linux/drivers/gpu/drm/i915/i915_gem_stolen.c

722 lines
19 KiB
C
Raw Normal View History

/*
* Copyright © 2008-2012 Intel Corporation
*
* 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
* THE AUTHORS OR COPYRIGHT HOLDERS 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.
*
* Authors:
* Eric Anholt <eric@anholt.net>
* Chris Wilson <chris@chris-wilson.co.uk>
*
*/
#include <drm/i915_drm.h>
#include "i915_drv.h"
/*
* The BIOS typically reserves some of the system's memory for the exclusive
* use of the integrated graphics. This memory is no longer available for
* use by the OS and so the user finds that his system has less memory
* available than he put in. We refer to this memory as stolen.
*
* The BIOS will allocate its framebuffer from the stolen memory. Our
* goal is try to reuse that object for our own fbcon which must always
* be available for panics. Anything else we can reuse the stolen memory
* for is a boon.
*/
int i915_gem_stolen_insert_node_in_range(struct drm_i915_private *dev_priv,
struct drm_mm_node *node, u64 size,
unsigned alignment, u64 start, u64 end)
{
int ret;
if (!drm_mm_initialized(&dev_priv->mm.stolen))
return -ENODEV;
drm/i915: Do NOT skip the first 4k of stolen memory for pre-allocated buffers v2 Before this commit the WaSkipStolenMemoryFirstPage workaround code was skipping the first 4k by passing 4096 as start of the address range passed to drm_mm_init(). This means that calling drm_mm_reserve_node() to try and reserve the firmware framebuffer so that we can inherit it would always fail, as the firmware framebuffer starts at address 0. Commit d43537610470 ("drm/i915: skip the first 4k of stolen memory on everything >= gen8") says in its commit message: "This is confirmed to fix Skylake screen flickering issues (probably caused by the fact that we initialized a ring in the first page of stolen, but I didn't 100% confirm this theory)." Which suggests that it is safe to use the first page for a linear framebuffer as the firmware is doing (see note below). This commit always passes 0 as start to drm_mm_init() and works around WaSkipStolenMemoryFirstPage in i915_gem_stolen_insert_node_in_range() by insuring the start address passed by to drm_mm_insert_node_in_range() is always 4k or more. All entry points to i915_gem_stolen.c go through i915_gem_stolen_insert_node_in_range(), so that any newly allocated objects such as ring-buffers will not be allocated in the first 4k. The one exception is i915_gem_object_create_stolen_for_preallocated() which directly calls drm_mm_reserve_node() which now will be able to use the first 4k. This fixes the i915 driver no longer being able to inherit the firmware framebuffer on gen8+, which fixes the video output changing from the vendor logo to a black screen as soon as the i915 driver is loaded (on systems without fbcon). Some notes about the mapping of the BIOS framebuffer: v1 led to some discussion if the assumption of the intel_display.c code that the firmware framebuffer is a linear mapping of the stolen memory starting at offset 0 is still correct, because that would mean that the GOP does not implement the WaSkipStolenMemoryFirstPage workaround. To verify this the following code was added at the end of i915_gem_object_create_stolen_for_preallocated() : pr_err("first ggtt entry before bind: 0x%016llx\n", readq(dev_priv->ggtt.gsm)); ret = i915_vma_bind(vma, HAS_LLC(dev_priv) ? I915_CACHE_LLC : I915_CACHE_NONE, PIN_UPDATE); pr_err("i915_vma_bind ret %d\n", ret); pr_err("first ggtt entry after bind: 0x%016llx\n", readq(dev_priv->ggtt.gsm)); Which prints the mapping of the first page, then does a vma_bind() to force update the mapping with our linear view of the framebuffer and then prints the mapping of the first page again. On an Asrock B150M Pro4S/D3 mainboard with i5-6500 CPU this prints: [ 1.651141] first ggtt entry before bind: 0x0000000078c00001 [ 1.651151] i915_vma_bind ret 0 [ 1.651152] first ggtt entry after bind: 0x0000000078c00083 And "sudo cat /proc/iomem | grep Stolen" gives: 78c00000-88bfffff : Graphics Stolen Memory There are no visual changes with this patch (BIOS vendor logo still stays in place when we inherit the BIOS framebuffer), so the vma_bind() does not impact which memory is being scanned out. The address of the first ggtt entry matches with the start of stolen and the i915_vma_bind call only changes the first gtt entry's flags, or-ing in _PAGE_RW (BIT(1)) and PPAT_CACHED (BIT(7)), which perfectly matches what we would expect based on gen8_pte_encode()'s behavior. So it seems that the GOP indeed does NOT implement the wa and the i915's code assuming a linear mapping at the start of stolen for the BIOS fb still holds true for gen8+. I've also tested this on a Cherry Trail based device (a GPD Win) with identical results (the flags are 0x1b after the vma_bind on CHT, which matches with I915_CACHE_NONE). Changed in v2: No code changes, extended the commit message with the verification that the intel_display.c BIOS framebuffer mapping is still correct. Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180420095933.16442-1-hdegoede@redhat.com
2018-04-20 17:59:33 +08:00
/* WaSkipStolenMemoryFirstPage:bdw+ */
if (INTEL_GEN(dev_priv) >= 8 && start < 4096)
start = 4096;
mutex_lock(&dev_priv->mm.stolen_lock);
