xserver/hw/xfree86/os-support/shared/drm_platform.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

228 lines
5.5 KiB
C
Raw Normal View History

xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
#ifdef HAVE_XORG_CONFIG_H
#include <xorg-config.h>
#endif
#ifdef XSERVER_PLATFORM_BUS
#include <xf86drm.h>
#include <fcntl.h>
#include <unistd.h>
#include <errno.h>
#include <string.h>
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
#include "config/hotplug_priv.h"
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
/* Linux platform device support */
#include "xf86_OSproc.h"
#include "xf86.h"
#include "xf86_os_support.h"
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
#include "xf86platformBus.h"
#include "xf86Bus.h"
#include "systemd-logind.h"
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
static Bool
get_drm_info(struct OdevAttributes *attribs, char *path, int delayed_index)
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
{
drmVersionPtr v;
int fd;
int err = 0;
Bool paused, server_fd = FALSE;
LogMessage(X_INFO, "Platform probe for %s\n", attribs->syspath);
fd = systemd_logind_take_fd(attribs->major, attribs->minor, path, &paused);
if (fd != -1) {
if (paused) {
LogMessage(X_ERROR,
"Error systemd-logind returned paused fd for drm node\n");
systemd_logind_release_fd(attribs->major, attribs->minor, -1);
return FALSE;
}
attribs->fd = fd;
server_fd = TRUE;
}
if (fd == -1)
fd = open(path, O_RDWR | O_CLOEXEC, 0);
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
if (fd == -1)
return FALSE;
/* for a delayed probe we've already added the device */
if (delayed_index == -1) {
xf86_add_platform_device(attribs, FALSE);
delayed_index = xf86_num_platform_devices - 1;
}
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
if (server_fd)
xf86_platform_devices[delayed_index].flags |= XF86_PDEV_SERVER_FD;
v = drmGetVersion(fd);
if (!v) {
LogMessageVerb(X_ERROR, 1, "%s: failed to query DRM version\n", path);
goto out;
}
xf86_platform_odev_attributes(delayed_index)->driver = XNFstrdup(v->name);
drmFreeVersion(v);
out:
if (!server_fd)
close(fd);
return (err == 0);
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
}
Bool
xf86PlatformDeviceCheckBusID(struct xf86_platform_device *device, const char *busid)
{
const char *syspath = device->attribs->syspath;
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
BusType bustype;
const char *id;
if (!syspath)
return FALSE;
bustype = StringToBusType(busid, &id);
if (bustype == BUS_PCI) {
struct pci_device *pPci = device->pdev;
if (!pPci)
return FALSE;
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
if (xf86ComparePciBusString(busid,
((pPci->domain << 8)
| pPci->bus),
pPci->dev, pPci->func)) {
return TRUE;
}
}
else if (bustype == BUS_PLATFORM) {
/* match on the minimum string */
int len = strlen(id);
if (strlen(syspath) < strlen(id))
len = strlen(syspath);
if (strncmp(id, syspath, len))
return FALSE;
return TRUE;
}
else if (bustype == BUS_USB) {
if (strcasecmp(busid, device->attribs->busid))
return FALSE;
return TRUE;
}
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
return FALSE;
}
void
xf86PlatformReprobeDevice(int index, struct OdevAttributes *attribs)
{
Bool ret;
char *dpath = attribs->path;
ret = get_drm_info(attribs, dpath, index);
if (ret == FALSE) {
xf86_remove_platform_device(index);
return;
}
xf86: allow DDX driver for GPU/PCI hot-plug The current X server infrastructure sets modesetting driver as default driver to handle PCI-hotplug of a GPU device. This prevents the respective DDX driver (like AMDGPU DDX driver) to take control of the card. This patch: - Adds a few functions and fine-tunes the GPU hotplug infrastructure to allow the DDX driver to be loaded, if it is configured in the X config file options as "hotplug-driver". - Scans and updates the PCI device list before adding the new GPU device in platform, so that the association of the platform device and PCI device is in place (dev->pdev). - Adds documentation of this new option An example usage in the config file would look like: Section "OutputClass" Identifier "AMDgpu" MatchDriver "amdgpu" Driver "amdgpu" HotplugDriver "amdgpu" EndSection V2: Fixed typo in commit message (Martin) Added R-B from Adam. Added ACK from Alex and Martin. V3: Added an output class based approach for finding the DDX driver (Aaron) Rebase V4: Addressed review comment from Aaron: GPU hot-plug handling driver's name to be read from the DDX config file options. In this way only the DDX drivers interested in handling GPU hot-plug will be picked and loaded, for others modesetting driver will be used as usual. V5: Addressed review comments from Aaron: - X config option to be listed in CamelCase. - Indentation fix at one place. - Code readability related optimization. V6: Addressed review comments from Aaron: - Squash the doc in the same patch - Doc formatting changes Cc: Alex Deucher <alexander.deucher@amd.com> Suggested-by: Aaron Plattner aplattner@nvidia.com (v3) Acked-by: Martin Roukala martin.roukala@mupuf.org(v1) Acked-by: Alex Deucher alexander.deucher@amd.com (v1) Reviewed-by: Adam Jackson ajax@redhat.com(v1) Signed-off-by: Shashank Sharma shashank.sharma@amd.com
