* [PATCH V5 01/11] PCI/VGA: Move vgaarb to drivers/pci
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
@ 2021-09-11 9:30 ` Huacai Chen
2021-09-11 9:30 ` [PATCH V5 02/11] PCI/VGA: Prefer vga_default_device() Huacai Chen
` (10 subsequent siblings)
11 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-11 9:30 UTC (permalink / raw)
To: David Airlie, Daniel Vetter, Bjorn Helgaas
Cc: linux-pci, dri-devel, Xuefeng Li, Huacai Chen
From: Bjorn Helgaas <bhelgaas@google.com>
The VGA arbiter is really PCI-specific and doesn't depend on any GPU
things. Move it to the PCI subsystem.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
drivers/gpu/vga/Kconfig | 19 -------------------
drivers/gpu/vga/Makefile | 1 -
drivers/pci/Kconfig | 19 +++++++++++++++++++
drivers/pci/Makefile | 1 +
drivers/{gpu/vga => pci}/vgaarb.c | 0
5 files changed, 20 insertions(+), 20 deletions(-)
rename drivers/{gpu/vga => pci}/vgaarb.c (100%)
diff --git a/drivers/gpu/vga/Kconfig b/drivers/gpu/vga/Kconfig
index 1ad4c4ef0b5e..eb8b14ab22c3 100644
--- a/drivers/gpu/vga/Kconfig
+++ b/drivers/gpu/vga/Kconfig
@@ -1,23 +1,4 @@
# SPDX-License-Identifier: GPL-2.0-only
-config VGA_ARB
- bool "VGA Arbitration" if EXPERT
- default y
- depends on (PCI && !S390)
- help
- Some "legacy" VGA devices implemented on PCI typically have the same
- hard-decoded addresses as they did on ISA. When multiple PCI devices
- are accessed at same time they need some kind of coordination. Please
- see Documentation/gpu/vgaarbiter.rst for more details. Select this to
- enable VGA arbiter.
-
-config VGA_ARB_MAX_GPUS
- int "Maximum number of GPUs"
- default 16
- depends on VGA_ARB
- help
- Reserves space in the kernel to maintain resource locking for
- multiple GPUS. The overhead for each GPU is very small.
-
config VGA_SWITCHEROO
bool "Laptop Hybrid Graphics - GPU switching support"
depends on X86
diff --git a/drivers/gpu/vga/Makefile b/drivers/gpu/vga/Makefile
index e92064442d60..9800620deda3 100644
--- a/drivers/gpu/vga/Makefile
+++ b/drivers/gpu/vga/Makefile
@@ -1,3 +1,2 @@
# SPDX-License-Identifier: GPL-2.0-only
-obj-$(CONFIG_VGA_ARB) += vgaarb.o
obj-$(CONFIG_VGA_SWITCHEROO) += vga_switcheroo.o
diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
index 0c473d75e625..7c9e56d7b857 100644
--- a/drivers/pci/Kconfig
+++ b/drivers/pci/Kconfig
@@ -252,6 +252,25 @@ config PCIE_BUS_PEER2PEER
endchoice
+config VGA_ARB
+ bool "VGA Arbitration" if EXPERT
+ default y
+ depends on (PCI && !S390)
+ help
+ Some "legacy" VGA devices implemented on PCI typically have the same
+ hard-decoded addresses as they did on ISA. When multiple PCI devices
+ are accessed at same time they need some kind of coordination. Please
+ see Documentation/gpu/vgaarbiter.rst for more details. Select this to
+ enable VGA arbiter.
+
+config VGA_ARB_MAX_GPUS
+ int "Maximum number of GPUs"
+ default 16
+ depends on VGA_ARB
+ help
+ Reserves space in the kernel to maintain resource locking for
+ multiple GPUS. The overhead for each GPU is very small.
+
source "drivers/pci/hotplug/Kconfig"
source "drivers/pci/controller/Kconfig"
source "drivers/pci/endpoint/Kconfig"
diff --git a/drivers/pci/Makefile b/drivers/pci/Makefile
index d62c4ac4ae1b..ebe720f69b15 100644
--- a/drivers/pci/Makefile
+++ b/drivers/pci/Makefile
@@ -29,6 +29,7 @@ obj-$(CONFIG_PCI_PF_STUB) += pci-pf-stub.o
obj-$(CONFIG_PCI_ECAM) += ecam.o
obj-$(CONFIG_PCI_P2PDMA) += p2pdma.o
obj-$(CONFIG_XEN_PCIDEV_FRONTEND) += xen-pcifront.o
+obj-$(CONFIG_VGA_ARB) += vgaarb.o
# Endpoint library must be initialized before its users
obj-$(CONFIG_PCI_ENDPOINT) += endpoint/
diff --git a/drivers/gpu/vga/vgaarb.c b/drivers/pci/vgaarb.c
similarity index 100%
rename from drivers/gpu/vga/vgaarb.c
rename to drivers/pci/vgaarb.c
--
2.27.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH V5 02/11] PCI/VGA: Prefer vga_default_device()
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
2021-09-11 9:30 ` [PATCH V5 01/11] PCI/VGA: Move vgaarb to drivers/pci Huacai Chen
@ 2021-09-11 9:30 ` Huacai Chen
2021-09-11 9:30 ` [PATCH V5 03/11] PCI/VGA: Move vga_arb_integrated_gpu() earlier in file Huacai Chen
` (9 subsequent siblings)
11 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-11 9:30 UTC (permalink / raw)
To: David Airlie, Daniel Vetter, Bjorn Helgaas
Cc: linux-pci, dri-devel, Xuefeng Li, Huacai Chen, Huacai Chen
Use the vga_default_device() interface consistently instead of directly
testing vga_default. No functional change intended.
