[RFC PATCH 0/2] libdrm: Add master <> render node helpers

[RFC PATCH 0/2] libdrm: Add master <> render node helpers

[Emil Velikov]

Hi all,

As mentioned a couple of days ago at #dri-devel some(most) users of 
render nodes tend to rely on strict mapping between the primary and 
render node. I.e. something along the lines of

  fstat(render_fd, &sbuf);
  sprintf(primary_node, "/dev/dri/card%d",
                        ((sbuf.st_rdev & 0x3f) | 0x80));

Currently the following are (ab)using the above code:
 - xf86-video-nouveau
 - xf86-video-intel
 - libva (vaapi)

As reminded by David Herrman, this is not the correct solution - thus 
I've added a couple of helpers which walk through sysfs of the 
respecitive device and return the correct device name.

I'm not 100% happy with the function names, so suggestions are greatly 
appreciated. Any other comments are also welcome :)

Note: BSD guys - you'll likely need your own version of these functions.

Cheers,
Emil

Cc: Thomas Klausner <wiz@NetBSD.org>
Cc: Konstantin Belousov <kib@freebsd.org>
Cc: Robert Millan <rmh@freebsd.org>
Cc: Mark Kettenis <kettenis@openbsd.org>
Cc: Jonathan Gray <jsg@jsg.id.au>
Cc: François Tigeot <ftigeot@wolfpond.org>
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH 1/2] drm: add drmGet{Device, Render}NameFrom{Render, Device}Fd helpers

[Emil Velikov]

Currently most places assume reliable primary <> render node mapping.
Although this may work in some cases, it is not correct.

Add a couple of helpers that hide the details and safes all the
guesswork for the user.

Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: David Herrmann <dh.herrmann@googlemail.com>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
---
 xf86drm.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 xf86drm.h |  3 +++
 2 files changed, 57 insertions(+)

diff --git a/xf86drm.c b/xf86drm.c
index 345325a..6af7ac0 100644
--- a/xf86drm.c
+++ b/xf86drm.c
@@ -2607,3 +2607,57 @@ int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle)
 	return 0;
 }
 
+char *drmGetDeviceNameFromRenderFD(int fd)
+{
+
+	struct stat sbuf;
+	char name[64], buf[64];
+	int maj, min, i;
+
+	if (!fstat(fd, &sbuf))
+		return NULL;
+
+	if (!S_ISCHR(sbuf.st_mode))
+		return NULL;
+
+	maj = major(sbuf.st_rdev);
+	min = minor(sbuf.st_rdev);
+
+	snprintf(buf, sizeof(buf), "/sys/dev/char/%d:%d/device/drm", maj, min);
+
+	for (i = 0; i < DRM_MAX_MINOR; i++) {
+		snprintf(name, sizeof(name), DRM_DEV_NAME, buf, i);
+		if (stat(name, &sbuf) == 0) {
+			snprintf(name, sizeof(name), DRM_DEV_NAME, DRM_DIR_NAME, i);
+			return strdup(name);
+		}
+	}
+	return NULL;
+}
+
+char *drmGetRenderNameFromDeviceFD(int fd)
+{
+	struct stat sbuf;
+	char name[64], buf[64];
+	int maj, min, i;
+
+	if (!fstat(fd, &sbuf))
+		return NULL;
+
+	if (!S_ISCHR(sbuf.st_mode))
+		return NULL;
+
+	maj = major(sbuf.st_rdev);
+	min = minor(sbuf.st_rdev);
+
+	snprintf(buf, sizeof(buf), "/sys/dev/char/%d:%d/device/drm", maj, min);
+
+	for (i = 128; i < (128 + DRM_MAX_MINOR); i++) {
+		snprintf(name, sizeof(name), DRM_DEV_NAME, buf, i);
+		if (stat(name, &sbuf) == 0) {
+			snprintf(name, sizeof(name), DRM_RENDER_DEV_NAME, DRM_DIR_NAME, i);
+			return strdup(name);
+		}
+	}
+	return NULL;
+}
diff --git a/xf86drm.h b/xf86drm.h
index bfd0670..bca5887 100644
--- a/xf86drm.h
+++ b/xf86drm.h
@@ -740,6 +740,9 @@ extern char *drmGetDeviceNameFromFd(int fd);
 extern int drmPrimeHandleToFD(int fd, uint32_t handle, uint32_t flags, int *prime_fd);
 extern int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle);
 
+extern char *drmGetRenderNameFromDeviceFD(int fd);
+extern char *drmGetDeviceNameFromRenderFD(int fd);
+
 #if defined(__cplusplus) || defined(c_plusplus)
 }
 #endif
-- 
2.2.2

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH 1/2] drm: add drmGet{Device, Render}NameFrom{Render, Device}Fd helpers

[Frank Binns]

Hi Emil,

On 02/02/15 00:14, Emil Velikov wrote:
> Currently most places assume reliable primary <> render node mapping.
> Although this may work in some cases, it is not correct.
>
> Add a couple of helpers that hide the details and safes all the
> guesswork for the user.
>
> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> Cc: David Herrmann <dh.herrmann@googlemail.com>
> Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
> ---
>  xf86drm.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  xf86drm.h |  3 +++
>  2 files changed, 57 insertions(+)
>
> diff --git a/xf86drm.c b/xf86drm.c
> index 345325a..6af7ac0 100644
> --- a/xf86drm.c
> +++ b/xf86drm.c
> @@ -2607,3 +2607,57 @@ int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle)
>  	return 0;
>  }
>  
> +char *drmGetDeviceNameFromRenderFD(int fd)
> +{
> +
> +	struct stat sbuf;
> +	char name[64], buf[64];
> +	int maj, min, i;
> +
> +	if (!fstat(fd, &sbuf))
> +		return NULL;
> +
fstat returns 0 on success and -1 otherwise so this should be

if (fstat(fd, &sbuf))
    return NULL;

> +	if (!S_ISCHR(sbuf.st_mode))
> +		return NULL;
> +
> +	maj = major(sbuf.st_rdev);
> +	min = minor(sbuf.st_rdev);
> +
> +	snprintf(buf, sizeof(buf), "/sys/dev/char/%d:%d/device/drm", maj, min);
> +
> +	for (i = 0; i < DRM_MAX_MINOR; i++) {
> +		snprintf(name, sizeof(name), DRM_DEV_NAME, buf, i);
> +		if (stat(name, &sbuf) == 0) {
> +			snprintf(name, sizeof(name), DRM_DEV_NAME, DRM_DIR_NAME, i);
> +			return strdup(name);
> +		}
> +	}
> +	return NULL;
> +}
> +
> +char *drmGetRenderNameFromDeviceFD(int fd)
> +{
> +	struct stat sbuf;
> +	char name[64], buf[64];
> +	int maj, min, i;
> +
> +	if (!fstat(fd, &sbuf))
> +		return NULL;
> +
Same here.

> +	if (!S_ISCHR(sbuf.st_mode))
> +		return NULL;
> +
> +	maj = major(sbuf.st_rdev);
> +	min = minor(sbuf.st_rdev);
> +
> +	snprintf(buf, sizeof(buf), "/sys/dev/char/%d:%d/device/drm", maj, min);
> +
> +	for (i = 128; i < (128 + DRM_MAX_MINOR); i++) {
> +		snprintf(name, sizeof(name), DRM_DEV_NAME, buf, i);
> +		if (stat(name, &sbuf) == 0) {
> +			snprintf(name, sizeof(name), DRM_RENDER_DEV_NAME, DRM_DIR_NAME, i);
> +			return strdup(name);
> +		}
> +	}
> +	return NULL;
> +}
> diff --git a/xf86drm.h b/xf86drm.h
> index bfd0670..bca5887 100644
> --- a/xf86drm.h
> +++ b/xf86drm.h
> @@ -740,6 +740,9 @@ extern char *drmGetDeviceNameFromFd(int fd);
>  extern int drmPrimeHandleToFD(int fd, uint32_t handle, uint32_t flags, int *prime_fd);
>  extern int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle);
>  
> +extern char *drmGetRenderNameFromDeviceFD(int fd);
> +extern char *drmGetDeviceNameFromRenderFD(int fd);
> +
>  #if defined(__cplusplus) || defined(c_plusplus)
>  }
>  #endif

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH libdrm v2] drm: add drmGet(Master|Render)NameFrom(Render|Master)FD functions

[Emil Velikov]

Currently most places assume reliable master <> render node mapping.
Although this may work in some cases, it is not correct.

Add a couple of helpers that hide the details and provide the name of
the master/render device name, given an render/master FD.

v2:
 - Rename Device and Primary to Master (aka the /dev/dri/cardX device).
 - Check for the file via readdir_r() rather than stat().
 - Wrap the check into a single function.
 - Return NULL for non-linux platforms.

Cc: Frank Binns <frank.binns@imgtec.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: David Herrmann <dh.herrmann@googlemail.com>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
---
 xf86drm.c | 82 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 xf86drm.h |  3 +++
 2 files changed, 85 insertions(+)

diff --git a/xf86drm.c b/xf86drm.c
index e117bc6..d4a4dc6 100644
--- a/xf86drm.c
+++ b/xf86drm.c
@@ -40,6 +40,8 @@
 #include <string.h>
 #include <strings.h>
 #include <ctype.h>
+#include <dirent.h>
+#include <stddef.h>
 #include <fcntl.h>
 #include <errno.h>
 #include <signal.h>
@@ -522,6 +524,20 @@ static int drmGetMinorType(int minor)
     }
 }
 
+static const char *drmGetMinorName(int type)
+{
+    switch (type) {
+    case DRM_NODE_PRIMARY:
+        return "card";
+    case DRM_NODE_CONTROL:
+        return "controlD";
+    case DRM_NODE_RENDER:
+        return "renderD";
+    default:
+        return NULL;
+    }
+}
+
 /**
  * Open the device by bus ID.
  *
@@ -2736,3 +2752,69 @@ int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle)
 	return 0;
 }
 
+static char *drmGetMinorNameForFD(int fd, int type)
+{
+#ifdef __linux__
+	DIR *sysdir;
+	struct dirent *pent, *ent;
+	struct stat sbuf;
+	const char *name = drmGetMinorName(type);
+	const int len = strlen(name);
+	char dev_name[64], buf[64];
+	long name_max;
+	int maj, min;
+
+	if (!name)
+		return NULL;
+
+	if (fstat(fd, &sbuf))
+		return NULL;
+
+	maj = major(sbuf.st_rdev);
+	min = minor(sbuf.st_rdev);
+
+	if (maj != DRM_MAJOR || !S_ISCHR(sbuf.st_mode))
+		return NULL;
+
+	snprintf(buf, sizeof(buf), "/sys/dev/char/%d:%d/device/drm", maj, min);
+
+	sysdir = opendir(buf);
+	if (!sysdir)
+		return NULL;
+
+	name_max = fpathconf(dirfd(sysdir), _PC_NAME_MAX);
+	if (name_max == -1)
+		goto out_close_dir;
+
+	pent = malloc(offsetof(struct dirent, d_name) + name_max + 1);
+	if (pent == NULL)
+		 goto out_close_dir;
+
+	while (readdir_r(sysdir, pent, &ent) == 0 && ent != NULL) {
+		if (strncmp(ent->d_name, name, len) == 0) {
+			free(pent);
+			closedir(sysdir);
+
+			snprintf(dev_name, sizeof(dev_name), DRM_DIR_NAME "/%s",
+				 ent->d_name);
+			return strdup(dev_name);
+		}
+	}
+
+	free(pent);
+
+out_close_dir:
+	closedir(sysdir);
+#endif
+	return NULL;
+}
+
+char *drmGetMasterNameFromRenderFD(int fd)
+{
+	return drmGetMinorNameForFD(fd, DRM_NODE_PRIMARY);
+}
+
+char *drmGetRenderNameFromMasterFD(int fd)
+{
+	return drmGetMinorNameForFD(fd, DRM_NODE_RENDER);
+}
diff --git a/xf86drm.h b/xf86drm.h
index afd38a1..5fdf27b 100644
--- a/xf86drm.h
+++ b/xf86drm.h
@@ -749,6 +749,9 @@ extern int drmGetNodeTypeFromFd(int fd);
 extern int drmPrimeHandleToFD(int fd, uint32_t handle, uint32_t flags, int *prime_fd);
 extern int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle);
 
