BTW: In response to your second question, the same rule should also apply to sub-surfaces which really map to the same underlying memory.
>-----Original Message----- >From: Strelchun, Timothy >Sent: Tuesday, September 01, 2009 4:43 PM >To: 'directfb-dev@directfb.org' >Subject: RE: [directfb-dev] question about locking > >Hi Anders, > >It is not a safe operation. The DFB core software and drivers >may depend upon knowing when CPU based pixel modifications are >complete for a variety of reasons (such as cache flushing for >the GPU or pushing the buffer to video RAM, etc.). Always >perform locking and unlocking has a paired operation. > >We just identified a fix for a customer today related to >visible caching artifacts, partially because lock/unlock were >not used properly (the buffer pointer was being cached). > >Regards, >Timothy > >-- > >Timothy Strelchun >CE Software Engineering >Digital Home Group >Intel Corporation > >The views expressed above are my own and not those of Intel > >>-----Original Message----- >>From: directfb-dev-boun...@directfb.org >>[mailto:directfb-dev-boun...@directfb.org] On Behalf Of Anders Bakken >>Sent: Tuesday, September 01, 2009 3:40 PM >>To: directfb-dev@directfb.org >>Subject: [directfb-dev] question about locking >> >>Is the following a "safe" operation: >> >>IDirectFBSurface *surfaceOne = ...; >>IDirectFBSurface *surfaceTwo = ...; >>int bpl; >>void *mem; >>surfaceOne->Lock(surfaceOne, DSLF_WRITE, &mem, &bpl); >>// change the pixels of mem using something other interface >>surfaceTwo->Blit(surfaceTwo, surfaceOne, 0, 0, 0); >> >>Do I need to call surfaceOne->Unlock(surfaceOne); before I blit >>surfaceOne onto surfaceTwo? >> >>If the answer is yes, what if I did this instead: >>IDirectFBSurface *surfaceOne = ...; >>IDirectFBSurface *surfaceTwo = ...; >>IDirectFBSurface *sub; >>surfaceOne->GetSubSurface(surfaceOne, 0, &sub); >>int bpl; >>void *mem; >>sub->Lock(sub, DSLF_WRITE, &mem, &bpl); >>// change the pixels of mem using something other interface >>surfaceTwo->Blit(surfaceTwo, surfaceOne, 0, 0, 0); >> >>Do I need to call sub->Unlock(sub); before I blit surfaceOne onto >>surfaceTwo? >> >>regards >>-- >>Anders Bakken _______________________________________________ directfb-dev mailing list directfb-dev@directfb.org http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