[bhelgaas: split to separate patch and extended]
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
drivers/pci/vgaarb.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 569930552957..6a5169d8578f 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -193,7 +193,7 @@ int vga_remove_vgacon(struct pci_dev *pdev)
{
int ret = 0;
- if (pdev != vga_default)
+ if (pdev != vga_default_device())
return 0;
vgaarb_info(&pdev->dev, "deactivate vga console\n");
@@ -695,7 +695,7 @@ static bool vga_arbiter_add_pci_device(struct pci_dev *pdev)
/* Deal with VGA default device. Use first enabled one
* by default if arch doesn't have it's own hook
*/
- if (vga_default == NULL &&
+ if (!vga_default_device() &&
((vgadev->owns & VGA_RSRC_LEGACY_MASK) == VGA_RSRC_LEGACY_MASK)) {
vgaarb_info(&pdev->dev, "setting as boot VGA device\n");
vga_set_default_device(pdev);
@@ -732,7 +732,7 @@ static bool vga_arbiter_del_pci_device(struct pci_dev *pdev)
goto bail;
}
- if (vga_default == pdev)
+ if (vga_default_device() == pdev)
vga_set_default_device(NULL);
if (vgadev->decodes & (VGA_RSRC_LEGACY_IO | VGA_RSRC_LEGACY_MEM))
--
2.27.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH V5 03/11] PCI/VGA: Move vga_arb_integrated_gpu() earlier in file
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
2021-09-11 9:30 ` [PATCH V5 01/11] PCI/VGA: Move vgaarb to drivers/pci Huacai Chen
2021-09-11 9:30 ` [PATCH V5 02/11] PCI/VGA: Prefer vga_default_device() Huacai Chen
@ 2021-09-11 9:30 ` Huacai Chen
2021-09-11 9:30 ` [PATCH V5 04/11] PCI/VGA: Split out vga_arb_update_default_device() Huacai Chen
` (8 subsequent siblings)
11 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-11 9:30 UTC (permalink / raw)
To: David Airlie, Daniel Vetter, Bjorn Helgaas
Cc: linux-pci, dri-devel, Xuefeng Li, Huacai Chen, Huacai Chen
Move vga_arb_integrated_gpu() earlier in file to prepare for future patch.
No functional change intended.
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
drivers/pci/vgaarb.c | 28 ++++++++++++++--------------
1 file changed, 14 insertions(+), 14 deletions(-)
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 6a5169d8578f..29e725ebaa43 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -565,6 +565,20 @@ void vga_put(struct pci_dev *pdev, unsigned int rsrc)
}
EXPORT_SYMBOL(vga_put);
+#if defined(CONFIG_ACPI)
+static bool vga_arb_integrated_gpu(struct device *dev)
+{
+ struct acpi_device *adev = ACPI_COMPANION(dev);
+
+ return adev && !strcmp(acpi_device_hid(adev), ACPI_VIDEO_HID);
+}
+#else
+static bool vga_arb_integrated_gpu(struct device *dev)
+{
+ return false;
+}
+#endif
+
/*
* Rules for using a bridge to control a VGA descendant decoding: if a bridge
* has only one VGA descendant then it can be used to control the VGA routing
@@ -1430,20 +1444,6 @@ static struct miscdevice vga_arb_device = {
MISC_DYNAMIC_MINOR, "vga_arbiter", &vga_arb_device_fops
};
-#if defined(CONFIG_ACPI)
-static bool vga_arb_integrated_gpu(struct device *dev)
-{
- struct acpi_device *adev = ACPI_COMPANION(dev);
-
- return adev && !strcmp(acpi_device_hid(adev), ACPI_VIDEO_HID);
-}
-#else
-static bool vga_arb_integrated_gpu(struct device *dev)
-{
- return false;
-}
-#endif
-
static void __init vga_arb_select_default_device(void)
{
struct pci_dev *pdev, *found = NULL;
--
2.27.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH V5 04/11] PCI/VGA: Split out vga_arb_update_default_device()
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
` (2 preceding siblings ...)
2021-09-11 9:30 ` [PATCH V5 03/11] PCI/VGA: Move vga_arb_integrated_gpu() earlier in file Huacai Chen
@ 2021-09-11 9:30 ` Huacai Chen
2021-09-11 9:30 ` [PATCH V5 05/11] PCI/VGA: Update default VGA device if a better one found Huacai Chen
` (7 subsequent siblings)
11 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-11 9:30 UTC (permalink / raw)
To: David Airlie, Daniel Vetter, Bjorn Helgaas
Cc: linux-pci, dri-devel, Xuefeng Li, Huacai Chen, Huacai Chen
This patch is the first step of the rework: If there's no default VGA
device, and we find a VGA device that owns the legacy VGA resources, we
make that device the default. Split this logic out from vga_arbiter_add_
pci_device() into a new function, vga_arb_update_default_device().
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
drivers/pci/vgaarb.c | 25 ++++++++++++++++---------
1 file changed, 16 insertions(+), 9 deletions(-)
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 29e725ebaa43..f8f95244d499 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -579,6 +579,21 @@ static bool vga_arb_integrated_gpu(struct device *dev)
}
#endif
+static void vga_arb_update_default_device(struct vga_device *vgadev)
+{
+ struct pci_dev *pdev = vgadev->pdev;
+
+ /*
+ * If we don't have a default VGA device yet, and this device owns
+ * the legacy VGA resources, make it the default.
+ */
+ if (!vga_default_device() &&
+ ((vgadev->owns & VGA_RSRC_LEGACY_MASK) == VGA_RSRC_LEGACY_MASK)) {
+ vgaarb_info(&pdev->dev, "setting as boot VGA device\n");
+ vga_set_default_device(pdev);
+ }
+}
+
/*
* Rules for using a bridge to control a VGA descendant decoding: if a bridge
* has only one VGA descendant then it can be used to control the VGA routing
@@ -706,15 +721,7 @@ static bool vga_arbiter_add_pci_device(struct pci_dev *pdev)
bus = bus->parent;
}
- /* Deal with VGA default device. Use first enabled one
- * by default if arch doesn't have it's own hook
- */
- if (!vga_default_device() &&
- ((vgadev->owns & VGA_RSRC_LEGACY_MASK) == VGA_RSRC_LEGACY_MASK)) {
- vgaarb_info(&pdev->dev, "setting as boot VGA device\n");
- vga_set_default_device(pdev);
- }
-
+ vga_arb_update_default_device(vgadev);
vga_arbiter_check_bridge_sharing(vgadev);
/* Add to the list */
--
2.27.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH V5 05/11] PCI/VGA: Update default VGA device if a better one found
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
` (3 preceding siblings ...)
2021-09-11 9:30 ` [PATCH V5 04/11] PCI/VGA: Split out vga_arb_update_default_device() Huacai Chen
@ 2021-09-11 9:30 ` Huacai Chen
2021-09-11 9:30 ` [PATCH V5 06/11] PCI/VGA: Update default VGA device again for X86/IA64 Huacai Chen
` (6 subsequent siblings)
11 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-11 9:30 UTC (permalink / raw)
To: David Airlie, Daniel Vetter, Bjorn Helgaas
Cc: linux-pci, dri-devel, Xuefeng Li, Huacai Chen, Huacai Chen
This patch is the second step of the rework: Update default VGA device if
a better one is found. What is a better one? A device with legacy I/O
resources enabled is better those not enabled. And the integrated GPU is
better than others.
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
drivers/pci/vgaarb.c | 24 +++++++++++++++++++++---
1 file changed, 21 insertions(+), 3 deletions(-)
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index f8f95244d499..02b68810273a 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -582,14 +582,32 @@ static bool vga_arb_integrated_gpu(struct device *dev)
static void vga_arb_update_default_device(struct vga_device *vgadev)
{
struct pci_dev *pdev = vgadev->pdev;
+ struct device *dev = &pdev->dev;
+ struct vga_device *vgadev_default;
/*
* If we don't have a default VGA device yet, and this device owns
* the legacy VGA resources, make it the default.
*/
- if (!vga_default_device() &&
- ((vgadev->owns & VGA_RSRC_LEGACY_MASK) == VGA_RSRC_LEGACY_MASK)) {
- vgaarb_info(&pdev->dev, "setting as boot VGA device\n");
+ if (!vga_default_device()) {
+ if ((vgadev->owns & VGA_RSRC_LEGACY_MASK) == VGA_RSRC_LEGACY_MASK)
+ vgaarb_info(dev, "setting as boot VGA device\n");
+ else
+ vgaarb_info(dev, "setting as boot device (VGA legacy resources not available)\n");
+ vga_set_default_device(pdev);
+ }
+
+ vgadev_default = vgadev_find(vga_default);
+
+ /* Overridden by a better device */
+ if (vgadev_default && ((vgadev_default->owns & VGA_RSRC_LEGACY_MASK) == 0)
+ && ((vgadev->owns & VGA_RSRC_LEGACY_MASK) == VGA_RSRC_LEGACY_MASK)) {
+ vgaarb_info(dev, "overriding boot VGA device\n");
+ vga_set_default_device(pdev);
+ }
+
+ if (vga_arb_integrated_gpu(dev)) {
+ vgaarb_info(dev, "overriding boot VGA device\n");
vga_set_default_device(pdev);
}
}
--
2.27.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH V5 06/11] PCI/VGA: Update default VGA device again for X86/IA64
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
` (4 preceding siblings ...)
2021-09-11 9:30 ` [PATCH V5 05/11] PCI/VGA: Update default VGA device if a better one found Huacai Chen
@ 2021-09-11 9:30 ` Huacai Chen
2021-09-11 9:30 ` [PATCH V5 07/11] PCI/VGA: Remove vga_arb_select_default_device() Huacai Chen
` (5 subsequent siblings)
11 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-11 9:30 UTC (permalink / raw)
To: David Airlie, Daniel Vetter, Bjorn Helgaas
Cc: linux-pci, dri-devel, Xuefeng Li, Huacai Chen, Huacai Chen
This patch is the third step of the rework: On X86 and IA64 platform,
update default VGA device if the new found device owns the firmware
framebuffer.
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
---
drivers/pci/vgaarb.c | 45 ++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 45 insertions(+)
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 02b68810273a..c768365138b6 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -584,6 +584,14 @@ static void vga_arb_update_default_device(struct vga_device *vgadev)
struct pci_dev *pdev = vgadev->pdev;
struct device *dev = &pdev->dev;
struct vga_device *vgadev_default;
+#if defined(CONFIG_X86) || defined(CONFIG_IA64)
+ int i;
+ unsigned long flags;
+ u64 base = screen_info.lfb_base;
+ u64 size = screen_info.lfb_size;
+ u64 limit;
+ resource_size_t start, end;
+#endif
/*
* If we don't have a default VGA device yet, and this device owns
@@ -610,6 +618,43 @@ static void vga_arb_update_default_device(struct vga_device *vgadev)
vgaarb_info(dev, "overriding boot VGA device\n");
vga_set_default_device(pdev);
}
+
+#if defined(CONFIG_X86) || defined(CONFIG_IA64)
+ if (screen_info.capabilities & VIDEO_CAPABILITY_64BIT_BASE)
+ base |= (u64)screen_info.ext_lfb_base << 32;
+
+ limit = base + size;
+
+ /*
+ * Override vga_arbiter_add_pci_device()'s I/O based detection
+ * as it may take the wrong device (e.g. on Apple system under
+ * EFI).
+ *
+ * Select the device owning the boot framebuffer if there is
+ * one.
