp_bind_memory(handle, start);
+}
+
+/** Wrapper around agp_unbind_memory() */
+int drm_unbind_agp(DRM_AGP_MEM *handle)
+{
+       return drm_agp_unbind_memory(handle);
+}
+#endif /* agp */
+#endif /* debug_memory */
diff --git a/linux-core/drm_memory_debug.c b/linux-core/drm_memory_debug.c
new file mode 100644 (file)
index 0000000..2c3564a
--- /dev/null
@@ -0,0 +1,459 @@
+/**
+ * \file drm_memory_debug.h 
+ * Memory management wrappers for DRM.
+ *
+ * \author Rickard E. (Rik) Faith <faith@valinux.com>
+ * \author Gareth Hughes <gareth@valinux.com>
+ */
+
+/*
+ * Copyright 1999 Precision Insight, Inc., Cedar Park, Texas.
+ * Copyright 2000 VA Linux Systems, Inc., Sunnyvale, California.
+ * 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:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * 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
+ * VA LINUX SYSTEMS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/config.h>
+#include "drmP.h"
+
+typedef struct drm_mem_stats {
+       const char        *name;
+       int               succeed_count;
+       int               free_count;
+       int               fail_count;
+       unsigned long     bytes_allocated;
+       unsigned long     bytes_freed;
+} drm_mem_stats_t;
+
+static spinlock_t        drm_mem_lock       = SPIN_LOCK_UNLOCKED;
+static unsigned long     drm_ram_available = 0; /* In pages */
+static unsigned long     drm_ram_used      = 0;
+static drm_mem_stats_t   drm_mem_stats[]   = {
+       [DRM_MEM_DMA]       = { "dmabufs"  },
+       [DRM_MEM_SAREA]     = { "sareas"   },
+       [DRM_MEM_DRIVER]    = { "driver"   },
+       [DRM_MEM_MAGIC]     = { "magic"    },
+       [DRM_MEM_IOCTLS]    = { "ioctltab" },
+       [DRM_MEM_MAPS]      = { "maplist"  },
+       [DRM_MEM_VMAS]      = { "vmalist"  },
+       [DRM_MEM_BUFS]      = { "buflist"  },
+       [DRM_MEM_SEGS]      = { "seglist"  },
+       [DRM_MEM_PAGES]     = { "pagelist" },
+       [DRM_MEM_FILES]     = { "files"    },
+       [DRM_MEM_QUEUES]    = { "queues"   },
+       [DRM_MEM_CMDS]      = { "commands" },
+       [DRM_MEM_MAPPINGS]  = { "mappings" },
+       [DRM_MEM_BUFLISTS]  = { "buflists" },
+       [DRM_MEM_AGPLISTS]  = { "agplist"  },
+       [DRM_MEM_SGLISTS]   = { "sglist"   },
+       [DRM_MEM_TOTALAGP]  = { "totalagp" },
+       [DRM_MEM_BOUNDAGP]  = { "boundagp" },
+       [DRM_MEM_CTXBITMAP] = { "ctxbitmap"},
+       [DRM_MEM_CTXLIST]   = { "ctxlist"  },
+       [DRM_MEM_STUB]      = { "stub"     },
+       { NULL, 0, }            /* Last entry must be null */
+};
+
+void drm_mem_init(void)
+{
+       drm_mem_stats_t *mem;
+       struct sysinfo  si;
+
+       for (mem = drm_mem_stats; mem->name; ++mem) {
+               mem->succeed_count   = 0;
+               mem->free_count      = 0;
+               mem->fail_count      = 0;
+               mem->bytes_allocated = 0;
+               mem->bytes_freed     = 0;
+       }
+
+       si_meminfo(&si);
+       drm_ram_available = si.totalram;
+       drm_ram_used       = 0;
+}
+
+/* drm_mem_info is called whenever a process reads /dev/drm/mem. */
+
+static int drm__mem_info(char *buf, char **start, off_t offset,
+                         int request, int *eof, void *data)
+{
+       drm_mem_stats_t *pt;
+       int             len = 0;
+
+       if (offset > DRM_PROC_LIMIT) {
+               *eof = 1;
+               return 0;