drm: Improve drm_mm search (and fix topdown allocation) with rbtrees The drm_mm range manager claimed to support top-down insertion, but it was neither searching for the top-most hole that could fit the allocation request nor fitting the request to the hole correctly. In order to search the range efficiently, we create a secondary index for the holes using either their size or their address. This index allows us to find the smallest hole or the hole at the bottom or top of the range efficiently, whilst keeping the hole stack to rapidly service evictions. v2: Search for holes both high and low. Rename flags to mode. v3: Discover rb_entry_safe() and use it! v4: Kerneldoc for enum drm_mm_insert_mode. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: "Christian König" <christian.koenig@amd.com> Cc: David Airlie <airlied@linux.ie> Cc: Russell King <rmk+kernel@armlinux.org.uk> Cc: Daniel Vetter <daniel.vetter@intel.com> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Sean Paul <seanpaul@chromium.org> Cc: Lucas Stach <l.stach@pengutronix.de> Cc: Christian Gmeiner <christian.gmeiner@gmail.com> Cc: Rob Clark <robdclark@gmail.com> Cc: Thierry Reding <thierry.reding@gmail.com> Cc: Stephen Warren <swarren@wwwdotorg.org> Cc: Alexandre Courbot <gnurou@gmail.com> Cc: Eric Anholt <eric@anholt.net> Cc: Sinclair Yeh <syeh@vmware.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Sinclair Yeh <syeh@vmware.com> # vmwgfx Reviewed-by: Lucas Stach <l.stach@pengutronix.de> #etnaviv Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20170202210438.28702-1-chris@chris-wilson.co.uk
2017-02-03 05:04:38 +08:00
ret = drm_mm_insert_node_in_range(&dev_priv->mm.stolen, node,
size, alignment, 0,
start, end, DRM_MM_INSERT_BEST);
mutex_unlock(&dev_priv->mm.stolen_lock);
return ret;
}
int i915_gem_stolen_insert_node(struct drm_i915_private *dev_priv,
struct drm_mm_node *node, u64 size,
unsigned alignment)
{
return i915_gem_stolen_insert_node_in_range(dev_priv, node, size,
2016-12-15 21:23:55 +08:00
alignment, 0, U64_MAX);
}
void i915_gem_stolen_remove_node(struct drm_i915_private *dev_priv,
struct drm_mm_node *node)
{
mutex_lock(&dev_priv->mm.stolen_lock);
drm_mm_remove_node(node);
mutex_unlock(&dev_priv->mm.stolen_lock);
}
static int i915_adjust_stolen(struct drm_i915_private *dev_priv,
struct resource *dsm)
{
struct i915_ggtt *ggtt = &dev_priv->ggtt;
struct resource *r;
drm/i915: Determine the stolen memory base address on gen2 There isn't an explicit stolen memory base register on gen2. Some old comment in the i915 code suggests we should get it via max_low_pfn_mapped, but that's clearly a bad idea on my MGM. The e820 map in said machine looks like this: [ 0.000000] BIOS-e820: [mem 0x0000000000000000-0x000000000009f7ff] usable [ 0.000000] BIOS-e820: [mem 0x000000000009f800-0x000000000009ffff] reserved [ 0.000000] BIOS-e820: [mem 0x00000000000ce000-0x00000000000cffff] reserved [ 0.000000] BIOS-e820: [mem 0x00000000000dc000-0x00000000000fffff] reserved [ 0.000000] BIOS-e820: [mem 0x0000000000100000-0x000000001f6effff] usable [ 0.000000] BIOS-e820: [mem 0x000000001f6f0000-0x000000001f6f7fff] ACPI data [ 0.000000] BIOS-e820: [mem 0x000000001f6f8000-0x000000001f6fffff] ACPI NVS [ 0.000000] BIOS-e820: [mem 0x000000001f700000-0x000000001fffffff] reserved [ 0.000000] BIOS-e820: [mem 0x00000000fec10000-0x00000000fec1ffff] reserved [ 0.000000] BIOS-e820: [mem 0x00000000ffb00000-0x00000000ffbfffff] reserved [ 0.000000] BIOS-e820: [mem 0x00000000fff00000-0x00000000ffffffff] reserved That makes max_low_pfn_mapped = 1f6f0000, so assuming our stolen memory would start there would place it on top of some ACPI memory regions. So not a good idea as already stated. The 9MB region after the ACPI regions at 0x1f700000 however looks promising given that the macine reports the stolen memory size to be 8MB. Looking at the PGTBL_CTL register, the GTT entries are at offset 0x1fee00000, and given that the GTT entries occupy 128KB, it looks like the stolen memory could start at 0x1f700000 and the GTT entries would occupy the last 128KB of the stolen memory. After some more digging through chipset documentation, I've determined the BIOS first allocates space for something called TSEG (something to do with SMM) from the top of memory, and then it allocates the graphics stolen memory below that. Accordind to the chipset documentation TSEG has a fixed size of 1MB on 855. So that explains the top 1MB in the e820 region. And it also confirms that the GTT entries are in fact at the end of the the stolen memory region. Derive the stolen memory base address on gen2 the same as the BIOS does (TOM-TSEG_SIZE-stolen_size). There are a few differences between the registers on various gen2 chipsets, so a few different codepaths are required. 865G is again bit more special since it seems to support enough memory to hit 4GB address space issues. This means the PCI allocations will also affect the location of the stolen memory. Fortunately there appears to be the TOUD register which may give us the correct answer directly. But the chipset docs are a bit unclear, so I'm not 100% sure that the graphics stolen memory is always the last thing the BIOS steals. Someone would need to verify it on a real system. I tested this on the my 830 and 855 machines, and so far everything looks peachy. v2: Rewrite to use the TOM-TSEG_SIZE-stolen_size and TOUD methods v3: Fix TSEG size for 830 v4: Add missing 'else' (Chris) Tested-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-10-08 17:08:20 +08:00
if (dsm->start == 0 || dsm->end <= dsm->start)
return -EINVAL;
/*
* TODO: We have yet too encounter the case where the GTT wasn't at the
* end of stolen. With that assumption we could simplify this.