+ */
+
+ /* Does firmware framebuffer belong to us? */
+ for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
+ flags = pci_resource_flags(vgadev->pdev, i);
+
+ if ((flags & IORESOURCE_MEM) == 0)
+ continue;
+
+ start = pci_resource_start(vgadev->pdev, i);
+ end = pci_resource_end(vgadev->pdev, i);
+
+ if (!start || !end)
+ continue;
+
+ if (base < start || limit >= end)
+ continue;
+
+ if (vgadev->pdev != vga_default_device())
+ vgaarb_info(dev, "overriding boot device\n");
+ vga_set_default_device(vgadev->pdev);
+ }
+#endif
}
/*
--
2.27.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH V5 07/11] PCI/VGA: Remove vga_arb_select_default_device()
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
` (5 preceding siblings ...)
2021-09-11 9:30 ` [PATCH V5 06/11] PCI/VGA: Update default VGA device again for X86/IA64 Huacai Chen
@ 2021-09-11 9:30 ` Huacai Chen
2021-09-11 9:30 ` [PATCH V5 08/11] PCI/VGA: Remove empty vga_arb_device_card_gone() Huacai Chen
` (4 subsequent siblings)
11 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-11 9:30 UTC (permalink / raw)
To: David Airlie, Daniel Vetter, Bjorn Helgaas
Cc: linux-pci, dri-devel, Xuefeng Li, Huacai Chen, Huacai Chen
This patch is the last step of the rework: Since vga_arb_update_default_
device() is complete, we can remove vga_arb_select_default_device() and
its call-site.
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
drivers/pci/vgaarb.c | 89 --------------------------------------------
1 file changed, 89 deletions(-)
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index c768365138b6..5a008d190ed8 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -1514,92 +1514,6 @@ static struct miscdevice vga_arb_device = {
MISC_DYNAMIC_MINOR, "vga_arbiter", &vga_arb_device_fops
};
-static void __init vga_arb_select_default_device(void)
-{
- struct pci_dev *pdev, *found = NULL;
- struct vga_device *vgadev;
-
-#if defined(CONFIG_X86) || defined(CONFIG_IA64)
- u64 base = screen_info.lfb_base;
- u64 size = screen_info.lfb_size;
- u64 limit;
- resource_size_t start, end;
- unsigned long flags;
- int i;
-
- if (screen_info.capabilities & VIDEO_CAPABILITY_64BIT_BASE)
- base |= (u64)screen_info.ext_lfb_base << 32;
-
- limit = base + size;
-
- list_for_each_entry(vgadev, &vga_list, list) {
- struct device *dev = &vgadev->pdev->dev;
- /*
- * Override vga_arbiter_add_pci_device()'s I/O based detection
- * as it may take the wrong device (e.g. on Apple system under
- * EFI).
- *
- * Select the device owning the boot framebuffer if there is
- * one.
- */
-
- /* Does firmware framebuffer belong to us? */
- for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
- flags = pci_resource_flags(vgadev->pdev, i);
-
- if ((flags & IORESOURCE_MEM) == 0)
- continue;
-
- start = pci_resource_start(vgadev->pdev, i);
- end = pci_resource_end(vgadev->pdev, i);
-
- if (!start || !end)
- continue;
-
- if (base < start || limit >= end)
- continue;
-
- if (!vga_default_device())
- vgaarb_info(dev, "setting as boot device\n");
- else if (vgadev->pdev != vga_default_device())
- vgaarb_info(dev, "overriding boot device\n");
- vga_set_default_device(vgadev->pdev);
- }
- }
-#endif
-
- if (!vga_default_device()) {
- list_for_each_entry_reverse(vgadev, &vga_list, list) {
- struct device *dev = &vgadev->pdev->dev;
- u16 cmd;
-
- pdev = vgadev->pdev;
- pci_read_config_word(pdev, PCI_COMMAND, &cmd);
- if (cmd & (PCI_COMMAND_IO | PCI_COMMAND_MEMORY)) {
- found = pdev;
- if (vga_arb_integrated_gpu(dev))
- break;
- }
- }
- }
-
- if (found) {
- vgaarb_info(&found->dev, "setting as boot device (VGA legacy resources not available)\n");
- vga_set_default_device(found);
- return;
- }
-
- if (!vga_default_device()) {
- vgadev = list_first_entry_or_null(&vga_list,
- struct vga_device, list);
- if (vgadev) {
- struct device *dev = &vgadev->pdev->dev;
- vgaarb_info(dev, "setting as boot device (VGA legacy resources not available)\n");
- vga_set_default_device(vgadev->pdev);
- }
- }
-}
-
static int __init vga_arb_device_init(void)
{
int rc;
@@ -1629,9 +1543,6 @@ static int __init vga_arb_device_init(void)
vgaarb_info(dev, "no bridge control possible\n");
}
- vga_arb_select_default_device();
-
- pr_info("loaded\n");
return rc;
}
subsys_initcall(vga_arb_device_init);
--
2.27.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH V5 08/11] PCI/VGA: Remove empty vga_arb_device_card_gone()
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
` (6 preceding siblings ...)
2021-09-11 9:30 ` [PATCH V5 07/11] PCI/VGA: Remove vga_arb_select_default_device() Huacai Chen
@ 2021-09-11 9:30 ` Huacai Chen
2021-09-11 9:30 ` [PATCH V5 09/11] PCI/VGA: Log bridge control messages when adding devices Huacai Chen
` (3 subsequent siblings)
11 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-11 9:30 UTC (permalink / raw)
To: David Airlie, Daniel Vetter, Bjorn Helgaas
Cc: linux-pci, dri-devel, Xuefeng Li, Huacai Chen
From: Bjorn Helgaas <bhelgaas@google.com>
vga_arb_device_card_gone() has always been empty. Remove it.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
drivers/pci/vgaarb.c | 16 +---------------
1 file changed, 1 insertion(+), 15 deletions(-)
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 5a008d190ed8..1a829cbacbc8 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -122,8 +122,6 @@ static int vga_str_to_iostate(char *buf, int str_size, int *io_state)
/* this is only used a cookie - it should not be dereferenced */
static struct pci_dev *vga_default;
-static void vga_arb_device_card_gone(struct pci_dev *pdev);
-
/* Find somebody in our list */
static struct vga_device *vgadev_find(struct pci_dev *pdev)
{
@@ -825,10 +823,6 @@ static bool vga_arbiter_del_pci_device(struct pci_dev *pdev)
/* Remove entry from list */
list_del(&vgadev->list);
vga_count--;
- /* Notify userland driver that the device is gone so it discards
- * it's copies of the pci_dev pointer
- */
- vga_arb_device_card_gone(pdev);
/* Wake up all possible waiters */
wake_up_all(&vga_wait_queue);
@@ -1078,9 +1072,7 @@ static ssize_t vga_arb_read(struct file *file, char __user *buf,
if (lbuf == NULL)
return -ENOMEM;
- /* Shields against vga_arb_device_card_gone (pci_dev going
- * away), and allows access to vga list
- */
+ /* Protects vga_list */
spin_lock_irqsave(&vga_lock, flags);
/* If we are targeting the default, use it */
@@ -1097,8 +1089,6 @@ static ssize_t vga_arb_read(struct file *file, char __user *buf,
/* Wow, it's not in the list, that shouldn't happen,
* let's fix us up and return invalid card
*/
- if (pdev == priv->target)
- vga_arb_device_card_gone(pdev);
spin_unlock_irqrestore(&vga_lock, flags);
len = sprintf(lbuf, "invalid");
goto done;
@@ -1442,10 +1432,6 @@ static int vga_arb_release(struct inode *inode, struct file *file)
return 0;
}
-static void vga_arb_device_card_gone(struct pci_dev *pdev)
-{
-}
-
/*
* callback any registered clients to let them know we have a
* change in VGA cards
--
2.27.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH V5 09/11] PCI/VGA: Log bridge control messages when adding devices
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
` (7 preceding siblings ...)
2021-09-11 9:30 ` [PATCH V5 08/11] PCI/VGA: Remove empty vga_arb_device_card_gone() Huacai Chen
@ 2021-09-11 9:30 ` Huacai Chen
2021-09-11 9:30 ` [PATCH V5 10/11] PCI/VGA: Use unsigned format string to print lock counts Huacai Chen
` (2 subsequent siblings)
11 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-11 9:30 UTC (permalink / raw)
To: David Airlie, Daniel Vetter, Bjorn Helgaas
Cc: linux-pci, dri-devel, Xuefeng Li, Huacai Chen, Huacai Chen
Previously vga_arb_device_init() iterated through all VGA devices and
indicated whether legacy VGA routing to each could be controlled by an
upstream bridge.
But we determine that information in vga_arbiter_add_pci_device(), which we
call for every device, so we can log it there without iterating through the
VGA devices again.
Note that we call vga_arbiter_check_bridge_sharing() before adding the
device to vga_list, so we have to handle the very first device separately.
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
drivers/pci/vgaarb.c | 19 ++++++++-----------
1 file changed, 8 insertions(+), 11 deletions(-)
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 1a829cbacbc8..5a042f9ac82c 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -672,8 +672,10 @@ static void vga_arbiter_check_bridge_sharing(struct vga_device *vgadev)
vgadev->bridge_has_one_vga = true;
- if (list_empty(&vga_list))
+ if (list_empty(&vga_list)) {
+ vgaarb_info(&vgadev->pdev->dev, "bridge control possible\n");
return;
+ }
/* okay iterate the new devices bridge hierarachy */
new_bus = vgadev->pdev->bus;
@@ -712,6 +714,11 @@ static void vga_arbiter_check_bridge_sharing(struct vga_device *vgadev)
}
new_bus = new_bus->parent;
}
+
+ if (vgadev->bridge_has_one_vga)
+ vgaarb_info(&vgadev->pdev->dev, "bridge control possible\n");
+ else
+ vgaarb_info(&vgadev->pdev->dev, "no bridge control possible\n");
}
/*
@@ -1504,7 +1511,6 @@ static int __init vga_arb_device_init(void)
{
int rc;
struct pci_dev *pdev;
- struct vga_device *vgadev;
rc = misc_register(&vga_arb_device);
if (rc < 0)
@@ -1520,15 +1526,6 @@ static int __init vga_arb_device_init(void)
PCI_ANY_ID, pdev)) != NULL)
vga_arbiter_add_pci_device(pdev);
- list_for_each_entry(vgadev, &vga_list, list) {
- struct device *dev = &vgadev->pdev->dev;
-
- if (vgadev->bridge_has_one_vga)
- vgaarb_info(dev, "bridge control possible\n");
- else
- vgaarb_info(dev, "no bridge control possible\n");
- }
-
return rc;
}
subsys_initcall(vga_arb_device_init);
--
2.27.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH V5 10/11] PCI/VGA: Use unsigned format string to print lock counts
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
` (8 preceding siblings ...)