+       }
+
+       *eof   = 0;
+       *start = &buf[offset];
+
+       DRM_PROC_PRINT("                  total counts                  "
+                      " |    outstanding  \n");
+       DRM_PROC_PRINT("type       alloc freed fail     bytes      freed"
+                      " | allocs      bytes\n\n");
+       DRM_PROC_PRINT("%-9.9s %5d %5d %4d %10lu kB         |\n",
+                      "system", 0, 0, 0,
+                      drm_ram_available << (PAGE_SHIFT - 10));
+       DRM_PROC_PRINT("%-9.9s %5d %5d %4d %10lu kB         |\n",
+                      "locked", 0, 0, 0, drm_ram_used >> 10);
+       DRM_PROC_PRINT("\n");
+       for (pt = drm_mem_stats; pt->name; pt++) {
+               DRM_PROC_PRINT("%-9.9s %5d %5d %4d %10lu %10lu | %6d %10ld\n",
+                              pt->name,
+                              pt->succeed_count,
+                              pt->free_count,
+                              pt->fail_count,
+                              pt->bytes_allocated,
+                              pt->bytes_freed,
+                              pt->succeed_count - pt->free_count,
+                              (long)pt->bytes_allocated
+                              - (long)pt->bytes_freed);
+       }
+
+       if (len > request + offset) return request;
+       *eof = 1;
+       return len - offset;
+}
+
+int drm_mem_info(char *buf, char **start, off_t offset,
+                 int len, int *eof, void *data)
+{
+       int ret;
+
+       spin_lock(&drm_mem_lock);
+       ret = drm__mem_info(buf, start, offset, len, eof, data);
+       spin_unlock(&drm_mem_lock);
+       return ret;
+}
+
+void *drm_alloc(size_t size, int area)
+{
+       void *pt;
+
+       if (!size) {
+               DRM_MEM_ERROR(area, "Allocating 0 bytes\n");
+               return NULL;
+       }
+
+       if (!(pt = kmalloc(size, GFP_KERNEL))) {
+               spin_lock(&drm_mem_lock);
+               ++drm_mem_stats[area].fail_count;
+               spin_unlock(&drm_mem_lock);
+               return NULL;
+       }
+       spin_lock(&drm_mem_lock);
+       ++drm_mem_stats[area].succeed_count;
+       drm_mem_stats[area].bytes_allocated += size;
+       spin_unlock(&drm_mem_lock);
+       return pt;
+}
+
+void *drm_calloc(size_t nmemb, size_t size, int area)
+{
+       void *addr;
+
+       addr = drm_alloc(nmemb * size, area);
+       if (addr != NULL)
+               memset((void *)addr, 0, size * nmemb);
+
+       return addr;
+}
+
+void *drm_realloc(void *oldpt, size_t oldsize, size_t size, int area)
+{
+       void *pt;
+
+       if (!(pt = drm_alloc(size, area))) return NULL;
+       if (oldpt && oldsize) {
+               memcpy(pt, oldpt, oldsize);
+               drm_free(oldpt, oldsize, area);
+       }
+       return pt;
+}
+
+void drm_free(void *pt, size_t size, int area)
+{
+       int alloc_count;
+       int free_count;
+
+       if (!pt) DRM_MEM_ERROR(area, "Attempt to free NULL pointer\n");
+       else     kfree(pt);
+       spin_lock(&drm_mem_lock);
+       drm_mem_stats[area].bytes_freed += size;
+       free_count  = ++drm_mem_stats[area].free_count;
+       alloc_count =   drm_mem_stats[area].succeed_count;
+       spin_unlock(&drm_mem_lock);
+       if (free_count > alloc_count) {
+               DRM_MEM_ERROR(area, "Excess frees: %d frees, %d allocs\n",