+extern char *drmGetRenderNameFromMasterFD(int fd);
+extern char *drmGetMasterNameFromRenderFD(int fd);
+
 #if defined(__cplusplus) || defined(c_plusplus)
 }
 #endif
-- 
2.3.0

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH libdrm v2] drm: add drmGet(Master|Render)NameFrom(Render|Master)FD functions

[Frank Binns]

Hi Emil,

On 23/02/15 12:22, Emil Velikov wrote:
> Currently most places assume reliable master <> render node mapping.
> Although this may work in some cases, it is not correct.
>
> Add a couple of helpers that hide the details and provide the name of
> the master/render device name, given an render/master FD.
>
> v2:
>  - Rename Device and Primary to Master (aka the /dev/dri/cardX device).
>  - Check for the file via readdir_r() rather than stat().
>  - Wrap the check into a single function.
>  - Return NULL for non-linux platforms.
>
> Cc: Frank Binns <frank.binns@imgtec.com>
> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> Cc: David Herrmann <dh.herrmann@googlemail.com>
> Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
> ---
>  xf86drm.c | 82 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  xf86drm.h |  3 +++
>  2 files changed, 85 insertions(+)
>
> diff --git a/xf86drm.c b/xf86drm.c
> index e117bc6..d4a4dc6 100644
> --- a/xf86drm.c
> +++ b/xf86drm.c
> @@ -40,6 +40,8 @@
>  #include <string.h>
>  #include <strings.h>
>  #include <ctype.h>
> +#include <dirent.h>
> +#include <stddef.h>
>  #include <fcntl.h>
>  #include <errno.h>
>  #include <signal.h>
> @@ -522,6 +524,20 @@ static int drmGetMinorType(int minor)
>      }
>  }
>  
> +static const char *drmGetMinorName(int type)
> +{
> +    switch (type) {
> +    case DRM_NODE_PRIMARY:
> +        return "card";
> +    case DRM_NODE_CONTROL:
> +        return "controlD";
> +    case DRM_NODE_RENDER:
> +        return "renderD";
> +    default:
> +        return NULL;
> +    }
> +}
> +
>  /**
>   * Open the device by bus ID.
>   *
> @@ -2736,3 +2752,69 @@ int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle)
>  	return 0;
>  }
>  
> +static char *drmGetMinorNameForFD(int fd, int type)
> +{
> +#ifdef __linux__
> +	DIR *sysdir;
> +	struct dirent *pent, *ent;
> +	struct stat sbuf;
> +	const char *name = drmGetMinorName(type);
> +	const int len = strlen(name);
This will cause a segfault if 'name' is NULL.

> +	char dev_name[64], buf[64];
> +	long name_max;
> +	int maj, min;
> +
> +	if (!name)
> +		return NULL;
> +
> +	if (fstat(fd, &sbuf))
> +		return NULL;
> +
> +	maj = major(sbuf.st_rdev);
> +	min = minor(sbuf.st_rdev);
> +
> +	if (maj != DRM_MAJOR || !S_ISCHR(sbuf.st_mode))
> +		return NULL;
> +
> +	snprintf(buf, sizeof(buf), "/sys/dev/char/%d:%d/device/drm", maj, min);
> +
> +	sysdir = opendir(buf);
> +	if (!sysdir)
> +		return NULL;
> +
> +	name_max = fpathconf(dirfd(sysdir), _PC_NAME_MAX);
> +	if (name_max == -1)
> +		goto out_close_dir;
> +
> +	pent = malloc(offsetof(struct dirent, d_name) + name_max + 1);
> +	if (pent == NULL)
> +		 goto out_close_dir;
> +
> +	while (readdir_r(sysdir, pent, &ent) == 0 && ent != NULL) {
> +		if (strncmp(ent->d_name, name, len) == 0) {
> +			free(pent);
> +			closedir(sysdir);
> +
> +			snprintf(dev_name, sizeof(dev_name), DRM_DIR_NAME "/%s",
> +				 ent->d_name);
> +			return strdup(dev_name);
> +		}
> +	}
> +
> +	free(pent);
> +
> +out_close_dir:
> +	closedir(sysdir);
> +#endif
> +	return NULL;
> +}
> +
> +char *drmGetMasterNameFromRenderFD(int fd)
I think drmGetPrimaryDeviceNameFromFd would be more appropriate given
the node type is 'primary', the type of the fd doesn't matter afaics and
for consistency with other drmGet* functions. However, given that's a
bit of a mouthful I guess the 'Device' part could be dropped.

> +{
> +	return drmGetMinorNameForFD(fd, DRM_NODE_PRIMARY);
> +}
> +
> +char *drmGetRenderNameFromMasterFD(int fd)
As above, maybe drmGetRenderDeviceNameFromFd?

With those things changed/fixed you can have a:
Reviewed-by: Frank Binns <frank.binns@imgtec.com>

Thanks
Frank

> +{
> +	return drmGetMinorNameForFD(fd, DRM_NODE_RENDER);
> +}
> diff --git a/xf86drm.h b/xf86drm.h
> index afd38a1..5fdf27b 100644
> --- a/xf86drm.h
> +++ b/xf86drm.h
> @@ -749,6 +749,9 @@ extern int drmGetNodeTypeFromFd(int fd);
>  extern int drmPrimeHandleToFD(int fd, uint32_t handle, uint32_t flags, int *prime_fd);
>  extern int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle);
>  
> +extern char *drmGetRenderNameFromMasterFD(int fd);
> +extern char *drmGetMasterNameFromRenderFD(int fd);
> +
>  #if defined(__cplusplus) || defined(c_plusplus)
>  }
>  #endif

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH libdrm v2] drm: add drmGet(Master|Render)NameFrom(Render|Master)FD functions

[Emil Velikov]

