* [PATCH] libv4lconvert: We support more then 32 bit src fmts now, so use 64 bit bitmasks
@ 2017-11-02 20:22 Hans de Goede
0 siblings, 0 replies; only message in thread
From: Hans de Goede @ 2017-11-02 20:22 UTC (permalink / raw)
To: Gregor Jasny; +Cc: Linux Media Mailing List, Hans de Goede
We support more then 32 bit src fmts now, so we can no longer re-use
struct v4l2_frmsizeenum.pixel_format to store a bitmask of all the
supported src-formats for a given frame-size.
This fixes a subtile bug where we would try to use SE401 as src fmt
instead of YUYV under certain circumstances.
BugLink: https://bugzilla.redhat.com/show_bug.cgi?id=1508706
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
lib/libv4lconvert/libv4lconvert-priv.h | 2 ++
lib/libv4lconvert/libv4lconvert.c | 9 ++++-----
2 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/lib/libv4lconvert/libv4lconvert-priv.h b/lib/libv4lconvert/libv4lconvert-priv.h
index e2389347..9a467e10 100644
--- a/lib/libv4lconvert/libv4lconvert-priv.h
+++ b/lib/libv4lconvert/libv4lconvert-priv.h
@@ -66,6 +66,8 @@ struct v4lconvert_data {
int cinfo_initialized;
#endif // HAVE_JPEG
struct v4l2_frmsizeenum framesizes[V4LCONVERT_MAX_FRAMESIZES];
+ /* Bitmask of all supported src_formats which can do for a size */
+ int64_t framesize_supported_src_formats[V4LCONVERT_MAX_FRAMESIZES];
unsigned int no_framesizes;
int bandwidth;
int fps;
diff --git a/lib/libv4lconvert/libv4lconvert.c b/lib/libv4lconvert/libv4lconvert.c
index 1a5ccec2..d666bd97 100644
--- a/lib/libv4lconvert/libv4lconvert.c
+++ b/lib/libv4lconvert/libv4lconvert.c
@@ -434,7 +434,8 @@ static int v4lconvert_do_try_format_uvc(struct v4lconvert_data *data,
for (i = 0; i < ARRAY_SIZE(supported_src_pixfmts); i++) {
/* is this format supported? */
- if (!(data->framesizes[best_framesize].pixel_format & (1 << i)))
+ if (!(data->framesize_supported_src_formats[best_framesize] &
+ (1ULL << i)))
continue;
/* Note the hardcoded use of discrete is based on this function
@@ -1647,9 +1648,7 @@ static void v4lconvert_get_framesizes(struct v4lconvert_data *data,
return;
}
data->framesizes[data->no_framesizes].type = frmsize.type;
- /* We use the pixel_format member to store a bitmask of all
- supported src_formats which can do this size */
- data->framesizes[data->no_framesizes].pixel_format = 1 << index;
+ data->framesize_supported_src_formats[data->no_framesizes] = 1ULL << index;
switch (frmsize.type) {
case V4L2_FRMSIZE_TYPE_DISCRETE:
@@ -1662,7 +1661,7 @@ static void v4lconvert_get_framesizes(struct v4lconvert_data *data,
}
data->no_framesizes++;
} else {
- data->framesizes[j].pixel_format |= 1 << index;
+ data->framesize_supported_src_formats[j] |= 1ULL << index;
}
}
}
--
2.14.3
^ permalink raw reply related [flat|nested] only message in thread
only message in thread, other threads:[~2017-11-02 20:22 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2017-11-02 20:22 [PATCH] libv4lconvert: We support more then 32 bit src fmts now, so use 64 bit bitmasks Hans de Goede
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).