+                             free_count, alloc_count);
+       }
+}
+
+unsigned long drm_alloc_pages(int order, int area)
+{
+       unsigned long address;
+       unsigned long bytes       = PAGE_SIZE << order;
+       unsigned long addr;
+       unsigned int  sz;
+
+       spin_lock(&drm_mem_lock);
+       if ((drm_ram_used >> PAGE_SHIFT)
+           > (DRM_RAM_PERCENT * drm_ram_available) / 100) {
+               spin_unlock(&drm_mem_lock);
+               return 0;
+       }
+       spin_unlock(&drm_mem_lock);
+
+       address = __get_free_pages(GFP_KERNEL, order);
+       if (!address) {
+               spin_lock(&drm_mem_lock);
+               ++drm_mem_stats[area].fail_count;
+               spin_unlock(&drm_mem_lock);
+               return 0;
+       }
+       spin_lock(&drm_mem_lock);
+       ++drm_mem_stats[area].succeed_count;
+       drm_mem_stats[area].bytes_allocated += bytes;
+       drm_ram_used                         += bytes;
+       spin_unlock(&drm_mem_lock);
+
+
+                               /* Zero outside the lock */
+       memset((void *)address, 0, bytes);
+
+                               /* Reserve */
+       for (addr = address, sz = bytes;
+            sz > 0;
+            addr += PAGE_SIZE, sz -= PAGE_SIZE) {
+               SetPageReserved(virt_to_page(addr));
+       }
+
+       return address;
+}
+
+void drm_free_pages(unsigned long address, int order, int area)
+{
+       unsigned long bytes = PAGE_SIZE << order;
+       int               alloc_count;
+       int               free_count;
+       unsigned long addr;
+       unsigned int  sz;
+
+       if (!address) {
+               DRM_MEM_ERROR(area, "Attempt to free address 0\n");
+       } else {
+                               /* Unreserve */
+               for (addr = address, sz = bytes;
+                    sz > 0;
+                    addr += PAGE_SIZE, sz -= PAGE_SIZE) {
+                       ClearPageReserved(virt_to_page(addr));
+               }
+               free_pages(address, order);
+       }
+
+       spin_lock(&drm_mem_lock);
+       free_count  = ++drm_mem_stats[area].free_count;
+       alloc_count =   drm_mem_stats[area].succeed_count;
+       drm_mem_stats[area].bytes_freed += bytes;
+       drm_ram_used                     -= bytes;
+       spin_unlock(&drm_mem_lock);
+       if (free_count > alloc_count) {
+               DRM_MEM_ERROR(area,
+                             "Excess frees: %d frees, %d allocs\n",
+                             free_count, alloc_count);
+       }
+}
+
+void *drm_ioremap(unsigned long offset, unsigned long size, drm_device_t *dev)
+{
+       void *pt;
+
+       if (!size) {
+               DRM_MEM_ERROR(DRM_MEM_MAPPINGS,
+                             "Mapping 0 bytes at 0x%08lx\n", offset);
+               return NULL;
+       }
+
+       if (!(pt = drm_ioremap(offset, size, dev))) {
+               spin_lock(&drm_mem_lock);
+               ++drm_mem_stats[DRM_MEM_MAPPINGS].fail_count;
+               spin_unlock(&drm_mem_lock);
+               return NULL;
+       }
+       spin_lock(&drm_mem_lock);
+       ++drm_mem_stats[DRM_MEM_MAPPINGS].succeed_count;
+       drm_mem_stats[DRM_MEM_MAPPINGS].bytes_allocated += size;
+       spin_unlock(&drm_mem_lock);
+       return pt;
+}
+
+void *drm_ioremap_nocache(unsigned long offset, unsigned long size, drm_device_t *dev)
+{
+       void *pt;