On 23 February 2015 at 14:35, Frank Binns <frank.binns@imgtec.com> wrote:
> Hi Emil,
>
> On 23/02/15 12:22, Emil Velikov wrote:
>> Currently most places assume reliable master <> render node mapping.
>> Although this may work in some cases, it is not correct.
>>
>> Add a couple of helpers that hide the details and provide the name of
>> the master/render device name, given an render/master FD.
>>
>> v2:
>>  - Rename Device and Primary to Master (aka the /dev/dri/cardX device).
>>  - Check for the file via readdir_r() rather than stat().
>>  - Wrap the check into a single function.
>>  - Return NULL for non-linux platforms.
>>
>> Cc: Frank Binns <frank.binns@imgtec.com>
>> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
>> Cc: David Herrmann <dh.herrmann@googlemail.com>
>> Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
>> ---
>>  xf86drm.c | 82 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>  xf86drm.h |  3 +++
>>  2 files changed, 85 insertions(+)
>>
>> diff --git a/xf86drm.c b/xf86drm.c
>> index e117bc6..d4a4dc6 100644
>> --- a/xf86drm.c
>> +++ b/xf86drm.c
>> @@ -40,6 +40,8 @@
>>  #include <string.h>
>>  #include <strings.h>
>>  #include <ctype.h>
>> +#include <dirent.h>
>> +#include <stddef.h>
>>  #include <fcntl.h>
>>  #include <errno.h>
>>  #include <signal.h>
>> @@ -522,6 +524,20 @@ static int drmGetMinorType(int minor)
>>      }
>>  }
>>
>> +static const char *drmGetMinorName(int type)
>> +{
>> +    switch (type) {
>> +    case DRM_NODE_PRIMARY:
>> +        return "card";
>> +    case DRM_NODE_CONTROL:
>> +        return "controlD";
>> +    case DRM_NODE_RENDER:
>> +        return "renderD";
>> +    default:
>> +        return NULL;
>> +    }
>> +}
>> +
>>  /**
>>   * Open the device by bus ID.
>>   *
>> @@ -2736,3 +2752,69 @@ int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle)
>>       return 0;
>>  }
>>
>> +static char *drmGetMinorNameForFD(int fd, int type)
>> +{
>> +#ifdef __linux__
>> +     DIR *sysdir;
>> +     struct dirent *pent, *ent;
>> +     struct stat sbuf;
>> +     const char *name = drmGetMinorName(type);
>> +     const int len = strlen(name);
> This will cause a segfault if 'name' is NULL.
>
>> +     char dev_name[64], buf[64];
>> +     long name_max;
>> +     int maj, min;
>> +
>> +     if (!name)
>> +             return NULL;
>> +
>> +     if (fstat(fd, &sbuf))
>> +             return NULL;
>> +
>> +     maj = major(sbuf.st_rdev);
>> +     min = minor(sbuf.st_rdev);
>> +
>> +     if (maj != DRM_MAJOR || !S_ISCHR(sbuf.st_mode))
>> +             return NULL;
>> +
>> +     snprintf(buf, sizeof(buf), "/sys/dev/char/%d:%d/device/drm", maj, min);
>> +
>> +     sysdir = opendir(buf);
>> +     if (!sysdir)
>> +             return NULL;
>> +
>> +     name_max = fpathconf(dirfd(sysdir), _PC_NAME_MAX);
>> +     if (name_max == -1)
>> +             goto out_close_dir;
>> +
>> +     pent = malloc(offsetof(struct dirent, d_name) + name_max + 1);
>> +     if (pent == NULL)
>> +              goto out_close_dir;
>> +
>> +     while (readdir_r(sysdir, pent, &ent) == 0 && ent != NULL) {
>> +             if (strncmp(ent->d_name, name, len) == 0) {
>> +                     free(pent);
>> +                     closedir(sysdir);
>> +
>> +                     snprintf(dev_name, sizeof(dev_name), DRM_DIR_NAME "/%s",
>> +                              ent->d_name);
>> +                     return strdup(dev_name);
>> +             }
>> +     }
>> +
>> +     free(pent);
>> +
>> +out_close_dir:
>> +     closedir(sysdir);
>> +#endif
>> +     return NULL;
>> +}
>> +
>> +char *drmGetMasterNameFromRenderFD(int fd)
> I think drmGetPrimaryDeviceNameFromFd would be more appropriate given
> the node type is 'primary',
Most places that I've seen call "/dev/dri/cardX" master node, although
with the introduction of DRM_NODE_PRIMARY I believe your suggestion
may be better.

> the type of the fd doesn't matter afaics and
> for consistency with other drmGet* functions. However, given that's a
> bit of a mouthful I guess the 'Device' part could be dropped.
>
I was thinking about this initially then decided to keep the
Render/MasterFD explicit. Mainly because I'm don't know how cumbersome
the *BSD implementation will be.

But with that said it none of the BSD guys objects within 1-2 weeks
I'll go with your suggestion.

Thanks
Emil
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH libdrm v3] drm: add drmGet(Master|Render)NameFrom(Render|Master)FD functions

[Emil Velikov]

Currently most places assume reliable master <> render node mapping.
Although this may work in some cases, it is not correct.

Add a couple of helpers that hide the details and provide the name of
the master/render device name, given an render/master FD.

v2:
 - Rename Device and Primary to Master (aka the /dev/dri/cardX device).
 - Check for the file via readdir_r() rather than stat().
 - Wrap the check into a single function.
 - Return NULL for non-linux platforms.

v3:
 - Don't segfault if name is NULL.
 - Update function names, as suggested by Frank Binns.

Cc: Frank Binns <frank.binns@imgtec.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: David Herrmann <dh.herrmann@googlemail.com>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
---

Hi all,

I'm planning to push this Monday afternoon. If you have any comments or 
suggestions please go ahead.