*/
/* Make sure we don't clobber the GTT if it's within stolen memory */
if (INTEL_GEN(dev_priv) <= 4 &&
!IS_G33(dev_priv) && !IS_PINEVIEW(dev_priv) && !IS_G4X(dev_priv)) {
struct resource stolen[2] = {*dsm, *dsm};
struct resource ggtt_res;
resource_size_t ggtt_start;
ggtt_start = I915_READ(PGTBL_CTL);
if (IS_GEN(dev_priv, 4))
ggtt_start = (ggtt_start & PGTBL_ADDRESS_LO_MASK) |
(ggtt_start & PGTBL_ADDRESS_HI_MASK) << 28;
else
ggtt_start &= PGTBL_ADDRESS_LO_MASK;
ggtt_res =
(struct resource) DEFINE_RES_MEM(ggtt_start,
ggtt_total_entries(ggtt) * 4);
if (ggtt_res.start >= stolen[0].start && ggtt_res.start < stolen[0].end)
stolen[0].end = ggtt_res.start;
if (ggtt_res.end > stolen[1].start && ggtt_res.end <= stolen[1].end)
stolen[1].start = ggtt_res.end;
/* Pick the larger of the two chunks */
if (resource_size(&stolen[0]) > resource_size(&stolen[1]))
*dsm = stolen[0];
else
*dsm = stolen[1];
if (stolen[0].start != stolen[1].start ||
stolen[0].end != stolen[1].end) {
DRM_DEBUG_DRIVER("GTT within stolen memory at %pR\n", &ggtt_res);
DRM_DEBUG_DRIVER("Stolen memory adjusted to %pR\n", dsm);
}
}
/*
* Verify that nothing else uses this physical address. Stolen
* memory should be reserved by the BIOS and hidden from the
* kernel. So if the region is already marked as busy, something
* is seriously wrong.
*/
r = devm_request_mem_region(dev_priv->drm.dev, dsm->start,
resource_size(dsm),
"Graphics Stolen Memory");
if (r == NULL) {
/*
* One more attempt but this time requesting region from
* start + 1, as we have seen that this resolves the region
* conflict with the PCI Bus.
* This is a BIOS w/a: Some BIOS wrap stolen in the root
* PCI bus, but have an off-by-one error. Hence retry the
* reservation starting from 1 instead of 0.
* There's also BIOS with off-by-one on the other end.
*/
r = devm_request_mem_region(dev_priv->drm.dev, dsm->start + 1,
resource_size(dsm) - 2,
"Graphics Stolen Memory");
/*
* GEN3 firmware likes to smash pci bridges into the stolen
* range. Apparently this works.
*/
if (r == NULL && !IS_GEN(dev_priv, 3)) {
DRM_ERROR("conflict detected with stolen region: %pR\n",
dsm);
return -EBUSY;
}
}
return 0;
}
void i915_gem_cleanup_stolen(struct drm_i915_private *dev_priv)
{
if (!drm_mm_initialized(&dev_priv->mm.stolen))
return;
drm_mm_takedown(&dev_priv->mm.stolen);
}
static void g4x_get_stolen_reserved(struct drm_i915_private *dev_priv,
resource_size_t *base,
resource_size_t *size)
{
u32 reg_val = I915_READ(IS_GM45(dev_priv) ?
CTG_STOLEN_RESERVED :
ELK_STOLEN_RESERVED);
resource_size_t stolen_top = dev_priv->dsm.end + 1;
DRM_DEBUG_DRIVER("%s_STOLEN_RESERVED = %08x\n",
IS_GM45(dev_priv) ? "CTG" : "ELK", reg_val);
if ((reg_val & G4X_STOLEN_RESERVED_ENABLE) == 0)
return;
/*
* Whether ILK really reuses the ELK register for this is unclear.
* Let's see if we catch anyone with this supposedly enabled on ILK.
*/
WARN(IS_GEN(dev_priv, 5), "ILK stolen reserved found? 0x%08x\n",
reg_val);
if (!(reg_val & G4X_STOLEN_RESERVED_ADDR2_MASK))
return;
*base = (reg_val & G4X_STOLEN_RESERVED_ADDR2_MASK) << 16;
WARN_ON((reg_val & G4X_STOLEN_RESERVED_ADDR1_MASK) < *base);
*size = stolen_top - *base;
}
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
static void gen6_get_stolen_reserved(struct drm_i915_private *dev_priv,
resource_size_t *base,
resource_size_t *size)
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
{
u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED);
DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val);
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE))
return;
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
*base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK;
switch (reg_val & GEN6_STOLEN_RESERVED_SIZE_MASK) {
case GEN6_STOLEN_RESERVED_1M:
*size = 1024 * 1024;
break;
case GEN6_STOLEN_RESERVED_512K:
*size = 512 * 1024;
break;
case GEN6_STOLEN_RESERVED_256K:
*size = 256 * 1024;
break;
case GEN6_STOLEN_RESERVED_128K:
*size = 128 * 1024;
break;
default:
*size = 1024 * 1024;
MISSING_CASE(reg_val & GEN6_STOLEN_RESERVED_SIZE_MASK);
}
}
static void vlv_get_stolen_reserved(struct drm_i915_private *dev_priv,
resource_size_t *base,
resource_size_t *size)
{
u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED);
resource_size_t stolen_top = dev_priv->dsm.end + 1;
DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val);
if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE))
return;
switch (reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK) {
default:
MISSING_CASE(reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK);
/* fall through */
case GEN7_STOLEN_RESERVED_1M:
*size = 1024 * 1024;
break;
}
/*
* On vlv, the ADDR_MASK portion is left as 0 and HW deduces the
* reserved location as (top - size).