+
+       if (!size) {
+               DRM_MEM_ERROR(DRM_MEM_MAPPINGS,
+                             "Mapping 0 bytes at 0x%08lx\n", offset);
+               return NULL;
+       }
+
+       if (!(pt = drm_ioremap_nocache(offset, size, dev))) {
+               spin_lock(&drm_mem_lock);
+               ++drm_mem_stats[DRM_MEM_MAPPINGS].fail_count;
+               spin_unlock(&drm_mem_lock);
+               return NULL;
+       }
+       spin_lock(&drm_mem_lock);
+       ++drm_mem_stats[DRM_MEM_MAPPINGS].succeed_count;
+       drm_mem_stats[DRM_MEM_MAPPINGS].bytes_allocated += size;
+       spin_unlock(&drm_mem_lock);
+       return pt;
+}
+
+void drm_ioremapfree(void *pt, unsigned long size, drm_device_t *dev)
+{
+       int alloc_count;
+       int free_count;
+
+       if (!pt)
+               DRM_MEM_ERROR(DRM_MEM_MAPPINGS,
+                             "Attempt to free NULL pointer\n");
+       else
+               drm_ioremapfree(pt, size, dev);
+
+       spin_lock(&drm_mem_lock);
+       drm_mem_stats[DRM_MEM_MAPPINGS].bytes_freed += size;
+       free_count  = ++drm_mem_stats[DRM_MEM_MAPPINGS].free_count;
+       alloc_count =   drm_mem_stats[DRM_MEM_MAPPINGS].succeed_count;
+       spin_unlock(&drm_mem_lock);
+       if (free_count > alloc_count) {
+               DRM_MEM_ERROR(DRM_MEM_MAPPINGS,
+                             "Excess frees: %d frees, %d allocs\n",
+                             free_count, alloc_count);
+       }
+}
+
+#if __OS_HAS_AGP
+
+DRM_AGP_MEM *drm_alloc_agp(int pages, u32 type)
+{
+       DRM_AGP_MEM *handle;
+
+       if (!pages) {
+               DRM_MEM_ERROR(DRM_MEM_TOTALAGP, "Allocating 0 pages\n");
+               return NULL;
+       }
+
+       if ((handle = drm_agp_allocate_memory(pages, type))) {
+               spin_lock(&drm_mem_lock);
+               ++drm_mem_stats[DRM_MEM_TOTALAGP].succeed_count;
+               drm_mem_stats[DRM_MEM_TOTALAGP].bytes_allocated
+                       += pages << PAGE_SHIFT;
+               spin_unlock(&drm_mem_lock);
+               return handle;
+       }
+       spin_lock(&drm_mem_lock);
+       ++drm_mem_stats[DRM_MEM_TOTALAGP].fail_count;
+       spin_unlock(&drm_mem_lock);
+       return NULL;
+}
+
+int drm_free_agp(DRM_AGP_MEM *handle, int pages)
+{
+       int           alloc_count;
+       int           free_count;
+       int           retval = -EINVAL;
+
+       if (!handle) {
+               DRM_MEM_ERROR(DRM_MEM_TOTALAGP,
+                             "Attempt to free NULL AGP handle\n");
+               return retval;
+       }
+
+       if (drm_agp_free_memory(handle)) {
+               spin_lock(&drm_mem_lock);
+               free_count  = ++drm_mem_stats[DRM_MEM_TOTALAGP].free_count;
+               alloc_count =   drm_mem_stats[DRM_MEM_TOTALAGP].succeed_count;
+               drm_mem_stats[DRM_MEM_TOTALAGP].bytes_freed
+                       += pages << PAGE_SHIFT;
+               spin_unlock(&drm_mem_lock);
+               if (free_count > alloc_count) {
+                       DRM_MEM_ERROR(DRM_MEM_TOTALAGP,
+                                     "Excess frees: %d frees, %d allocs\n",
+                                     free_count, alloc_count);
+               }
+               return 0;
+       }
+       return retval;
+}
+
+int drm_bind_agp(DRM_AGP_MEM *handle, unsigned int start)
+{
+       int retcode = -EINVAL;