Cheers,
Emil


 xf86drm.c | 84 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 xf86drm.h |  3 +++
 2 files changed, 87 insertions(+)

diff --git a/xf86drm.c b/xf86drm.c
index e117bc6..194cd35 100644
--- a/xf86drm.c
+++ b/xf86drm.c
@@ -40,6 +40,8 @@
 #include <string.h>
 #include <strings.h>
 #include <ctype.h>
+#include <dirent.h>
+#include <stddef.h>
 #include <fcntl.h>
 #include <errno.h>
 #include <signal.h>
@@ -522,6 +524,20 @@ static int drmGetMinorType(int minor)
     }
 }
 
+static const char *drmGetMinorName(int type)
+{
+    switch (type) {
+    case DRM_NODE_PRIMARY:
+        return "card";
+    case DRM_NODE_CONTROL:
+        return "controlD";
+    case DRM_NODE_RENDER:
+        return "renderD";
+    default:
+        return NULL;
+    }
+}
+
 /**
  * Open the device by bus ID.
  *
@@ -2736,3 +2752,71 @@ int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle)
 	return 0;
 }
 
+static char *drmGetMinorNameForFD(int fd, int type)
+{
+#ifdef __linux__
+	DIR *sysdir;
+	struct dirent *pent, *ent;
+	struct stat sbuf;
+	const char *name = drmGetMinorName(type);
+	int len;
+	char dev_name[64], buf[64];
+	long name_max;
+	int maj, min;
+
+	if (!name)
+		return NULL;
+
+	len = strlen(name);
+
+	if (fstat(fd, &sbuf))
+		return NULL;
+
+	maj = major(sbuf.st_rdev);
+	min = minor(sbuf.st_rdev);
+
+	if (maj != DRM_MAJOR || !S_ISCHR(sbuf.st_mode))
+		return NULL;
+
+	snprintf(buf, sizeof(buf), "/sys/dev/char/%d:%d/device/drm", maj, min);
+
+	sysdir = opendir(buf);
+	if (!sysdir)
+		return NULL;
+
+	name_max = fpathconf(dirfd(sysdir), _PC_NAME_MAX);
+	if (name_max == -1)
+		goto out_close_dir;
+
+	pent = malloc(offsetof(struct dirent, d_name) + name_max + 1);
+	if (pent == NULL)
+		 goto out_close_dir;
+
+	while (readdir_r(sysdir, pent, &ent) == 0 && ent != NULL) {
+		if (strncmp(ent->d_name, name, len) == 0) {
+			free(pent);
+			closedir(sysdir);
+
+			snprintf(dev_name, sizeof(dev_name), DRM_DIR_NAME "/%s",
+				 ent->d_name);
+			return strdup(dev_name);
+		}
+	}
+
+	free(pent);
+
+out_close_dir:
+	closedir(sysdir);
+#endif
+	return NULL;
+}
+
+char *drmGetPrimaryDeviceNameFromFd(int fd)
+{
+	return drmGetMinorNameForFD(fd, DRM_NODE_PRIMARY);
+}
+
+char *drmGetRenderDeviceNameFromFd(int fd)
+{
+	return drmGetMinorNameForFD(fd, DRM_NODE_RENDER);
+}
diff --git a/xf86drm.h b/xf86drm.h
index afd38a1..40c55c9 100644
--- a/xf86drm.h
+++ b/xf86drm.h
@@ -749,6 +749,9 @@ extern int drmGetNodeTypeFromFd(int fd);
 extern int drmPrimeHandleToFD(int fd, uint32_t handle, uint32_t flags, int *prime_fd);
 extern int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle);
 
+extern char *drmGetPrimaryDeviceNameFromFd(int fd);
+extern char *drmGetRenderDeviceNameFromFd(int fd);
+
 #if defined(__cplusplus) || defined(c_plusplus)
 }
 #endif
-- 
2.3.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH libdrm v3] drm: add drmGet(Master|Render)NameFrom(Render|Master)FD functions

[Emil Velikov]

On 7 March 2015 at 00:58, Emil Velikov <emil.l.velikov@gmail.com> wrote:
> Currently most places assume reliable master <> render node mapping.
> Although this may work in some cases, it is not correct.
>
> Add a couple of helpers that hide the details and provide the name of
> the master/render device name, given an render/master FD.
>
> v2:
>  - Rename Device and Primary to Master (aka the /dev/dri/cardX device).
>  - Check for the file via readdir_r() rather than stat().
>  - Wrap the check into a single function.
>  - Return NULL for non-linux platforms.
>
> v3:
>  - Don't segfault if name is NULL.
>  - Update function names, as suggested by Frank Binns.
And seemingly I've forgot to update the patch summary to reflect the
change. I'll do so before pushing

-Emil
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH 2/2] libdrm: add drmGetNodeType() helper

[Emil Velikov]

The add a simple helper which returns the node type of the opened fd.
Likely to be used in conjunction with the previous two helpers.

Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: David Herrmann <dh.herrmann@googlemail.com>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
---
 xf86drm.c | 43 +++++++++++++++++++++++++++++++++++++++----
 xf86drm.h |  7 +++++++
 2 files changed, 46 insertions(+), 4 deletions(-)

diff --git a/xf86drm.c b/xf86drm.c
index 6af7ac0..a70c0dd 100644
--- a/xf86drm.c
+++ b/xf86drm.c
@@ -85,10 +85,6 @@
 
 #define DRM_MSG_VERBOSITY 3
 
-#define DRM_NODE_CONTROL 0
-#define DRM_NODE_PRIMARY 1
-#define DRM_NODE_RENDER 2
-
 static drmServerInfoPtr drm_server_info;
 
 void drmSetServerInfo(drmServerInfoPtr info)
@@ -2607,6 +2603,45 @@ int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle)
 	return 0;
 }
 
+int drmGetNodeType(int fd, drmNodeType *type)
+{
+	struct stat sbuf;
+	char name[64];
+	dev_t d;
+	int i;
+
+	if (!fstat(fd, &sbuf))
+		return -errno;
+
+	d = sbuf.st_rdev;
+
+	for (i = 0; i < DRM_MAX_MINOR; i++) {
+		snprintf(name, sizeof(name), DRM_DEV_NAME, DRM_DIR_NAME, i);
+		if (stat(name, &sbuf) == 0 && sbuf.st_rdev == d) {
+			*type = DRM_NODE_PRIMARY;
+			return 0;
+		}
+	}
+
+	for (i = 64; i < (64 + DRM_MAX_MINOR); i++) {
+		snprintf(name, sizeof(name), DRM_CONTROL_DEV_NAME, DRM_DIR_NAME, i);
+		if (stat(name, &sbuf) == 0 && sbuf.st_rdev == d) {
+			*type = DRM_NODE_CONTROL;
+			return 0;
+		}
+	}
+
+	for (i = 128; i < (128 + DRM_MAX_MINOR); i++) {
+		snprintf(name, sizeof(name), DRM_RENDER_DEV_NAME, DRM_DIR_NAME, i);
+		if (stat(name, &sbuf) == 0 && sbuf.st_rdev == d) {
+			*type = DRM_NODE_RENDER;
+			return 0;
+		}
+	}
+
+	return -EINVAL;
+}
+
 char *drmGetDeviceNameFromRenderFD(int fd)
 {
 
diff --git a/xf86drm.h b/xf86drm.h
index bca5887..7d67df9 100644
--- a/xf86drm.h
+++ b/xf86drm.h
@@ -740,6 +740,13 @@ extern char *drmGetDeviceNameFromFd(int fd);
 extern int drmPrimeHandleToFD(int fd, uint32_t handle, uint32_t flags, int *prime_fd);
 extern int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle);
 