*/
*base = stolen_top - *size;
}
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
static void gen7_get_stolen_reserved(struct drm_i915_private *dev_priv,
resource_size_t *base,
resource_size_t *size)
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
{
u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED);
DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val);
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE))
return;
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
*base = reg_val & GEN7_STOLEN_RESERVED_ADDR_MASK;
switch (reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK) {
case GEN7_STOLEN_RESERVED_1M:
*size = 1024 * 1024;
break;
case GEN7_STOLEN_RESERVED_256K:
*size = 256 * 1024;
break;
default:
*size = 1024 * 1024;
MISSING_CASE(reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK);
}
}
static void chv_get_stolen_reserved(struct drm_i915_private *dev_priv,
resource_size_t *base,
resource_size_t *size)
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
{
u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED);
DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val);
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE))
return;
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
*base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK;
switch (reg_val & GEN8_STOLEN_RESERVED_SIZE_MASK) {
case GEN8_STOLEN_RESERVED_1M:
*size = 1024 * 1024;
break;
case GEN8_STOLEN_RESERVED_2M:
*size = 2 * 1024 * 1024;
break;
case GEN8_STOLEN_RESERVED_4M:
*size = 4 * 1024 * 1024;
break;
case GEN8_STOLEN_RESERVED_8M:
*size = 8 * 1024 * 1024;
break;
default:
*size = 8 * 1024 * 1024;
MISSING_CASE(reg_val & GEN8_STOLEN_RESERVED_SIZE_MASK);
}
}
static void bdw_get_stolen_reserved(struct drm_i915_private *dev_priv,
resource_size_t *base,
resource_size_t *size)
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
{
u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED);
resource_size_t stolen_top = dev_priv->dsm.end + 1;
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val);
if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE))
return;
if (!(reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK))
return;
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
*base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK;
*size = stolen_top - *base;
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
}
static void icl_get_stolen_reserved(struct drm_i915_private *dev_priv,
resource_size_t *base,
resource_size_t *size)
{
u64 reg_val = I915_READ64(GEN6_STOLEN_RESERVED);
DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = 0x%016llx\n", reg_val);
*base = reg_val & GEN11_STOLEN_RESERVED_ADDR_MASK;
switch (reg_val & GEN8_STOLEN_RESERVED_SIZE_MASK) {
case GEN8_STOLEN_RESERVED_1M:
*size = 1024 * 1024;
break;
case GEN8_STOLEN_RESERVED_2M:
*size = 2 * 1024 * 1024;
break;
case GEN8_STOLEN_RESERVED_4M:
*size = 4 * 1024 * 1024;
break;
case GEN8_STOLEN_RESERVED_8M:
*size = 8 * 1024 * 1024;
break;
default:
*size = 8 * 1024 * 1024;
MISSING_CASE(reg_val & GEN8_STOLEN_RESERVED_SIZE_MASK);
}
}
int i915_gem_init_stolen(struct drm_i915_private *dev_priv)
{
resource_size_t reserved_base, stolen_top;
resource_size_t reserved_total, reserved_size;
mutex_init(&dev_priv->mm.stolen_lock);
if (intel_vgpu_active(dev_priv)) {
DRM_INFO("iGVT-g active, disabling use of stolen memory\n");
return 0;
}
if (intel_vtd_active() && INTEL_GEN(dev_priv) < 8) {
DRM_INFO("DMAR active, disabling use of stolen memory\n");
return 0;
}
if (resource_size(&intel_graphics_stolen_res) == 0)
return 0;
dev_priv->dsm = intel_graphics_stolen_res;
if (i915_adjust_stolen(dev_priv, &dev_priv->dsm))
return 0;
GEM_BUG_ON(dev_priv->dsm.start == 0);
GEM_BUG_ON(dev_priv->dsm.end <= dev_priv->dsm.start);
stolen_top = dev_priv->dsm.end + 1;
reserved_base = stolen_top;
reserved_size = 0;
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
switch (INTEL_GEN(dev_priv)) {
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
case 2:
case 3:
break;
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
case 4:
if (!IS_G4X(dev_priv))
break;
/* fall through */
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
case 5:
g4x_get_stolen_reserved(dev_priv,
&reserved_base, &reserved_size);
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
break;
case 6:
gen6_get_stolen_reserved(dev_priv,
&reserved_base, &reserved_size);
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
break;
case 7:
if (IS_VALLEYVIEW(dev_priv))
vlv_get_stolen_reserved(dev_priv,
&reserved_base, &reserved_size);
else
gen7_get_stolen_reserved(dev_priv,
&reserved_base, &reserved_size);
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
break;
case 8:
case 9:
case 10:
if (IS_LP(dev_priv))
chv_get_stolen_reserved(dev_priv,
&reserved_base, &reserved_size);
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
else
bdw_get_stolen_reserved(dev_priv,
&reserved_base, &reserved_size);
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
break;
case 11:
default:
icl_get_stolen_reserved(dev_priv, &reserved_base,
&reserved_size);
break;
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
}
/*
* Our expectation is that the reserved space is at the top of the
* stolen region and *never* at the bottom. If we see !reserved_base,
* it likely means we failed to read the registers correctly.
