ChanServ changed the topic of #etnaviv to: #etnaviv - the home of the reverse-engineered Vivante GPU driver - Logs https://oftc.irclog.whitequark.org/etnaviv
mvlad has joined #etnaviv
pcercuei has joined #etnaviv
<tomeu> dumping this here in case somebody can save me some digging:
<tomeu> - 0x284d2200, /* [01028] PS.INST_ADDR := ADDR_12_A */
<tomeu> + 0x658d2200, /* [01028] PS.INST_ADDR := ADDR_12_A */
<tomeu> - 0x00000100, /* [00A30] PA.ATTRIBUTE_ELEMENT_COUNT := UNK0=0x0,COUNT=0x1 */
<tomeu> ...
<tomeu> + 0x00000000, /* [00A30] PA.ATTRIBUTE_ELEMENT_COUNT := UNK0=0x0,COUNT=0x0 */
<tomeu> the previous dump was for a kernel that used get_global_id()
<tomeu> the new one has instead a constant (7)
<tomeu> I'm trying to figure out what needs to be in the command stream for the global_id to be available to the kernel
<tomeu> in the kernel, the global_id would be in t0.xxxx
<tomeu> has anybody already poked into those bits?
<tomeu> ah, the different value for PS.INST_ADDR just means a different pointer for the shader bo
<cphealy> Is there any mechanism with etnaviv to track how much GPU memory a process is using?
<austriancoder> tomeu: exactly.. all bits you have in the diff are documented.. soo.. nothing special
<tomeu> yeah, I'm wondering what could determine then that t0.xxxx has to contain the global_id
<tomeu> afaics, PA_ATTRIBUTE_ELEMENT_COUNT is used for varyings in VS and FS
<tomeu> hmm, guess it could be that for compute shaders, t0 contains the global_id, initially
adjtm has joined #etnaviv
mvlad has quit [Quit: Leaving]
pcercuei has quit [Read error: Connection reset by peer]
pcercuei has joined #etnaviv
adjtm is now known as Guest1107
Guest1107 has quit [Read error: Connection reset by peer]
adjtm has joined #etnaviv
adjtm is now known as Guest1118
Guest1118 has quit [Read error: Connection reset by peer]
adjtm has joined #etnaviv
adjtm is now known as Guest1119
Guest1119 has quit [Read error: Connection reset by peer]
adjtm has joined #etnaviv
pcercuei has quit [Quit: dodo]