+typedef enum _drmNodeType {
+    DRM_NODE_CONTROL = 0,
+    DRM_NODE_PRIMARY = 1,
+    DRM_NODE_RENDER  = 2
+} drmNodeType, *drmNodeTypePtr;
+
+extern int drmGetNodeType(int fd, drmNodeType *type);
 extern char *drmGetRenderNameFromDeviceFD(int fd);
 extern char *drmGetDeviceNameFromRenderFD(int fd);
 
-- 
2.2.2

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH 2/2] libdrm: add drmGetNodeType() helper

[Frank Binns]

Hi Emil,

On 02/02/15 00:14, Emil Velikov wrote:
> The add a simple helper which returns the node type of the opened fd.
> Likely to be used in conjunction with the previous two helpers.
>
> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> Cc: David Herrmann <dh.herrmann@googlemail.com>
> Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
> ---
>  xf86drm.c | 43 +++++++++++++++++++++++++++++++++++++++----
>  xf86drm.h |  7 +++++++
>  2 files changed, 46 insertions(+), 4 deletions(-)
>
> diff --git a/xf86drm.c b/xf86drm.c
> index 6af7ac0..a70c0dd 100644
> --- a/xf86drm.c
> +++ b/xf86drm.c
> @@ -85,10 +85,6 @@
>  
>  #define DRM_MSG_VERBOSITY 3
>  
> -#define DRM_NODE_CONTROL 0
> -#define DRM_NODE_PRIMARY 1
> -#define DRM_NODE_RENDER 2
> -
>  static drmServerInfoPtr drm_server_info;
>  
>  void drmSetServerInfo(drmServerInfoPtr info)
> @@ -2607,6 +2603,45 @@ int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle)
>  	return 0;
>  }
>  
> +int drmGetNodeType(int fd, drmNodeType *type)
> +{
> +	struct stat sbuf;
> +	char name[64];
> +	dev_t d;
> +	int i;
> +
> +	if (!fstat(fd, &sbuf))
> +		return -errno;
> +
> +	d = sbuf.st_rdev;
> +
> +	for (i = 0; i < DRM_MAX_MINOR; i++) {
> +		snprintf(name, sizeof(name), DRM_DEV_NAME, DRM_DIR_NAME, i);
> +		if (stat(name, &sbuf) == 0 && sbuf.st_rdev == d) {
> +			*type = DRM_NODE_PRIMARY;
> +			return 0;
> +		}
> +	}
> +
> +	for (i = 64; i < (64 + DRM_MAX_MINOR); i++) {
> +		snprintf(name, sizeof(name), DRM_CONTROL_DEV_NAME, DRM_DIR_NAME, i);
> +		if (stat(name, &sbuf) == 0 && sbuf.st_rdev == d) {
> +			*type = DRM_NODE_CONTROL;
> +			return 0;
> +		}
> +	}
> +
> +	for (i = 128; i < (128 + DRM_MAX_MINOR); i++) {
> +		snprintf(name, sizeof(name), DRM_RENDER_DEV_NAME, DRM_DIR_NAME, i);
> +		if (stat(name, &sbuf) == 0 && sbuf.st_rdev == d) {
> +			*type = DRM_NODE_RENDER;
> +			return 0;
> +		}
> +	}
It would seem easier (and faster) to use minor(sbuf.st_rdev) to
determine the type of node.

I'm not sure what the etiquette is here but I've prepared a patch that
does it this way instead, which will follow this email. Please feel free
to add your own signed-off-by, suggested-by, etc as I don't want to
steal all your credit :)

Thanks
Frank

> +
> +	return -EINVAL;
> +}
> +
>  char *drmGetDeviceNameFromRenderFD(int fd)
>  {
>  
> diff --git a/xf86drm.h b/xf86drm.h
> index bca5887..7d67df9 100644
> --- a/xf86drm.h
> +++ b/xf86drm.h
> @@ -740,6 +740,13 @@ extern char *drmGetDeviceNameFromFd(int fd);
>  extern int drmPrimeHandleToFD(int fd, uint32_t handle, uint32_t flags, int *prime_fd);
>  extern int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle);
>  
> +typedef enum _drmNodeType {
> +    DRM_NODE_CONTROL = 0,
> +    DRM_NODE_PRIMARY = 1,
> +    DRM_NODE_RENDER  = 2
> +} drmNodeType, *drmNodeTypePtr;
> +
> +extern int drmGetNodeType(int fd, drmNodeType *type);
>  extern char *drmGetRenderNameFromDeviceFD(int fd);
>  extern char *drmGetDeviceNameFromRenderFD(int fd);
>  

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [RFC PATCH 0/2] libdrm: Add master <> render node helpers

[Emil Velikov]

On 02/02/15 00:14, Emil Velikov wrote:
> Hi all,
> 
> As mentioned a couple of days ago at #dri-devel some(most) users of 
> render nodes tend to rely on strict mapping between the primary and 
> render node. I.e. something along the lines of
> 
>   fstat(render_fd, &sbuf);
>   sprintf(primary_node, "/dev/dri/card%d",
>                         ((sbuf.st_rdev & 0x3f) | 0x80));
> 
> Currently the following are (ab)using the above code:
>  - xf86-video-nouveau
>  - xf86-video-intel
>  - libva (vaapi)
> 
> As reminded by David Herrman, this is not the correct solution - thus 
> I've added a couple of helpers which walk through sysfs of the 
> respecitive device and return the correct device name.
> 
> I'm not 100% happy with the function names, so suggestions are greatly 
> appreciated. Any other comments are also welcome :)
> 
> Note: BSD guys - you'll likely need your own version of these functions.
> 
David, Daniel

Can you please take a look at these two patches. Would be great if we
can minimize the above assumptions before they get too wide spread.

