Unverified Commit 5504b29d authored by Connor Abbott's avatar Connor Abbott Committed by GitHub
Browse files

midgard: Initial decoding of compact branches

parent d7d2198d
# T6xx Architecture
# Utgard Architecture
## Publicly available information
......@@ -66,8 +66,8 @@ The first (32-bit) word is a control word which, in addition to the usual 8-bit
21: vector add ALU (48 bits)
23: scalar multiply ALU (32 bits)
25: LUT / multiply ALU 2 (48 bits)
26: output write/discard? (16 bits)
27: branch (48 bits)
26: compact output write/branch (16 bits)
27: output write/branch (48 bits)
It's not clear why only every other bit is used for the ALU's (fp64?).
......@@ -219,6 +219,24 @@ Pseudocode for how atan/atan2 is implemented:
To do atan instead of atan2, replace y with 1.0. asin and acos are implemented just like in the Mali 200 PP.
### Compact branch/framebuffer
This field is used for encoding branches, as well as framebuffer writes. So far, we've only figured out branches.
0-2: unknown/opcode?
010 for conditional branches
3-6: Type of instruction to branch to
Note: this complements the next-instruction-type field that every instruction has, which gives the type of the next instruction to execute if there is no branching.
7-13: Offset to branch to
This gives the low 7 bits of the offset in units of quadwords (16 bytes) relative to the next instruction that would be executed.
14-15: Opcode?/sign extending
00 - unknown
01 - extend branch offset with all 1's (use for negative offsets)
10 - extend branch offset with all 0's (use for positive offsets)
11 - unknown, seen used for framebuffer writes.
I'm a little confused about the offset bit -- the way I've described things, it should be possible to encode positive branch offsets of at most 2^7 - 1 quadwords. But the blob refuses to encode branch offsets of more than 2^6 - 1 quadwords, switching to the larger branch encoding instead, as if the offset field is sign-extended. But then it also sets high two bits differently depending on whether the branch offset is positive or negative, which seems redundant. Is the blob just being inefficient, or am I missing something?
## Load/store words
The load/store word consists of the standard 8-bit tag, followed by two 60-bit instructions whose format is described below. Each instruction can load or store up to 128 bits at once.
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment