diff for duplicates of <4E18ADEB.3000904@freescale.com> diff --git a/a/1.txt b/N1/1.txt index e8ac534..4386a9b 100644 --- a/a/1.txt +++ b/N1/1.txt @@ -1,5 +1,6 @@ Anatolij Gustschin wrote: ->> > ad->pix_fmt >> > - diu_ops.get_pixel_format(var->bits_per_pixel, +>> > ad->pix_fmt =3D +>> > - diu_ops.get_pixel_format(var->bits_per_pixel, >> > - machine_data->monitor_port); >> > + cpu_to_be32(diu_ops.get_pixel_format(machine_data->monitor_port, >> > + var->bits_per_pixel)); @@ -23,23 +24,24 @@ struct diu_ad { __be32 pix_fmt; /* hard coding pixel format */ -I see other places where the endianness of pix_fmt is assumed. I'll have +I see other places where the endianness of pix_fmt is assumed. I'll have=20 to fix those later. -When I look at the 5121 and 1022 reference manuals, it says that the area -descriptor is little-endian. I think York made this register big-endian -so that it matches the endianness of get_pixel_format(), which doesn't -make much sense. I'll have to fix the whole endianness thing with pix_fmt +When I look at the 5121 and 1022 reference manuals, it says that the area=20 +descriptor is little-endian. I think York made this register big-endian=20 +so that it matches the endianness of get_pixel_format(), which doesn't=20 +make much sense. I'll have to fix the whole endianness thing with pix_fmt= +=20 in another patch. > get_pixel_format() returns LE32 value. -I don't think the endianness of get_pixel_format() is fixed. It's +I don't think the endianness of get_pixel_format() is fixed. It's=20 whatever the CPU native endiannes is. That's why I added a cpu_to_be32(). -Anyway, I'm not quite sure what the right answer is, but it seems obvious +Anyway, I'm not quite sure what the right answer is, but it seems obvious=20 that cpu_to_be32() is wrong. I'll post another patch which removes it. --- +--=20 Timur Tabi -Linux kernel developer at Freescale +Linux kernel developer at Freescale= diff --git a/a/content_digest b/N1/content_digest index 7c9099c..5a91b68 100644 --- a/a/content_digest +++ b/N1/content_digest @@ -1,8 +1,8 @@ "ref\01310057466-14599-1-git-send-email-timur@freescale.com\0" "ref\020110709163809.53f18fea@wker\0" "From\0Tabi Timur-B04825 <B04825@freescale.com>\0" - "Subject\0Re: [PATCH] drivers/video: use strings to specify the Freescale DIU\0" - "Date\0Sat, 09 Jul 2011 19:37:16 +0000\0" + "Subject\0Re: [PATCH] drivers/video: use strings to specify the Freescale DIU monitor port\0" + "Date\0Sat, 9 Jul 2011 19:37:16 +0000\0" "To\0Anatolij Gustschin <agust@denx.de>\0" "Cc\0linuxppc-dev@ozlabs.org <linuxppc-dev@ozlabs.org>" Gala Kumar-B11780 <B11780@freescale.com> @@ -12,7 +12,8 @@ "\00:1\0" "b\0" "Anatolij Gustschin wrote:\n" - ">> > \tad->pix_fmt >> > -\t\tdiu_ops.get_pixel_format(var->bits_per_pixel,\n" + ">> > \tad->pix_fmt =3D\n" + ">> > -\t\tdiu_ops.get_pixel_format(var->bits_per_pixel,\n" ">> > -\t\t\t\t\t machine_data->monitor_port);\n" ">> > +\t\tcpu_to_be32(diu_ops.get_pixel_format(machine_data->monitor_port,\n" ">> > +\t\t\t\t\t\t var->bits_per_pixel));\n" @@ -36,25 +37,26 @@ "\n" "\t__be32 pix_fmt; /* hard coding pixel format */\n" "\n" - "I see other places where the endianness of pix_fmt is assumed. I'll have \n" + "I see other places where the endianness of pix_fmt is assumed. I'll have=20\n" "to fix those later.\n" "\n" - "When I look at the 5121 and 1022 reference manuals, it says that the area \n" - "descriptor is little-endian. I think York made this register big-endian \n" - "so that it matches the endianness of get_pixel_format(), which doesn't \n" - "make much sense. I'll have to fix the whole endianness thing with pix_fmt \n" + "When I look at the 5121 and 1022 reference manuals, it says that the area=20\n" + "descriptor is little-endian. I think York made this register big-endian=20\n" + "so that it matches the endianness of get_pixel_format(), which doesn't=20\n" + "make much sense. I'll have to fix the whole endianness thing with pix_fmt=\n" + "=20\n" "in another patch.\n" "\n" " > get_pixel_format() returns LE32 value.\n" "\n" - "I don't think the endianness of get_pixel_format() is fixed. It's \n" + "I don't think the endianness of get_pixel_format() is fixed. It's=20\n" "whatever the CPU native endiannes is. That's why I added a cpu_to_be32().\n" "\n" - "Anyway, I'm not quite sure what the right answer is, but it seems obvious \n" + "Anyway, I'm not quite sure what the right answer is, but it seems obvious=20\n" "that cpu_to_be32() is wrong. I'll post another patch which removes it.\n" "\n" - "-- \n" + "--=20\n" "Timur Tabi\n" - Linux kernel developer at Freescale + Linux kernel developer at Freescale= -708388f36e1fefe96246db70bf50cb2130099e8c80583593d7420e256d08456a +34521b2d6e8176a06c3d5f7620ff0e9b01a1d74276e8bb43cd2ac83512c9ca47
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.