*/
if (!reserved_base) {
DRM_ERROR("inconsistent reservation %pa + %pa; ignoring\n",
&reserved_base, &reserved_size);
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
reserved_base = stolen_top;
reserved_size = 0;
}
dev_priv->dsm_reserved =
(struct resource) DEFINE_RES_MEM(reserved_base, reserved_size);
if (!resource_contains(&dev_priv->dsm, &dev_priv->dsm_reserved)) {
DRM_ERROR("Stolen reserved area %pR outside stolen memory %pR\n",
&dev_priv->dsm_reserved, &dev_priv->dsm);
return 0;
drm/i915: fix stolen bios_reserved checks I started digging this when I noticed that the BDW code was just reserving 1mb by coincidence since it was reading reserved fields. Then I noticed we didn't have any values set for SNB and earlier, and that the HSW sizes were wrong. After that, I noticed that the reserved area has a specific start, and may not exactly end where the stolen memory ends. I also noticed the base pointer can be zero. So I decided to just write a single patch fixing everything instead of 20 patches that would be much harder to review. This patch may solve random stolen memory corruption/problems on almost all platforms. Notice that since this is always dealing with the top of the stolen memory, the problems are not so easy to reproduce - especially since FBC is still disabled by default. One of the major differences of this patch is that we now look at both the size and base address. By only looking at the size we were assuming that the reserved area was always at the very top of stolen, which is not always true. After we merge the patch series that allows user space to allocate stolen memory we'll be able to write IGT tests that maybe catch the bugs fixed by this patch. v2: - s/BIOS reserved/stolen reserved/g (Chris) - Don't DRM_ERROR if we can't do anything about it (Chris) - Improve debug messages (Chris). - Use the gen7 version instead of gen6 on HSW. Tom found some documentation problems, so I think with gen7 we're on the safer side (Tom). Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-08-11 01:57:32 +08:00
}
/* It is possible for the reserved area to end before the end of stolen
* memory, so just consider the start. */
reserved_total = stolen_top - reserved_base;
DRM_DEBUG_DRIVER("Memory reserved for graphics device: %lluK, usable: %lluK\n",
(u64)resource_size(&dev_priv->dsm) >> 10,
((u64)resource_size(&dev_priv->dsm) - reserved_total) >> 10);
dev_priv->stolen_usable_size =
drm/i915: Do NOT skip the first 4k of stolen memory for pre-allocated buffers v2 Before this commit the WaSkipStolenMemoryFirstPage workaround code was skipping the first 4k by passing 4096 as start of the address range passed to drm_mm_init(). This means that calling drm_mm_reserve_node() to try and reserve the firmware framebuffer so that we can inherit it would always fail, as the firmware framebuffer starts at address 0. Commit d43537610470 ("drm/i915: skip the first 4k of stolen memory on everything >= gen8") says in its commit message: "This is confirmed to fix Skylake screen flickering issues (probably caused by the fact that we initialized a ring in the first page of stolen, but I didn't 100% confirm this theory)." Which suggests that it is safe to use the first page for a linear framebuffer as the firmware is doing (see note below). This commit always passes 0 as start to drm_mm_init() and works around WaSkipStolenMemoryFirstPage in i915_gem_stolen_insert_node_in_range() by insuring the start address passed by to drm_mm_insert_node_in_range() is always 4k or more. All entry points to i915_gem_stolen.c go through i915_gem_stolen_insert_node_in_range(), so that any newly allocated objects such as ring-buffers will not be allocated in the first 4k. The one exception is i915_gem_object_create_stolen_for_preallocated() which directly calls drm_mm_reserve_node() which now will be able to use the first 4k. This fixes the i915 driver no longer being able to inherit the firmware framebuffer on gen8+, which fixes the video output changing from the vendor logo to a black screen as soon as the i915 driver is loaded (on systems without fbcon). Some notes about the mapping of the BIOS framebuffer: v1 led to some discussion if the assumption of the intel_display.c code that the firmware framebuffer is a linear mapping of the stolen memory starting at offset 0 is still correct, because that would mean that the GOP does not implement the WaSkipStolenMemoryFirstPage workaround. To verify this the following code was added at the end of i915_gem_object_create_stolen_for_preallocated() : pr_err("first ggtt entry before bind: 0x%016llx\n", readq(dev_priv->ggtt.gsm)); ret = i915_vma_bind(vma, HAS_LLC(dev_priv) ? I915_CACHE_LLC : I915_CACHE_NONE, PIN_UPDATE); pr_err("i915_vma_bind ret %d\n", ret); pr_err("first ggtt entry after bind: 0x%016llx\n", readq(dev_priv->ggtt.gsm)); Which prints the mapping of the first page, then does a vma_bind() to force update the mapping with our linear view of the framebuffer and then prints the mapping of the first page again. On an Asrock B150M Pro4S/D3 mainboard with i5-6500 CPU this prints: [ 1.651141] first ggtt entry before bind: 0x0000000078c00001 [ 1.651151] i915_vma_bind ret 0 [ 1.651152] first ggtt entry after bind: 0x0000000078c00083 And "sudo cat /proc/iomem | grep Stolen" gives: 78c00000-88bfffff : Graphics Stolen Memory There are no visual changes with this patch (BIOS vendor logo still stays in place when we inherit the BIOS framebuffer), so the vma_bind() does not impact which memory is being scanned out. The address of the first ggtt entry matches with the start of stolen and the i915_vma_bind call only changes the first gtt entry's flags, or-ing in _PAGE_RW (BIT(1)) and PPAT_CACHED (BIT(7)), which perfectly matches what we would expect based on gen8_pte_encode()'s behavior. So it seems that the GOP indeed does NOT implement the wa and the i915's code assuming a linear mapping at the start of stolen for the BIOS fb still holds true for gen8+. I've also tested this on a Cherry Trail based device (a GPD Win) with identical results (the flags are 0x1b after the vma_bind on CHT, which matches with I915_CACHE_NONE). Changed in v2: No code changes, extended the commit message with the verification that the intel_display.c BIOS framebuffer mapping is still correct. Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180420095933.16442-1-hdegoede@redhat.com
2018-04-20 17:59:33 +08:00
resource_size(&dev_priv->dsm) - reserved_total;
2016-12-15 21:23:55 +08:00
/* Basic memrange allocator for stolen space. */
drm/i915: Do NOT skip the first 4k of stolen memory for pre-allocated buffers v2 Before this commit the WaSkipStolenMemoryFirstPage workaround code was skipping the first 4k by passing 4096 as start of the address range passed to drm_mm_init(). This means that calling drm_mm_reserve_node() to try and reserve the firmware framebuffer so that we can inherit it would always fail, as the firmware framebuffer starts at address 0. Commit d43537610470 ("drm/i915: skip the first 4k of stolen memory on everything >= gen8") says in its commit message: "This is confirmed to fix Skylake screen flickering issues (probably caused by the fact that we initialized a ring in the first page of stolen, but I didn't 100% confirm this theory)." Which suggests that it is safe to use the first page for a linear framebuffer as the firmware is doing (see note below). This commit always passes 0 as start to drm_mm_init() and works around WaSkipStolenMemoryFirstPage in i915_gem_stolen_insert_node_in_range() by insuring the start address passed by to drm_mm_insert_node_in_range() is always 4k or more. All entry points to i915_gem_stolen.c go through i915_gem_stolen_insert_node_in_range(), so that any newly allocated objects such as ring-buffers will not be allocated in the first 4k. The one exception is i915_gem_object_create_stolen_for_preallocated() which directly calls drm_mm_reserve_node() which now will be able to use the first 4k. This fixes the i915 driver no longer being able to inherit the firmware framebuffer on gen8+, which fixes the video output changing from the vendor logo to a black screen as soon as the i915 driver is loaded (on systems without fbcon). Some notes about the mapping of the BIOS framebuffer: v1 led to some discussion if the assumption of the intel_display.c code that the firmware framebuffer is a linear mapping of the stolen memory starting at offset 0 is still correct, because that would mean that the GOP does not implement the WaSkipStolenMemoryFirstPage workaround. To verify this the following code was added at the end of i915_gem_object_create_stolen_for_preallocated() : pr_err("first ggtt entry before bind: 0x%016llx\n", readq(dev_priv->ggtt.gsm)); ret = i915_vma_bind(vma, HAS_LLC(dev_priv) ? I915_CACHE_LLC : I915_CACHE_NONE, PIN_UPDATE); pr_err("i915_vma_bind ret %d\n", ret); pr_err("first ggtt entry after bind: 0x%016llx\n", readq(dev_priv->ggtt.gsm)); Which prints the mapping of the first page, then does a vma_bind() to force update the mapping with our linear view of the framebuffer and then prints the mapping of the first page again. On an Asrock B150M Pro4S/D3 mainboard with i5-6500 CPU this prints: [ 1.651141] first ggtt entry before bind: 0x0000000078c00001 [ 1.651151] i915_vma_bind ret 0 [ 1.651152] first ggtt entry after bind: 0x0000000078c00083 And "sudo cat /proc/iomem | grep Stolen" gives: 78c00000-88bfffff : Graphics Stolen Memory There are no visual changes with this patch (BIOS vendor logo still stays in place when we inherit the BIOS framebuffer), so the vma_bind() does not impact which memory is being scanned out. The address of the first ggtt entry matches with the start of stolen and the i915_vma_bind call only changes the first gtt entry's flags, or-ing in _PAGE_RW (BIT(1)) and PPAT_CACHED (BIT(7)), which perfectly matches what we would expect based on gen8_pte_encode()'s behavior. So it seems that the GOP indeed does NOT implement the wa and the i915's code assuming a linear mapping at the start of stolen for the BIOS fb still holds true for gen8+. I've also tested this on a Cherry Trail based device (a GPD Win) with identical results (the flags are 0x1b after the vma_bind on CHT, which matches with I915_CACHE_NONE). Changed in v2: No code changes, extended the commit message with the verification that the intel_display.c BIOS framebuffer mapping is still correct. Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180420095933.16442-1-hdegoede@redhat.com
2018-04-20 17:59:33 +08:00
drm_mm_init(&dev_priv->mm.stolen, 0, dev_priv->stolen_usable_size);
return 0;
}
static struct sg_table *
i915_pages_create_for_stolen(struct drm_device *dev,
resource_size_t offset, resource_size_t size)
{
struct drm_i915_private *dev_priv = to_i915(dev);
struct sg_table *st;
struct scatterlist *sg;
GEM_BUG_ON(range_overflows(offset, size, resource_size(&dev_priv->dsm)));
/* We hide that we have no struct page backing our stolen object
* by wrapping the contiguous physical allocation with a fake
* dma mapping in a single scatterlist.