2021-09-11 9:30 ` [PATCH V5 09/11] PCI/VGA: Log bridge control messages when adding devices Huacai Chen
@ 2021-09-11 9:30 ` Huacai Chen
2021-09-11 9:30 ` [PATCH V5 11/11] PCI/VGA: Replace full MIT license text with SPDX identifier Huacai Chen
2021-09-14 19:18 ` [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Bjorn Helgaas
11 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-11 9:30 UTC (permalink / raw)
To: David Airlie, Daniel Vetter, Bjorn Helgaas
Cc: linux-pci, dri-devel, Xuefeng Li, Huacai Chen
From: Bjorn Helgaas <bhelgaas@google.com>
In struct vga_device, io_lock_cnt and mem_lock_cnt are unsigned, but we
previously printed them with "%d", the signed decimal format. Print them
with the unsigned format "%u" instead.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
drivers/pci/vgaarb.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 5a042f9ac82c..f9581775939b 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -1103,7 +1103,7 @@ static ssize_t vga_arb_read(struct file *file, char __user *buf,
/* Fill the buffer with infos */
len = snprintf(lbuf, 1024,
- "count:%d,PCI:%s,decodes=%s,owns=%s,locks=%s(%d:%d)\n",
+ "count:%d,PCI:%s,decodes=%s,owns=%s,locks=%s(%u:%u)\n",
vga_decode_count, pci_name(pdev),
vga_iostate_to_str(vgadev->decodes),
vga_iostate_to_str(vgadev->owns),
--
2.27.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH V5 11/11] PCI/VGA: Replace full MIT license text with SPDX identifier
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
` (9 preceding siblings ...)
2021-09-11 9:30 ` [PATCH V5 10/11] PCI/VGA: Use unsigned format string to print lock counts Huacai Chen
@ 2021-09-11 9:30 ` Huacai Chen
2021-09-14 19:18 ` [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Bjorn Helgaas
11 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-11 9:30 UTC (permalink / raw)
To: David Airlie, Daniel Vetter, Bjorn Helgaas
Cc: linux-pci, dri-devel, Xuefeng Li, Huacai Chen
From: Bjorn Helgaas <bhelgaas@google.com>
Per Documentation/process/license-rules.rst, the SPDX MIT identifier is
equivalent to including the entire MIT license text from
LICENSES/preferred/MIT.
Replace the MIT license text with the equivalent SPDX identifier.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
drivers/pci/vgaarb.c | 23 +----------------------
1 file changed, 1 insertion(+), 22 deletions(-)
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index f9581775939b..54d0b451b335 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -1,32 +1,11 @@
+// SPDX-License-Identifier: MIT
/*
* vgaarb.c: Implements the VGA arbitration. For details refer to
* Documentation/gpu/vgaarbiter.rst
*
- *
* (C) Copyright 2005 Benjamin Herrenschmidt <benh@kernel.crashing.org>
* (C) Copyright 2007 Paulo R. Zanoni <przanoni@gmail.com>
* (C) Copyright 2007, 2009 Tiago Vignatti <vignatti@freedesktop.org>
- *
- * 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.
- *
*/
#define pr_fmt(fmt) "vgaarb: " fmt
--
2.27.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* Re: [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection
2021-09-11 9:30 [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Huacai Chen
` (10 preceding siblings ...)
2021-09-11 9:30 ` [PATCH V5 11/11] PCI/VGA: Replace full MIT license text with SPDX identifier Huacai Chen
@ 2021-09-14 19:18 ` Bjorn Helgaas
2021-09-16 8:28 ` Huacai Chen
11 siblings, 1 reply; 15+ messages in thread
From: Bjorn Helgaas @ 2021-09-14 19:18 UTC (permalink / raw)
To: Huacai Chen
Cc: David Airlie, Daniel Vetter, Bjorn Helgaas, linux-pci, dri-devel,
Xuefeng Li, Huacai Chen
On Sat, Sep 11, 2021 at 05:30:45PM +0800, Huacai Chen wrote:
> My original work is at [1].
>
> Current default VGA device selection fails in some cases:
>
> - On BMC system, the AST2500 bridge [1a03:1150] does not implement
> PCI_BRIDGE_CTL_VGA [1]. This is perfectly legal but means the
> legacy VGA resources won't reach downstream devices unless they're
> included in the usual bridge windows.
>
> - vga_arb_select_default_device() will set a device below such a
> bridge as the default VGA device as long as it has PCI_COMMAND_IO
> and PCI_COMMAND_MEMORY enabled.
>
> - vga_arbiter_add_pci_device() is called for every VGA device,
> either at boot-time or at hot-add time, and it will also set the
> device as the default VGA device, but ONLY if all bridges leading
> to it implement PCI_BRIDGE_CTL_VGA.
>
> - This difference between vga_arb_select_default_device() and
> vga_arbiter_add_pci_device() means that a device below an AST2500
> or similar bridge can only be set as the default if it is
> enumerated before vga_arb_device_init().
>
> - On ACPI-based systems, PCI devices are enumerated by acpi_init(),
> which runs before vga_arb_device_init().
>
> - On non-ACPI systems, like on MIPS system, they are enumerated by
> pcibios_init(), which typically runs *after*
> vga_arb_device_init().
>
> So I made vga_arb_update_default_device() to replace the current vga_
> arb_select_default_device(), which will be call from vga_arbiter_add_
> pci_device(), set the default device even if it does not own the VGA
> resources because an upstream bridge doesn't implement PCI_BRIDGE_CTL_
> VGA. And the default VGA device is updated if a better one is found
> (device with legacy resources enabled is better, device owns the
> firmware framebuffer is even better).
>
> Bjorn do some rework and extension in V2. It moves the VGA arbiter to
> the PCI subsystem, fixes a few nits, and breaks a few pieces to make
> the main patch a little smaller.
>
> V3 rewrite the commit log of the last patch (which is also summarized
> by Bjorn).
>
> V4 split the last patch to two steps.
>
> V5 split big patches again and sort the patches.
Not sure if I'm missing something, or if this is an interim version
and you're working on a v6.
From https://lore.kernel.org/r/20210909175926.GA996660@bjorn-Precision-5520,
I was looking for:
BUT as I mentioned, I want the very first patch to be the very
simple 2-line change to vga_arb_update_default_device() that actually
fixes your problem.
That doesn't seem to be what we have here.
