Am 11.12.20 um 01:55 schrieb Darren Salt:
This allows BAR0 resizing to be done for cards which don't advertise support
for a size large enough to cover the VRAM but which do advertise at least
one size larger than the default. For example, my RX 5600 XT, which
advertises 256MB, 512MB and 1GB.
[v2] rewritten to use PCI helper functions; some extra log text.
Signed-off-by: Darren Salt <devspam@xxxxxxxxxxxxxxxxx>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 53 ++++++++++++++++++----
1 file changed, 43 insertions(+), 10 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 6637b84aeb85..1e99ca62a4d2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -1106,21 +1106,24 @@ void amdgpu_device_wb_free(struct amdgpu_device *adev, u32 wb)
*/
int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev)
{
- u64 space_needed = roundup_pow_of_two(adev->gmc.real_vram_size);
- u32 rbar_size = order_base_2(((space_needed >> 20) | 1)) - 1;
+ int rbar_size, current_size;
+ u32 available_sizes;
struct pci_bus *root;
struct resource *res;
unsigned i;
u16 cmd;
int r;
+ bool nospc = false;
/* Bypass for VF */
if (amdgpu_sriov_vf(adev))
return 0;
- /* skip if the bios has already enabled large BAR */
- if (adev->gmc.real_vram_size &&
- (pci_resource_len(adev->pdev, 0) >= adev->gmc.real_vram_size))
+ rbar_size = pci_rebar_bytes_to_size(adev->gmc.real_vram_size);
+ current_size = pci_rebar_get_current_size(adev->pdev, 0);
+
+ /* Skip if the BIOS has already enabled large BAR, covering the VRAM */
+ if (current_size >= rbar_size)
You should probably keep the comparison as it is and check the resource
length against the VRAM size instead.
return 0;
/* Check if the root BUS has 64bit memory resources */
@@ -1138,6 +1141,14 @@ int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev)
if (!res)
return 0;
+ available_sizes = pci_rebar_get_possible_sizes(adev->pdev, 0);
+ if (available_sizes == 0)
+ return 0;
+
+ dev_dbg(adev->dev, "BIOS-allocated BAR0 was %lluMB; trying to get %lluMB",
+ current_size < 0 ? 0 : (pci_rebar_size_to_bytes(current_size) >> 20),
+ pci_rebar_size_to_bytes(rbar_size) >> 20);
Please no extra debugging output, we spam syslog that enough with the
existing resize.
+
/* Disable memory decoding while we change the BAR addresses and size */
pci_read_config_word(adev->pdev, PCI_COMMAND, &cmd);
pci_write_config_word(adev->pdev, PCI_COMMAND,
@@ -1150,11 +1161,33 @@ int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev)
pci_release_resource(adev->pdev, 0);
- r = pci_resize_resource(adev->pdev, 0, rbar_size);
- if (r == -ENOSPC)
- DRM_INFO("Not enough PCI address space for a large BAR.");
- else if (r && r != -ENOTSUPP)
- DRM_ERROR("Problem resizing BAR0 (%d).", r);
+ r = 0;
+ for (; rbar_size >= 0 && rbar_size > current_size; --rbar_size) {
Well exactly that try and error is a rather big NAK.
What you need to do instead is to look at the return value from
pci_rebar_get_possible_sizes() and determine the size closed to the
desired one.
E.g. when need a size of 13 is needed you first check if any bit >= 13
are set. You can use the ffs() for this.
If that isn't the case use fls() to get the highest set bit < 13.
Regards,
Christian.
+ /* Skip this size if it isn't advertised.
+ * This avoids pci_resize_resources returning -EINVAL for that reason.
+ */
+ if (!(available_sizes & BIT(rbar_size)))
+ continue;
+
+ r = pci_resize_resource(adev->pdev, 0, rbar_size);
+ if (r == 0) {
+ dev_dbg(adev->dev, "Succeeded in resizing to %lluMB.",
+ pci_rebar_size_to_bytes(rbar_size) >> 20);
+ break;
+ } else if (r == -ENOTSUPP) {
+ dev_info(adev->dev, "BAR resizing not supported.");
+ break;
+ } else if (r == -ENOSPC) {
+ if (!nospc) {
+ /* Warn only the first time */
+ dev_info(adev->dev, "Not enough PCI address space for a large BAR.");
+ nospc = true;
+ }
+ } else {
+ dev_err(adev->dev, "Problem resizing BAR0 (%d).", r);
+ break;
+ }
+ }
pci_assign_unassigned_bus_resources(adev->pdev->bus);
_______________________________________________
amd-gfx mailing list
amd-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