*/
st = kmalloc(sizeof(*st), GFP_KERNEL);
if (st == NULL)
return ERR_PTR(-ENOMEM);
if (sg_alloc_table(st, 1, GFP_KERNEL)) {
kfree(st);
return ERR_PTR(-ENOMEM);
}
sg = st->sgl;
sg->offset = 0;
sg->length = size;
sg_dma_address(sg) = (dma_addr_t)dev_priv->dsm.start + offset;
sg_dma_len(sg) = size;
return st;
}
static int i915_gem_object_get_pages_stolen(struct drm_i915_gem_object *obj)
{
struct sg_table *pages =
i915_pages_create_for_stolen(obj->base.dev,
obj->stolen->start,
obj->stolen->size);
if (IS_ERR(pages))
return PTR_ERR(pages);
__i915_gem_object_set_pages(obj, pages, obj->stolen->size);
return 0;
}
static void i915_gem_object_put_pages_stolen(struct drm_i915_gem_object *obj,
struct sg_table *pages)
{
/* Should only be called from i915_gem_object_release_stolen() */
sg_free_table(pages);
kfree(pages);
}
static void
i915_gem_object_release_stolen(struct drm_i915_gem_object *obj)
{
struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
struct drm_mm_node *stolen = fetch_and_zero(&obj->stolen);
GEM_BUG_ON(!stolen);
__i915_gem_object_unpin_pages(obj);
i915_gem_stolen_remove_node(dev_priv, stolen);
kfree(stolen);
}
static const struct drm_i915_gem_object_ops i915_gem_object_stolen_ops = {
.get_pages = i915_gem_object_get_pages_stolen,
.put_pages = i915_gem_object_put_pages_stolen,
.release = i915_gem_object_release_stolen,
};
static struct drm_i915_gem_object *
_i915_gem_object_create_stolen(struct drm_i915_private *dev_priv,
struct drm_mm_node *stolen)
{
struct drm_i915_gem_object *obj;
drm/i915: Split obj->cache_coherent to track r/w Another month, another story in the cache coherency saga. This time, we come to the realisation that i915_gem_object_is_coherent() has been reporting whether we can read from the target without requiring a cache invalidate; but we were using it in places for testing whether we could write into the object without requiring a cache flush. So split the tracking into two, one to decide before reads, one after writes. See commit e27ab73d17ef ("drm/i915: Mark CPU cache as dirty on every transition for CPU writes") for the previous entry in this saga. v2: Be verbose v3: Remove unused function (i915_gem_object_is_coherent) v4: Fix inverted coherency check prior to execbuf (from v2) v5: Add comment for nasty code where we are optimising on gcc's behalf. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101109 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101555 Testcase: igt/kms_mmap_write_crc Testcase: igt/kms_pwrite_crc Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Dongwon Kim <dongwon.kim@intel.com> Cc: Matt Roper <matthew.d.roper@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Tested-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20170811111116.10373-1-chris@chris-wilson.co.uk Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
2017-08-11 19:11:16 +08:00
unsigned int cache_level;
obj = i915_gem_object_alloc(dev_priv);
if (obj == NULL)
return NULL;
drm_gem_private_object_init(&dev_priv->drm, &obj->base, stolen->size);
i915_gem_object_init(obj, &i915_gem_object_stolen_ops);
obj->stolen = stolen;
obj->read_domains = I915_GEM_DOMAIN_CPU | I915_GEM_DOMAIN_GTT;
drm/i915: Split obj->cache_coherent to track r/w Another month, another story in the cache coherency saga. This time, we come to the realisation that i915_gem_object_is_coherent() has been reporting whether we can read from the target without requiring a cache invalidate; but we were using it in places for testing whether we could write into the object without requiring a cache flush. So split the tracking into two, one to decide before reads, one after writes. See commit e27ab73d17ef ("drm/i915: Mark CPU cache as dirty on every transition for CPU writes") for the previous entry in this saga. v2: Be verbose v3: Remove unused function (i915_gem_object_is_coherent) v4: Fix inverted coherency check prior to execbuf (from v2) v5: Add comment for nasty code where we are optimising on gcc's behalf. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101109 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101555 Testcase: igt/kms_mmap_write_crc Testcase: igt/kms_pwrite_crc Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Dongwon Kim <dongwon.kim@intel.com> Cc: Matt Roper <matthew.d.roper@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Tested-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20170811111116.10373-1-chris@chris-wilson.co.uk Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
2017-08-11 19:11:16 +08:00
cache_level = HAS_LLC(dev_priv) ? I915_CACHE_LLC : I915_CACHE_NONE;
i915_gem_object_set_cache_coherency(obj, cache_level);
if (i915_gem_object_pin_pages(obj))
goto cleanup;
return obj;
cleanup:
i915_gem_object_free(obj);
return NULL;
}
struct drm_i915_gem_object *
i915_gem_object_create_stolen(struct drm_i915_private *dev_priv,
resource_size_t size)
{
struct drm_i915_gem_object *obj;
struct drm_mm_node *stolen;
int ret;
if (!drm_mm_initialized(&dev_priv->mm.stolen))
return NULL;
if (size == 0)
return NULL;
stolen = kzalloc(sizeof(*stolen), GFP_KERNEL);
if (!stolen)
return NULL;
ret = i915_gem_stolen_insert_node(dev_priv, stolen, size, 4096);
if (ret) {
kfree(stolen);
return NULL;
}
obj = _i915_gem_object_create_stolen(dev_priv, stolen);
if (obj)
return obj;
i915_gem_stolen_remove_node(dev_priv, stolen);
kfree(stolen);
return NULL;
}
struct drm_i915_gem_object *
i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv,
resource_size_t stolen_offset,
resource_size_t gtt_offset,
resource_size_t size)
{
struct i915_ggtt *ggtt = &dev_priv->ggtt;