> All comments welcome!
>
> [1] https://lore.kernel.org/dri-devel/20210705100503.1120643-1-chenhuacai@loongson.cn/
>
> Bjorn Helgaas (4):
> PCI/VGA: Move vgaarb to drivers/pci
> PCI/VGA: Remove empty vga_arb_device_card_gone()
> PCI/VGA: Use unsigned format string to print lock counts
> PCI/VGA: Replace full MIT license text with SPDX identifier
>
> Huacai Chen (7):
> PCI/VGA: Prefer vga_default_device()
> PCI/VGA: Move vga_arb_integrated_gpu() earlier in file
> PCI/VGA: Split out vga_arb_update_default_device()
> PCI/VGA: Update default VGA device if a better one found
> PCI/VGA: Update default VGA device again for X86/IA64
> PCI/VGA: Remove vga_arb_select_default_device()
> PCI/VGA: Log bridge control messages when adding devices
>
> Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
> ---
> drivers/gpu/vga/Kconfig | 19 ---
> drivers/gpu/vga/Makefile | 1 -
> drivers/pci/Kconfig | 19 +++
> drivers/pci/Makefile | 1 +
> drivers/{gpu/vga => pci}/vgaarb.c | 269 ++++++++++++------------------
> 5 files changed, 126 insertions(+), 183 deletions(-)
> rename drivers/{gpu/vga => pci}/vgaarb.c (90%)
> --
> 2.27.0
>
^ permalink raw reply [flat|nested] 15+ messages in thread
* Re: [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection
2021-09-14 19:18 ` [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection Bjorn Helgaas
@ 2021-09-16 8:28 ` Huacai Chen
0 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-16 8:28 UTC (permalink / raw)
To: Bjorn Helgaas
Cc: Huacai Chen, David Airlie, Daniel Vetter, Bjorn Helgaas,
linux-pci, Maling list - DRI developers, Xuefeng Li
Hi, Bjorn,
On Wed, Sep 15, 2021 at 3:18 AM Bjorn Helgaas <helgaas@kernel.org> wrote:
>
> On Sat, Sep 11, 2021 at 05:30:45PM +0800, Huacai Chen wrote:
> > My original work is at [1].
> >
> > Current default VGA device selection fails in some cases:
> >
> > - On BMC system, the AST2500 bridge [1a03:1150] does not implement
> > PCI_BRIDGE_CTL_VGA [1]. This is perfectly legal but means the
> > legacy VGA resources won't reach downstream devices unless they're
> > included in the usual bridge windows.
> >
> > - vga_arb_select_default_device() will set a device below such a
> > bridge as the default VGA device as long as it has PCI_COMMAND_IO
> > and PCI_COMMAND_MEMORY enabled.
> >
> > - vga_arbiter_add_pci_device() is called for every VGA device,
> > either at boot-time or at hot-add time, and it will also set the
> > device as the default VGA device, but ONLY if all bridges leading
> > to it implement PCI_BRIDGE_CTL_VGA.
> >
> > - This difference between vga_arb_select_default_device() and
> > vga_arbiter_add_pci_device() means that a device below an AST2500
> > or similar bridge can only be set as the default if it is
> > enumerated before vga_arb_device_init().
> >
> > - On ACPI-based systems, PCI devices are enumerated by acpi_init(),
> > which runs before vga_arb_device_init().
> >
> > - On non-ACPI systems, like on MIPS system, they are enumerated by
> > pcibios_init(), which typically runs *after*
> > vga_arb_device_init().
> >
> > So I made vga_arb_update_default_device() to replace the current vga_
> > arb_select_default_device(), which will be call from vga_arbiter_add_
> > pci_device(), set the default device even if it does not own the VGA
> > resources because an upstream bridge doesn't implement PCI_BRIDGE_CTL_
> > VGA. And the default VGA device is updated if a better one is found
> > (device with legacy resources enabled is better, device owns the
> > firmware framebuffer is even better).
> >
> > Bjorn do some rework and extension in V2. It moves the VGA arbiter to
> > the PCI subsystem, fixes a few nits, and breaks a few pieces to make
> > the main patch a little smaller.
> >
> > V3 rewrite the commit log of the last patch (which is also summarized
> > by Bjorn).
> >
> > V4 split the last patch to two steps.
> >
> > V5 split big patches again and sort the patches.
>
> Not sure if I'm missing something, or if this is an interim version
> and you're working on a v6.
>
> From https://lore.kernel.org/r/20210909175926.GA996660@bjorn-Precision-5520,
> I was looking for:
>
> BUT as I mentioned, I want the very first patch to be the very
> simple 2-line change to vga_arb_update_default_device() that actually
> fixes your problem.
>
> That doesn't seem to be what we have here.
I try my best to split patches. But it seems very difficult to achieve
the goal of "simple 2-line change" because I cannot break a single
functional change. E.g., I think I can try to split Patch-5 again,
but Patch-6 seems impossible.
Huacai
>
> > All comments welcome!