+
+       if (!handle) {
+               DRM_MEM_ERROR(DRM_MEM_BOUNDAGP,
+                             "Attempt to bind NULL AGP handle\n");
+               return retcode;
+       }
+
+       if (!(retcode = drm_agp_bind_memory(handle, start))) {
+               spin_lock(&drm_mem_lock);
+               ++drm_mem_stats[DRM_MEM_BOUNDAGP].succeed_count;
+               drm_mem_stats[DRM_MEM_BOUNDAGP].bytes_allocated
+                       += handle->page_count << PAGE_SHIFT;
+               spin_unlock(&drm_mem_lock);
+               return retcode;
+       }
+       spin_lock(&drm_mem_lock);
+       ++drm_mem_stats[DRM_MEM_BOUNDAGP].fail_count;
+       spin_unlock(&drm_mem_lock);
+       return retcode;
+}
+
+int drm_unbind_agp(DRM_AGP_MEM *handle)
+{
+       int alloc_count;
+       int free_count;
+       int retcode = -EINVAL;
+
+       if (!handle) {
+               DRM_MEM_ERROR(DRM_MEM_BOUNDAGP,
+                             "Attempt to unbind NULL AGP handle\n");
+               return retcode;
+       }
+
+       if ((retcode = drm_agp_unbind_memory(handle))) return retcode;
+       spin_lock(&drm_mem_lock);
+       free_count  = ++drm_mem_stats[DRM_MEM_BOUNDAGP].free_count;
+       alloc_count = drm_mem_stats[DRM_MEM_BOUNDAGP].succeed_count;
+       drm_mem_stats[DRM_MEM_BOUNDAGP].bytes_freed
+               += handle->page_count << PAGE_SHIFT;
+       spin_unlock(&drm_mem_lock);
+       if (free_count > alloc_count) {
+               DRM_MEM_ERROR(DRM_MEM_BOUNDAGP,
+                             "Excess frees: %d frees, %d allocs\n",
+                             free_count, alloc_count);
+       }
+       return retcode;
+}
+#endif
diff --git a/linux-core/drm_module.h b/linux-core/drm_module.h
new file mode 100644 (file)
index 0000000..84a4c43
--- /dev/null
@@ -0,0 +1,67 @@
+/*
+ * Copyright 2004 Jon Smirl <jonsmirl@gmail.com>
+ * Copyright 1998-2003 VIA Technologies, Inc. All Rights Reserved.
+ * Copyright 2001-2003 S3 Graphics, Inc. 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, sub license,
+ * 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:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * 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 NON-INFRINGEMENT. IN NO EVENT SHALL
+ * VIA, S3 GRAPHICS, AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#include "drm_pciids.h"
+
+int DRM(postinit)( struct drm_device *dev, unsigned long flags )
+{
+       DRM_INFO( "Initialized %s %d.%d.%d %s on minor %d: %s\n",
+               DRIVER_NAME,
+               DRIVER_MAJOR,
+               DRIVER_MINOR,
+               DRIVER_PATCHLEVEL,
+               DRIVER_DATE,
+               dev->minor,
+               pci_pretty_name(pdev)
+               );
+       return 0;
+}
+
+
+static struct pci_device_id DRM(pciidlist)[] = {
+       DRM(PCI_IDS)
+};
+
+static struct pci_driver DRM(driver) = {
+       .name          = DRIVER_NAME,
+       .id_table      = DRM(pciidlist),
+       .probe         = drm_probe,
+       .remove        = __devexit_p(drm_cleanup_pci),
+};
+
+static int __init DRM(init)(void)
+{
+       return drm_init(&DRM(driver), DRM(pciidlist));
+}
+
+static void __exit DRM(exit)(void)
+{
+       drm_exit(&DRM(driver));
+}
+
+#define drm_module_init(x) module_init(x)
+#define drm_module_exit(x) module_exit(x)
+drm_module_init(DRM(init));
+drm_module_exit(DRM(exit));