2022-08-16 13:39:59 +02:00
ret = xf86platformAddDevice(xf86PlatformFindHotplugDriver(index), index);
if (ret == -1)
xf86_remove_platform_device(index);
}
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
void
xf86PlatformDeviceProbe(struct OdevAttributes *attribs)
{
int i;
char *path = attribs->path;
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
Bool ret;
if (!path)
goto out_free;
for (i = 0; i < xf86_num_platform_devices; i++) {
char *dpath = xf86_platform_odev_attributes(i)->path;
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
if (dpath && !strcmp(path, dpath))
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
break;
}
if (i != xf86_num_platform_devices)
goto out_free;
LogMessage(X_INFO, "xfree86: Adding drm device (%s)\n", path);
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
if (!xf86VTOwner()) {
/* if we don't currently own the VT then don't probe the device,
just mark it as unowned for later use */
xf86_add_platform_device(attribs, TRUE);
return;
}
ret = get_drm_info(attribs, path, -1);
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
if (ret == FALSE)
goto out_free;
return;
out_free:
config_odev_free_attributes(attribs);
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
}
void NewGPUDeviceRequest(struct OdevAttributes *attribs)
{
int old_num = xf86_num_platform_devices;
int ret;
xf86: allow DDX driver for GPU/PCI hot-plug The current X server infrastructure sets modesetting driver as default driver to handle PCI-hotplug of a GPU device. This prevents the respective DDX driver (like AMDGPU DDX driver) to take control of the card. This patch: - Adds a few functions and fine-tunes the GPU hotplug infrastructure to allow the DDX driver to be loaded, if it is configured in the X config file options as "hotplug-driver". - Scans and updates the PCI device list before adding the new GPU device in platform, so that the association of the platform device and PCI device is in place (dev->pdev). - Adds documentation of this new option An example usage in the config file would look like: Section "OutputClass" Identifier "AMDgpu" MatchDriver "amdgpu" Driver "amdgpu" HotplugDriver "amdgpu" EndSection V2: Fixed typo in commit message (Martin) Added R-B from Adam. Added ACK from Alex and Martin. V3: Added an output class based approach for finding the DDX driver (Aaron) Rebase V4: Addressed review comment from Aaron: GPU hot-plug handling driver's name to be read from the DDX config file options. In this way only the DDX drivers interested in handling GPU hot-plug will be picked and loaded, for others modesetting driver will be used as usual. V5: Addressed review comments from Aaron: - X config option to be listed in CamelCase. - Indentation fix at one place. - Code readability related optimization. V6: Addressed review comments from Aaron: - Squash the doc in the same patch - Doc formatting changes Cc: Alex Deucher <alexander.deucher@amd.com> Suggested-by: Aaron Plattner aplattner@nvidia.com (v3) Acked-by: Martin Roukala martin.roukala@mupuf.org(v1) Acked-by: Alex Deucher alexander.deucher@amd.com (v1) Reviewed-by: Adam Jackson ajax@redhat.com(v1) Signed-off-by: Shashank Sharma shashank.sharma@amd.com
2022-08-16 13:39:59 +02:00
const char *driver_name;
xf86PlatformDeviceProbe(attribs);
if (old_num == xf86_num_platform_devices)
return;
if (xf86_get_platform_device_unowned(xf86_num_platform_devices - 1) == TRUE)
return;
xf86: allow DDX driver for GPU/PCI hot-plug The current X server infrastructure sets modesetting driver as default driver to handle PCI-hotplug of a GPU device. This prevents the respective DDX driver (like AMDGPU DDX driver) to take control of the card. This patch: - Adds a few functions and fine-tunes the GPU hotplug infrastructure to allow the DDX driver to be loaded, if it is configured in the X config file options as "hotplug-driver". - Scans and updates the PCI device list before adding the new GPU device in platform, so that the association of the platform device and PCI device is in place (dev->pdev). - Adds documentation of this new option An example usage in the config file would look like: Section "OutputClass" Identifier "AMDgpu" MatchDriver "amdgpu" Driver "amdgpu" HotplugDriver "amdgpu" EndSection V2: Fixed typo in commit message (Martin) Added R-B from