> >
> > [1] https://lore.kernel.org/dri-devel/20210705100503.1120643-1-chenhuacai@loongson.cn/
> >
> > Bjorn Helgaas (4):
> > PCI/VGA: Move vgaarb to drivers/pci
> > PCI/VGA: Remove empty vga_arb_device_card_gone()
> > PCI/VGA: Use unsigned format string to print lock counts
> > PCI/VGA: Replace full MIT license text with SPDX identifier
> >
> > Huacai Chen (7):
> > PCI/VGA: Prefer vga_default_device()
> > PCI/VGA: Move vga_arb_integrated_gpu() earlier in file
> > PCI/VGA: Split out vga_arb_update_default_device()
> > PCI/VGA: Update default VGA device if a better one found
> > PCI/VGA: Update default VGA device again for X86/IA64
> > PCI/VGA: Remove vga_arb_select_default_device()
> > PCI/VGA: Log bridge control messages when adding devices
> >
> > Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
> > Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
> > ---
> > drivers/gpu/vga/Kconfig | 19 ---
> > drivers/gpu/vga/Makefile | 1 -
> > drivers/pci/Kconfig | 19 +++
> > drivers/pci/Makefile | 1 +
> > drivers/{gpu/vga => pci}/vgaarb.c | 269 ++++++++++++------------------
> > 5 files changed, 126 insertions(+), 183 deletions(-)
> > rename drivers/{gpu/vga => pci}/vgaarb.c (90%)
> > --
> > 2.27.0
> >
^ permalink raw reply [flat|nested] 15+ messages in thread
* Re: [PATCH V5 00/11] PCI/VGA: Rework default VGA device selection
@ 2021-09-16 8:28 ` Huacai Chen
0 siblings, 0 replies; 15+ messages in thread
From: Huacai Chen @ 2021-09-16 8:28 UTC (permalink / raw)
To: Bjorn Helgaas
Cc: Huacai Chen, David Airlie, Daniel Vetter, Bjorn Helgaas,
linux-pci, Maling list - DRI developers, Xuefeng Li
Hi, Bjorn,
On Wed, Sep 15, 2021 at 3:18 AM Bjorn Helgaas <helgaas@kernel.org> wrote:
>
> On Sat, Sep 11, 2021 at 05:30:45PM +0800, Huacai Chen wrote:
> > My original work is at [1].
> >
> > Current default VGA device selection fails in some cases:
> >
> > - On BMC system, the AST2500 bridge [1a03:1150] does not implement
> > PCI_BRIDGE_CTL_VGA [1]. This is perfectly legal but means the
> > legacy VGA resources won't reach downstream devices unless they're
> > included in the usual bridge windows.
> >
> > - vga_arb_select_default_device() will set a device below such a
> > bridge as the default VGA device as long as it has PCI_COMMAND_IO
> > and PCI_COMMAND_MEMORY enabled.
> >
> > - vga_arbiter_add_pci_device() is called for every VGA device,
> > either at boot-time or at hot-add time, and it will also set the
> > device as the default VGA device, but ONLY if all bridges leading
> > to it implement PCI_BRIDGE_CTL_VGA.
> >
> > - This difference between vga_arb_select_default_device() and
> > vga_arbiter_add_pci_device() means that a device below an AST2500
> > or similar bridge can only be set as the default if it is
> > enumerated before vga_arb_device_init().
> >
> > - On ACPI-based systems, PCI devices are enumerated by acpi_init(),
> > which runs before vga_arb_device_init().
> >
> > - On non-ACPI systems, like on MIPS system, they are enumerated by
> > pcibios_init(), which typically runs *after*
> > vga_arb_device_init().
> >
> > So I made vga_arb_update_default_device() to replace the current vga_
> > arb_select_default_device(), which will be call from vga_arbiter_add_
> > pci_device(), set the default device even if it does not own the VGA
> > resources because an upstream bridge doesn't implement PCI_BRIDGE_CTL_
> > VGA. And the default VGA device is updated if a better one is found
> > (device with legacy resources enabled is better, device owns the
> > firmware framebuffer is even better).
> >
> > Bjorn do some rework and extension in V2. It moves the VGA arbiter to
> > the PCI subsystem, fixes a few nits, and breaks a few pieces to make
> > the main patch a little smaller.
> >
> > V3 rewrite the commit log of the last patch (which is also summarized
> > by Bjorn).
> >
> > V4 split the last patch to two steps.
> >
> > V5 split big patches again and sort the patches.
>
> Not sure if I'm missing something, or if this is an interim version
> and you're working on a v6.
>
> From https://lore.kernel.org/r/20210909175926.GA996660@bjorn-Precision-5520,
> I was looking for:
>
> BUT as I mentioned, I want the very first patch to be the very
> simple 2-line change to vga_arb_update_default_device() that actually
> fixes your problem.
>
> That doesn't seem to be what we have here.
I try my best to split patches. But it seems very difficult to achieve
the goal of "simple 2-line change" because I cannot break a single
functional change. E.g., I think I can try to split Patch-5 again,
but Patch-6 seems impossible.
Huacai
>
> > All comments welcome!
> >
> > [1] https://lore.kernel.org/dri-devel/20210705100503.1120643-1-chenhuacai@loongson.cn/
> >
> > Bjorn Helgaas (4):
> > PCI/VGA: Move vgaarb to drivers/pci
> > PCI/VGA: Remove empty vga_arb_device_card_gone()
> > PCI/VGA: Use unsigned format string to print lock counts
> > PCI/VGA: Replace full MIT license text with SPDX identifier
> >
> > Huacai Chen (7):
> > PCI/VGA: Prefer vga_default_device()
> > PCI/VGA: Move vga_arb_integrated_gpu() earlier in file
> > PCI/VGA: Split out vga_arb_update_default_device()
> > PCI/VGA: Update default VGA device if a better one found
> > PCI/VGA: Update default VGA device again for X86/IA64
> > PCI/VGA: Remove vga_arb_select_default_device()
> > PCI/VGA: Log bridge control messages when adding devices
> >
> > Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
> > Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
> > ---
> > drivers/gpu/vga/Kconfig | 19 ---
> > drivers/gpu/vga/Makefile | 1 -
> > drivers/pci/Kconfig | 19 +++
> > drivers/pci/Makefile | 1 +
> > drivers/{gpu/vga => pci}/vgaarb.c | 269 ++++++++++++------------------
> > 5 files changed, 126 insertions(+), 183 deletions(-)
> > rename drivers/{gpu/vga => pci}/vgaarb.c (90%)
> > --
> > 2.27.0
> >
^ permalink raw reply [flat|nested] 15+ messages in thread