gsdx: fix regression on 24 bit format (gust game)

bpp is the size of the container format but not the size of the format itself.
For PSM24, bpp is 32, trbpp is 24
This commit is contained in:
Gregory Hainaut 2016-05-06 13:47:53 +02:00
parent c94fecdadb
commit f34e77f08c
1 changed files with 3 additions and 3 deletions

View File

@ -385,7 +385,7 @@ GSTextureCache::Target* GSTextureCache::LookupTarget(const GIFRegTEX0& TEX0, int
dst->Update();
dst->m_dirty_alpha |= (psm_s.bpp != 24);
dst->m_dirty_alpha |= (psm_s.trbpp != 24);
} else if (CanConvertDepth()) {
@ -1725,8 +1725,8 @@ GSTextureCache::Target::Target(GSRenderer* r, const GIFRegTEX0& TEX0, uint8* tem
, m_depth_supported(depth_supported)
{
m_TEX0 = TEX0;
m_32_bits_fmt |= (GSLocalMemory::m_psm[TEX0.PSM].bpp != 16);
m_dirty_alpha = (TEX0.PSM != PSM_PSMCT24) && (TEX0.PSM != PSM_PSMZ24);
m_32_bits_fmt |= (GSLocalMemory::m_psm[TEX0.PSM].trbpp != 16);
m_dirty_alpha = GSLocalMemory::m_psm[TEX0.PSM].trbpp != 24;
m_valid = GSVector4i::zero();
}