Bug 940 - SVP64 LD/ST auto-increment mode (working with LDST-update)
Summary: SVP64 LD/ST auto-increment mode (working with LDST-update)
Status: CONFIRMED
Alias: None
Product: Libre-SOC's first SoC
Classification: Unclassified
Component: Specification (show other bugs)
Version: unspecified
Hardware: Other Linux
: --- enhancement
Assignee: Luke Kenneth Casson Leighton
URL: https://libre-soc.org/openpower/sv/ldst/
Depends on:
Blocks:
 
Reported: 2022-10-03 17:16 BST by Luke Kenneth Casson Leighton
Modified: 2022-10-12 10:27 BST (History)
1 user (show)

See Also:
NLnet milestone: ---
total budget (EUR) for completion of task and all subtasks: 0
budget (EUR) for this task, excluding subtasks' budget: 0
parent task for budget allocation:
child tasks for budget allocation:
The table of payments (in EUR) for this task; TOML format:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Luke Kenneth Casson Leighton 2022-10-03 17:16:20 BST 
the 68000 PDP8 PDP11 and 6600 all had auto-increment addressing modes.
Power ISA has LDST-with-update which is not quite the same but almost:
update-with-immediate on a Vector advances by the immediate, but has
an initial "one too many adds"

autoincrementing should use the EA *before* the add is performed
(i.e. just RA) as an option.  thus on loop exit RA is set to
point to where the *next* batch starts.
Comment 1 Luke Kenneth Casson Leighton 2022-10-03 17:45:18 BST 
in LD/ST-immediate there is a mode-combination reserved, which can be used
for saying "use RA only as the EA, but write out (RA+imm) as the update".

this is a significant change as it means modifying the pseudocode:

    ea <- RA+D
    if preincrement then EA <- ea
    else                 EA <- RA
    RT -< MEM(EA, sz)
    RA <- ea

with a default "preincrement=1" in Scalar ISA