struct drm_i915_gem_object *obj;
struct drm_mm_node *stolen;
struct i915_vma *vma;
int ret;
if (!drm_mm_initialized(&dev_priv->mm.stolen))
return NULL;
lockdep_assert_held(&dev_priv->drm.struct_mutex);
DRM_DEBUG_DRIVER("creating preallocated stolen object: stolen_offset=%pa, gtt_offset=%pa, size=%pa\n",
&stolen_offset, &gtt_offset, &size);
/* KISS and expect everything to be page-aligned */
if (WARN_ON(size == 0) ||
WARN_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE)) ||
WARN_ON(!IS_ALIGNED(stolen_offset, I915_GTT_MIN_ALIGNMENT)))
return NULL;
stolen = kzalloc(sizeof(*stolen), GFP_KERNEL);
if (!stolen)
return NULL;
stolen->start = stolen_offset;
stolen->size = size;
mutex_lock(&dev_priv->mm.stolen_lock);
ret = drm_mm_reserve_node(&dev_priv->mm.stolen, stolen);
mutex_unlock(&dev_priv->mm.stolen_lock);
if (ret) {
DRM_DEBUG_DRIVER("failed to allocate stolen space\n");
kfree(stolen);
return NULL;
}
obj = _i915_gem_object_create_stolen(dev_priv, stolen);
if (obj == NULL) {
DRM_DEBUG_DRIVER("failed to allocate stolen object\n");
i915_gem_stolen_remove_node(dev_priv, stolen);
kfree(stolen);
return NULL;
}
/* Some objects just need physical mem from stolen space */
if (gtt_offset == I915_GTT_OFFSET_NONE)
return obj;
ret = i915_gem_object_pin_pages(obj);
if (ret)
goto err;
vma = i915_vma_instance(obj, &ggtt->vm, NULL);
if (IS_ERR(vma)) {
ret = PTR_ERR(vma);
goto err_pages;
}
/* To simplify the initialisation sequence between KMS and GTT,
* we allow construction of the stolen object prior to
* setting up the GTT space. The actual reservation will occur
* later.
*/
ret = i915_gem_gtt_reserve(&ggtt->vm, &vma->node,
size, gtt_offset, obj->cache_level,
0);
if (ret) {
DRM_DEBUG_DRIVER("failed to allocate stolen GTT space\n");
goto err_pages;
}
GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
vma->pages = obj->mm.pages;
vma->flags |= I915_VMA_GLOBAL_BIND;
__i915_vma_set_map_and_fenceable(vma);
drm/i915: Stop tracking MRU activity on VMA Our goal is to remove struct_mutex and replace it with fine grained locking. One of the thorny issues is our eviction logic for reclaiming space for an execbuffer (or GTT mmaping, among a few other examples). While eviction itself is easy to move under a per-VM mutex, performing the activity tracking is less agreeable. One solution is not to do any MRU tracking and do a simple coarse evaluation during eviction of active/inactive, with a loose temporal ordering of last insertion/evaluation. That keeps all the locking constrained to when we are manipulating the VM itself, neatly avoiding the tricky handling of possible recursive locking during execbuf and elsewhere. Note that discarding the MRU (currently implemented as a pair of lists, to avoid scanning the active list for a NONBLOCKING search) is unlikely to impact upon our efficiency to reclaim VM space (where we think a LRU model is best) as our current strategy is to use random idle replacement first before doing a search, and over time the use of softpinned 48b per-ppGTT is growing (thereby eliminating any need to perform any eviction searches, in theory at least) with the remaining users being found on much older devices (gen2-gen6). v2: Changelog and commentary rewritten to elaborate on the duality of a single list being both an inactive and active list. v3: Consolidate bool parameters into a single set of flags; don't comment on the duality of a single variable being a multiplicity of bits. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20190128102356.15037-1-chris@chris-wilson.co.uk
2019-01-28 18:23:52 +08:00
mutex_lock(&ggtt->vm.mutex);
drm/i915: Stop tracking MRU activity on VMA Our goal is to remove struct_mutex and replace it with fine grained locking. One of the thorny issues is our eviction logic for reclaiming space for an execbuffer (or GTT mmaping, among a few other examples). While eviction itself is easy to move under a per-VM mutex, performing the activity tracking is less agreeable. One solution is not to do any MRU tracking and do a simple coarse evaluation during eviction of active/inactive, with a loose temporal ordering of last insertion/evaluation. That keeps all the locking constrained to when we are manipulating the VM itself, neatly avoiding the tricky handling of possible recursive locking during execbuf and elsewhere. Note that discarding the MRU (currently implemented as a pair of lists, to avoid scanning the active list for a NONBLOCKING search) is unlikely to impact upon our efficiency to reclaim VM space (where we think a LRU model is best) as our current strategy is to use random idle replacement first before doing a search, and over time the use of softpinned 48b per-ppGTT is growing (thereby eliminating any need to perform any eviction searches, in theory at least) with the remaining users being found on much older devices (gen2-gen6). v2: Changelog and commentary rewritten to elaborate on the duality of a single list being both an inactive and active list. v3: Consolidate bool parameters into a single set of flags; don't comment on the duality of a single variable being a multiplicity of bits. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20190128102356.15037-1-chris@chris-wilson.co.uk
2019-01-28 18:23:52 +08:00
list_move_tail(&vma->vm_link, &ggtt->vm.bound_list);
mutex_unlock(&ggtt->vm.mutex);
spin_lock(&dev_priv->mm.obj_lock);
list_move_tail(&obj->mm.link, &dev_priv->mm.bound_list);
obj->bind_count++;
spin_unlock(&dev_priv->mm.obj_lock);
return obj;
err_pages:
i915_gem_object_unpin_pages(obj);
err:
i915_gem_object_put(obj);
return NULL;
}