Adam. Added ACK from Alex and Martin. V3: Added an output class based approach for finding the DDX driver (Aaron) Rebase V4: Addressed review comment from Aaron: GPU hot-plug handling driver's name to be read from the DDX config file options. In this way only the DDX drivers interested in handling GPU hot-plug will be picked and loaded, for others modesetting driver will be used as usual. V5: Addressed review comments from Aaron: - X config option to be listed in CamelCase. - Indentation fix at one place. - Code readability related optimization. V6: Addressed review comments from Aaron: - Squash the doc in the same patch - Doc formatting changes Cc: Alex Deucher <alexander.deucher@amd.com> Suggested-by: Aaron Plattner aplattner@nvidia.com (v3) Acked-by: Martin Roukala martin.roukala@mupuf.org(v1) Acked-by: Alex Deucher alexander.deucher@amd.com (v1) Reviewed-by: Adam Jackson ajax@redhat.com(v1) Signed-off-by: Shashank Sharma shashank.sharma@amd.com
2022-08-16 13:39:59 +02:00
/* Scan and update PCI devices before adding new platform device */
xf86PlatformScanPciDev();
driver_name = xf86PlatformFindHotplugDriver(xf86_num_platform_devices - 1);
ret = xf86platformAddDevice(driver_name, xf86_num_platform_devices-1);
if (ret == -1)
xf86_remove_platform_device(xf86_num_platform_devices-1);
ErrorF("xf86: found device %d\n", xf86_num_platform_devices);
return;
}
void DeleteGPUDeviceRequest(struct OdevAttributes *attribs)
{
int index;
char *syspath = attribs->syspath;
if (!syspath)
goto out;
for (index = 0; index < xf86_num_platform_devices; index++) {
char *dspath = xf86_platform_odev_attributes(index)->syspath;
if (dspath && !strcmp(syspath, dspath))
break;
}
if (index == xf86_num_platform_devices)
goto out;
ErrorF("xf86: remove device %d %s\n", index, syspath);
if (xf86_get_platform_device_unowned(index) == TRUE)
xf86_remove_platform_device(index);
else
xf86platformRemoveDevice(index);
out:
config_odev_free_attributes(attribs);
}
xfree86: use udev to provide device enumeration for kms devices (v10) On Linux in order for future hotplug work, we are required to interface to udev to detect device creation/removal. In order to try and get some earlier testing on this, this patch adds the ability to use udev for device enumeration on Linux. At startup the list of drm/kms devices is probed and this info is used to load drivers. A new driver probing method is introduced that passes the udev device info to the driver for probing. The probing integrates with the pci probing code and will fallback to the pci probe and old school probe functions in turn. The flags parameter to the probe function will be used later to provide hotplug and gpu screen flags for the driver to behave in a different way. This patch changes the driver ABI, all drivers should at least be set with a NULL udev probe function after this commit. v2: rename to platform bus, now with 100% less udev specific, this version passes config_odev_attribs around which are an array of id/string pairs, then the udev code can attach the set of attribs it understands, the OS specific code can attach its attrib, and then the core/drivers can lookup the required attribs. also add MATCH_PCI_DEVICES macro. This version is mainly to address concerns raised by ajax. v3: Address comments from Peter. fix whitespace that snuck in. rework to use a linked list with some core functions that xf86 wraps. v4: add free list, fix struct whitespace. ajax this address most of your issues? v5: drop probe ifdef, fix logic issue v6: some overhaul after more testing. Implement primaryBus for platform devices. document hotplug.h dev attribs - drop sysname attrib fix build with udev kms disabled make probing work like the PCI probe code, match against bus id if one exists, or primary device. RFC: add new bus id support "PLAT:syspath". we probably want to match on this a bit different, or use a different property maybe. I was mainly wanting this for use with specifying usb devices in xorg.conf directly, but PLAT:path could also work I suppose. v6.1: add missing noop platform function v7: fix two interactions with pci probing and slot claiming, prevents pci and platform trying to load two drivers for same slot. v8: test with zaphod mode on -ati driver, fixup resulting issue clean up common probe code into another function, change busid matching to allow dropping end of strings. v9: fix platform probing logic so it actually works. v9.1: fix pdev init to NULL properly. v10: address most of Keith's concerns. v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com> v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> Reviewed-by: Keith Packard <keithp@keithp.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-05-08 14:26:47 +01:00
#endif