*BSD guys,

Please take a look and/or forward this to the relevant people. I fear
that my *BSD knowledge is virtually none :\

Cheers,
Emil
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [RFC PATCH 0/2] libdrm: Add master <> render node helpers

[Jonathan Gray]

On Tue, Feb 10, 2015 at 10:37:16PM +0000, Emil Velikov wrote:
> On 02/02/15 00:14, Emil Velikov wrote:
> > Hi all,
> > 
> > As mentioned a couple of days ago at #dri-devel some(most) users of 
> > render nodes tend to rely on strict mapping between the primary and 
> > render node. I.e. something along the lines of
> > 
> >   fstat(render_fd, &sbuf);
> >   sprintf(primary_node, "/dev/dri/card%d",
> >                         ((sbuf.st_rdev & 0x3f) | 0x80));
> > 
> > Currently the following are (ab)using the above code:
> >  - xf86-video-nouveau
> >  - xf86-video-intel
> >  - libva (vaapi)
> > 
> > As reminded by David Herrman, this is not the correct solution - thus 
> > I've added a couple of helpers which walk through sysfs of the 
> > respecitive device and return the correct device name.
> > 
> > I'm not 100% happy with the function names, so suggestions are greatly 
> > appreciated. Any other comments are also welcome :)
> > 
> > Note: BSD guys - you'll likely need your own version of these functions.
> > 
> David, Daniel
> 
> Can you please take a look at these two patches. Would be great if we
> can minimize the above assumptions before they get too wide spread.
> 
> *BSD guys,
> 
> Please take a look and/or forward this to the relevant people. I fear
> that my *BSD knowledge is virtually none :\
> 
> Cheers,
> Emil

I can't speak for other projects but OpenBSD drm is currently based
on Linux 3.8 with no render nodes or prime.

There is no sysfs/udev etc so when/if render nodes work it will
likely involve either a fixed mapping or ioctls.
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [RFC PATCH 0/2] libdrm: Add master <> render node helpers

[David Herrmann]

Hi

On Tue, Feb 10, 2015 at 11:37 PM, Emil Velikov <emil.l.velikov@gmail.com> wrote:
> On 02/02/15 00:14, Emil Velikov wrote:
>> Hi all,
>>
>> As mentioned a couple of days ago at #dri-devel some(most) users of
>> render nodes tend to rely on strict mapping between the primary and
>> render node. I.e. something along the lines of
>>
>>   fstat(render_fd, &sbuf);
>>   sprintf(primary_node, "/dev/dri/card%d",
>>                         ((sbuf.st_rdev & 0x3f) | 0x80));
>>
>> Currently the following are (ab)using the above code:
>>  - xf86-video-nouveau
>>  - xf86-video-intel
>>  - libva (vaapi)
>>
>> As reminded by David Herrman, this is not the correct solution - thus
>> I've added a couple of helpers which walk through sysfs of the
>> respecitive device and return the correct device name.
>>
>> I'm not 100% happy with the function names, so suggestions are greatly
>> appreciated. Any other comments are also welcome :)
>>
>> Note: BSD guys - you'll likely need your own version of these functions.
>>
> David, Daniel
>
> Can you please take a look at these two patches. Would be great if we
> can minimize the above assumptions before they get too wide spread.

The patches look ok. I really dislike the direct file-probing, though.
I'd recommend using libudev or readdir_r() on /dev/dri/, but I'm not
the one to ask about coding-style for libdrm, so fine with me.

Thanks
David
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [RFC PATCH 0/2] libdrm: Add master <> render node helpers

[Emil Velikov]

On 12/02/15 18:21, David Herrmann wrote:
> Hi
> 
> On Tue, Feb 10, 2015 at 11:37 PM, Emil Velikov <emil.l.velikov@gmail.com> wrote:
>> On 02/02/15 00:14, Emil Velikov wrote:
>>> Hi all,
>>>
>>> As mentioned a couple of days ago at #dri-devel some(most) users of
>>> render nodes tend to rely on strict mapping between the primary and
>>> render node. I.e. something along the lines of
>>>
>>>   fstat(render_fd, &sbuf);
>>>   sprintf(primary_node, "/dev/dri/card%d",
>>>                         ((sbuf.st_rdev & 0x3f) | 0x80));
>>>
>>> Currently the following are (ab)using the above code:
>>>  - xf86-video-nouveau
>>>  - xf86-video-intel
>>>  - libva (vaapi)
>>>
>>> As reminded by David Herrman, this is not the correct solution - thus
>>> I've added a couple of helpers which walk through sysfs of the
>>> respecitive device and return the correct device name.
>>>
>>> I'm not 100% happy with the function names, so suggestions are greatly
>>> appreciated. Any other comments are also welcome :)
>>>
>>> Note: BSD guys - you'll likely need your own version of these functions.
>>>
>> David, Daniel
>>
>> Can you please take a look at these two patches. Would be great if we
>> can minimize the above assumptions before they get too wide spread.
> 
> The patches look ok. I really dislike the direct file-probing, though.
> I'd recommend using libudev or readdir_r() on /dev/dri/, but I'm not
> the one to ask about coding-style for libdrm, so fine with me.
> 
Thanks for having a look. libudev is no used in libdrm so I will stay
away for now. I would prefer to use readdir on /dev/dri/ but AFAICT
there is no way to get the mapping from those alone. Am I missing
something ?

