| Message ID | alpine.LRH.2.02.1808251549550.8284@file01.intranet.prod.int.rdu2.redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [1/3] mach64: fix display corruption on big endian machines | expand |
On Sat, Aug 25, 2018 at 03:51:00PM -0400, Mikulas Patocka wrote: > The code for manual bit triple is not endian-clean. It builds the variable > "hostdword" using byte accesses, therefore we must read the variable with > "le32_to_cpu". > > The patch also enables (hardware or software) bit triple only if the image > is monochrome (image->depth). If we want to blit full-color image, we > shouldn't use the triple code. Makes sense. Reviewed-by: Ville Syrjälä <syrjala@sci.fi> > > Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> > Cc: stable@vger.kernel.org > > --- > drivers/video/fbdev/aty/mach64_accel.c | 6 +++--- > 1 file changed, 3 insertions(+), 3 deletions(-) > > Index: linux-stable/drivers/video/fbdev/aty/mach64_accel.c > =================================================================== > --- linux-stable.orig/drivers/video/fbdev/aty/mach64_accel.c 2018-08-24 17:31:21.000000000 +0200 > +++ linux-stable/drivers/video/fbdev/aty/mach64_accel.c 2018-08-24 19:12:40.000000000 +0200 > @@ -345,7 +345,7 @@ void atyfb_imageblit(struct fb_info *inf > * since Rage 3D IIc we have DP_HOST_TRIPLE_EN bit > * this hwaccelerated triple has an issue with not aligned data > */ > - if (M64_HAS(HW_TRIPLE) && image->width % 8 == 0) > + if (image->depth == 1 && M64_HAS(HW_TRIPLE) && image->width % 8 == 0) > pix_width |= DP_HOST_TRIPLE_EN; > } > > @@ -382,7 +382,7 @@ void atyfb_imageblit(struct fb_info *inf > src_bytes = (((image->width * image->depth) + 7) / 8) * image->height; > > /* manual triple each pixel */ > - if (info->var.bits_per_pixel == 24 && !(pix_width & DP_HOST_TRIPLE_EN)) { > + if (image->depth == 1 && info->var.bits_per_pixel == 24 && !(pix_width & DP_HOST_TRIPLE_EN)) { > int inbit, outbit, mult24, byte_id_in_dword, width; > u8 *pbitmapin = (u8*)image->data, *pbitmapout; > u32 hostdword; > @@ -415,7 +415,7 @@ void atyfb_imageblit(struct fb_info *inf > } > } > wait_for_fifo(1, par); > - aty_st_le32(HOST_DATA0, hostdword, par); > + aty_st_le32(HOST_DATA0, le32_to_cpu(hostdword), par); > } > } else { > u32 *pbitmap, dwords = (src_bytes + 3) / 4;
On 08/27/2018 02:54 PM, Ville Syrjälä wrote: > On Sat, Aug 25, 2018 at 03:51:00PM -0400, Mikulas Patocka wrote: >> The code for manual bit triple is not endian-clean. It builds the variable >> "hostdword" using byte accesses, therefore we must read the variable with >> "le32_to_cpu". >> >> The patch also enables (hardware or software) bit triple only if the image >> is monochrome (image->depth). If we want to blit full-color image, we >> shouldn't use the triple code. > > Makes sense. > Reviewed-by: Ville Syrjälä <syrjala@sci.fi> Patch queued for 4.20, thanks. Best regards, -- Bartlomiej Zolnierkiewicz Samsung R&D Institute Poland Samsung Electronics
Index: linux-stable/drivers/video/fbdev/aty/mach64_accel.c =================================================================== --- linux-stable.orig/drivers/video/fbdev/aty/mach64_accel.c 2018-08-24 17:31:21.000000000 +0200 +++ linux-stable/drivers/video/fbdev/aty/mach64_accel.c 2018-08-24 19:12:40.000000000 +0200 @@ -345,7 +345,7 @@ void atyfb_imageblit(struct fb_info *inf * since Rage 3D IIc we have DP_HOST_TRIPLE_EN bit * this hwaccelerated triple has an issue with not aligned data */ - if (M64_HAS(HW_TRIPLE) && image->width % 8 == 0) + if (image->depth == 1 && M64_HAS(HW_TRIPLE) && image->width % 8 == 0) pix_width |= DP_HOST_TRIPLE_EN; } @@ -382,7 +382,7 @@ void atyfb_imageblit(struct fb_info *inf src_bytes = (((image->width * image->depth) + 7) / 8) * image->height; /* manual triple each pixel */ - if (info->var.bits_per_pixel == 24 && !(pix_width & DP_HOST_TRIPLE_EN)) { + if (image->depth == 1 && info->var.bits_per_pixel == 24 && !(pix_width & DP_HOST_TRIPLE_EN)) { int inbit, outbit, mult24, byte_id_in_dword, width; u8 *pbitmapin = (u8*)image->data, *pbitmapout; u32 hostdword; @@ -415,7 +415,7 @@ void atyfb_imageblit(struct fb_info *inf } } wait_for_fifo(1, par); - aty_st_le32(HOST_DATA0, hostdword, par); + aty_st_le32(HOST_DATA0, le32_to_cpu(hostdword), par); } } else { u32 *pbitmap, dwords = (src_bytes + 3) / 4;
The code for manual bit triple is not endian-clean. It builds the variable "hostdword" using byte accesses, therefore we must read the variable with "le32_to_cpu". The patch also enables (hardware or software) bit triple only if the image is monochrome (image->depth). If we want to blit full-color image, we shouldn't use the triple code. Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Cc: stable@vger.kernel.org --- drivers/video/fbdev/aty/mach64_accel.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)