I'm also thinking that changing the prototypes, as below, might be
useful - i.e. return 0 on success, -ENODEV when the corresponding device
lacks render/primary device, and other negative value on error ?

char *drmGetRenderNameFromDeviceFD(int fd);
char *drmGetDeviceNameFromRenderFD(int fd);

int drmGetRenderNameFromDeviceFD(int fd, char **render);
int drmGetDeviceNameFromRenderFD(int fd, char **device);


Cheers,
Emil

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [RFC PATCH 0/2] libdrm: Add master <> render node helpers

[Frank Binns]

Hi Emil,

On 13/02/15 09:18, Emil Velikov wrote:
> On 12/02/15 18:21, David Herrmann wrote:
>> Hi
>>
>> On Tue, Feb 10, 2015 at 11:37 PM, Emil Velikov <emil.l.velikov@gmail.com> wrote:
>>> On 02/02/15 00:14, Emil Velikov wrote:
>>>> Hi all,
>>>>
>>>> As mentioned a couple of days ago at #dri-devel some(most) users of
>>>> render nodes tend to rely on strict mapping between the primary and
>>>> render node. I.e. something along the lines of
>>>>
>>>>   fstat(render_fd, &sbuf);
>>>>   sprintf(primary_node, "/dev/dri/card%d",
>>>>                         ((sbuf.st_rdev & 0x3f) | 0x80));
>>>>
>>>> Currently the following are (ab)using the above code:
>>>>  - xf86-video-nouveau
>>>>  - xf86-video-intel
>>>>  - libva (vaapi)
>>>>
>>>> As reminded by David Herrman, this is not the correct solution - thus
>>>> I've added a couple of helpers which walk through sysfs of the
>>>> respecitive device and return the correct device name.
>>>>
>>>> I'm not 100% happy with the function names, so suggestions are greatly
>>>> appreciated. Any other comments are also welcome :)
>>>>
>>>> Note: BSD guys - you'll likely need your own version of these functions.
>>>>
>>> David, Daniel
>>>
>>> Can you please take a look at these two patches. Would be great if we
>>> can minimize the above assumptions before they get too wide spread.
>> The patches look ok. I really dislike the direct file-probing, though.
>> I'd recommend using libudev or readdir_r() on /dev/dri/, but I'm not
>> the one to ask about coding-style for libdrm, so fine with me.
>>
> Thanks for having a look. libudev is no used in libdrm so I will stay
> away for now. I would prefer to use readdir on /dev/dri/ but AFAICT
> there is no way to get the mapping from those alone. Am I missing
> something ?
Can you you not just use use readdir on /dev/dri and stat on each
entry's d_name comparing the result to an fstat on the fd? Once you find
a match you just
strdup the d_name.

>
> I'm also thinking that changing the prototypes, as below, might be
> useful - i.e. return 0 on success, -ENODEV when the corresponding device
> lacks render/primary device, and other negative value on error ?
>
> char *drmGetRenderNameFromDeviceFD(int fd);
> char *drmGetDeviceNameFromRenderFD(int fd);
I personally prefer this as it fits in better with what we already have.

>
> int drmGetRenderNameFromDeviceFD(int fd, char **render);
> int drmGetDeviceNameFromRenderFD(int fd, char **device);
>
>
> Cheers,
> Emil
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/dri-devel

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [RFC PATCH 0/2] libdrm: Add master <> render node helpers

[Emil Velikov]

On 13/02/15 11:18, Frank Binns wrote:
> Hi Emil,
> 
> On 13/02/15 09:18, Emil Velikov wrote:
>> On 12/02/15 18:21, David Herrmann wrote:
>>> Hi
>>>
>>> On Tue, Feb 10, 2015 at 11:37 PM, Emil Velikov <emil.l.velikov@gmail.com> wrote:
>>>> On 02/02/15 00:14, Emil Velikov wrote:
>>>>> Hi all,
>>>>>
>>>>> As mentioned a couple of days ago at #dri-devel some(most) users of
>>>>> render nodes tend to rely on strict mapping between the primary and
>>>>> render node. I.e. something along the lines of
>>>>>
>>>>>   fstat(render_fd, &sbuf);
>>>>>   sprintf(primary_node, "/dev/dri/card%d",
>>>>>                         ((sbuf.st_rdev & 0x3f) | 0x80));
>>>>>
>>>>> Currently the following are (ab)using the above code:
>>>>>  - xf86-video-nouveau
>>>>>  - xf86-video-intel
>>>>>  - libva (vaapi)
>>>>>
>>>>> As reminded by David Herrman, this is not the correct solution - thus
>>>>> I've added a couple of helpers which walk through sysfs of the
>>>>> respecitive device and return the correct device name.
>>>>>
>>>>> I'm not 100% happy with the function names, so suggestions are greatly
>>>>> appreciated. Any other comments are also welcome :)
>>>>>
>>>>> Note: BSD guys - you'll likely need your own version of these functions.
>>>>>
>>>> David, Daniel
>>>>
>>>> Can you please take a look at these two patches. Would be great if we
>>>> can minimize the above assumptions before they get too wide spread.
>>> The patches look ok. I really dislike the direct file-probing, though.
>>> I'd recommend using libudev or readdir_r() on /dev/dri/, but I'm not
>>> the one to ask about coding-style for libdrm, so fine with me.
>>>
>> Thanks for having a look. libudev is no used in libdrm so I will stay
>> away for now. I would prefer to use readdir on /dev/dri/ but AFAICT
>> there is no way to get the mapping from those alone. Am I missing
>> something ?
> Can you you not just use use readdir on /dev/dri and stat on each
> entry's d_name comparing the result to an fstat on the fd?
Unless I've misunderstood something, a primary device will have
226:{0-63}, while render device 226:{128-191}.

One cannot just mask out the 7th bit, as this is the (wrong) solution
what I've mentioned above. Or maybe I'm missing something ?
Afaict the information/"link" between the two is present at the drm
module level - thus the ugly sysfs dance.

> Once you find a match you just strdup the d_name.
> 
>>
>> I'm also thinking that changing the prototypes, as below, might be
>> useful - i.e. return 0 on success, -ENODEV when the corresponding device
>> lacks render/primary device, and other negative value on error ?
>>
>> char *drmGetRenderNameFromDeviceFD(int fd);
>> char *drmGetDeviceNameFromRenderFD(int fd);
> I personally prefer this as it fits in better with what we already have.
> 
The recent inspiration came from the prime helpers. I don't feel
strongly either way.

Cheers,
Emil
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel