2.0: Remove SuccessWithValue #2430

bradjc · 2021-02-12T21:03:04Z

Pull Request Overview

This pull request removes SuccessWithValue for 2.0. Other than IPC, nothing was using SuccessWithValue anymore.

Testing Strategy

travis

TODO or Help Wanted

Need the IPC updates first.

Documentation Updated

Updated the relevant files in /docs, or no updates are required.

Formatting

Ran make prepush.

hudson-ayers

Looks good modulo the IPC change, of course

phil-levis · 2021-02-16T01:02:42Z

kernel/src/returncode.rs

@@ -42,7 +40,6 @@ pub enum ReturnCode {
 impl From<ReturnCode> for isize {
    fn from(original: ReturnCode) -> isize {
        match original {
-            ReturnCode::SuccessWithValue { value } => value as isize,


bradjc · 2021-02-17T20:22:14Z

Ok new plan...we leave SuccessWithValue but vow to never use it.

But seriously, this change seems to increase the imix kernel size by 4512 bytes.

Before:

make
   Compiling kernel v0.1.0 (/Users/bradjc/git/tock/kernel)
   Compiling cortexm v0.1.0 (/Users/bradjc/git/tock/arch/cortex-m)
   Compiling capsules v0.1.0 (/Users/bradjc/git/tock/capsules)
   Compiling cortexm4 v0.1.0 (/Users/bradjc/git/tock/arch/cortex-m4)
   Compiling sam4l v0.1.0 (/Users/bradjc/git/tock/chips/sam4l)
   Compiling components v0.1.0 (/Users/bradjc/git/tock/boards/components)
   Compiling imix v0.1.0 (/Users/bradjc/git/tock/boards/imix)
    Finished release [optimized + debuginfo] target(s) in 41.70s
   text	   data	    bss	    dec	    hex	filename
 188373	     64	  32704	 221141	  35fd5	/Users/bradjc/git/tock/target/thumbv7em-none-eabi/release/imix
dd516105645d93567b70c598c75fe2e2cb3d4de9b83817aa4f4b49bdbd88746e	/Users/bradjc/git/tock/target/thumbv7em-none-eabi/release/imix.bin

after

make
   Compiling imix v0.1.0 (/Users/bradjc/git/tock/boards/imix)
    Finished release [optimized + debuginfo] target(s) in 18.77s
   text	   data	    bss	    dec	    hex	filename
 192885	     64	  32704	 225653	  37175	/Users/bradjc/git/tock/target/thumbv7em-none-eabi/release/imix
3ba74f8fd1b72bdbb0c9dddf57609caf2ce8834b561b12eda0aebd07c589f592	/Users/bradjc/git/tock/target/thumbv7em-none-eabi/release/imix.bin

bradjc · 2021-02-17T20:34:56Z

Just adding the SuccessWithValue enum value back "fixes" the issue and reduces the size relative to the 2.0 branch:

make
   Compiling kernel v0.1.0 (/Users/bradjc/git/tock/kernel)
   Compiling cortexm v0.1.0 (/Users/bradjc/git/tock/arch/cortex-m)
   Compiling capsules v0.1.0 (/Users/bradjc/git/tock/capsules)
   Compiling cortexm4 v0.1.0 (/Users/bradjc/git/tock/arch/cortex-m4)
   Compiling sam4l v0.1.0 (/Users/bradjc/git/tock/chips/sam4l)
   Compiling components v0.1.0 (/Users/bradjc/git/tock/boards/components)
   Compiling imix v0.1.0 (/Users/bradjc/git/tock/boards/imix)
    Finished release [optimized + debuginfo] target(s) in 34.42s
   text	   data	    bss	    dec	    hex	filename
 188309	     64	  32704	 221077	  35f95	/Users/bradjc/git/tock/target/thumbv7em-none-eabi/release/imix
3e6578da0741e58e8f0b802d445558df4c1394f8f898dd064ea0d39bde04f7ae	/Users/bradjc/git/tock/target/thumbv7em-none-eabi/release/imix.bin

phil-levis · 2021-02-18T19:54:59Z

I will look into this. I'll try to improve the print tool to see what's happening.

hudson-ayers · 2021-02-18T20:46:22Z

Rebased ontop of 2.0-dev now that I merged master in

hudson-ayers · 2021-02-18T21:03:20Z

After the rebase, this PR adds just under 2kB of code, but Imix now fits.

hudson-ayers · 2021-02-18T22:31:30Z

Some tidbits from my investigating the size changes:

Increase seems spread across many functions
Marking all returnCode functions inline(never) → small improvement
Marking inline(always) → little worse
mem::size_of(Returncode now): 1 byte. Before it was 8 bytes!

pub enum ReturnCode {
    SuccessWithValue { value },
    SUCCESS

Total Imix size: 184996 bytes

pub enum ReturnCode {
    SUCCESS,

Total Imix size: 186980

#[repr(isize)]
pub enum ReturnCode {
    SUCCESS,

Total Imix size: 184485

#[repr(rust)]
pub enum ReturnCode {
    SUCCESS,
// ...
#[repr(rust)] // previously was #[repr(usize)]
pub enum ErrorCode {
    SUCCESS,

Total Imix size: 185541

Hypothesis: Making ReturnCode 1 byte instead of 4 decreases stack memory use, however doing so also increases code size, because now kernel has to sign extend values when moving ReturnCodes in/out of a word-size register at the userspace boundary.

Looking at the difference in stack frame sizes, most functions are unchanged. the biggest changes are:

UDP::Driver::command() stack frame 480 bytes --> 536 bytes
IPC::Driver::command() stack frame 232 bytes --> 280 bytes
nrfSerializiation::Driver::command() stack frame 136 bytes --> 216 bytes

alevy · 2021-02-19T00:03:48Z

Following some side discussion, @hudson-ayers's theory seems to hold water and his proposal to just force ReturnCode to usize/isize seems strictly positive---there is no semantic difference within the kernel and it ends up resulting in an overall improvement in code space for this PR.

A note: I'm very skeptical that we get any actual memory savings from the ReturnCode being smaller than a word since likely the compiler is using a word for it on the stack anyway in most cases.

ppannuto

bors r+

alevy · 2021-02-19T17:24:01Z

bors r+

bors · 2021-02-19T17:24:03Z

Already running a review

bors · 2021-02-19T17:35:19Z

Build succeeded:

phil-levis · 2021-02-19T18:56:39Z

A note: I'm very skeptical that we get any actual memory savings from the ReturnCode being smaller than a word since likely the compiler is using a word for it on the stack anyway in most cases.

This seems to be a case where you can trade off RAM for flash; given we are much more cramped for flash, a possible few extra bytes of RAM here or there are well worth it.

2430: 2.0: Remove SuccessWithValue r=ppannuto a=bradjc ### Pull Request Overview This pull request removes `SuccessWithValue` for 2.0. Other than IPC, nothing was using SuccessWithValue anymore. ### Testing Strategy travis ### TODO or Help Wanted Need the IPC updates first. ### Documentation Updated - [x] Updated the relevant files in `/docs`, or no updates are required. ### Formatting - [x] Ran `make prepush`. 2443: 2.0: lps25hb r=alevy a=bradjc ### Pull Request Overview Update the last capsule to 2.0. ### Testing Strategy travis. ### TODO or Help Wanted n/a ### Documentation Updated - [x] Updated the relevant files in `/docs`, or no updates are required. ### Formatting - [x] Ran `make prepush`. Co-authored-by: Brad Campbell <[email protected]> Co-authored-by: Hudson Ayers <[email protected]>

bradjc added blocked Waiting on something, like a different PR or a dependency. tock-2.0 Issues and PRs related to Tock version 2.0. labels Feb 12, 2021

github-actions bot added the kernel label Feb 12, 2021

bradjc mentioned this pull request Feb 12, 2021

Release 2.0 #2429

Closed

23 tasks

ppannuto force-pushed the 2.0-no-successwithvalue branch from 645d2a4 to 5c3de11 Compare February 12, 2021 22:06

hudson-ayers approved these changes Feb 15, 2021

View reviewed changes

phil-levis reviewed Feb 16, 2021

View reviewed changes

bradjc force-pushed the 2.0-no-successwithvalue branch from 5c3de11 to 2edf62d Compare February 17, 2021 19:56

bradjc removed the blocked Waiting on something, like a different PR or a dependency. label Feb 17, 2021

Remove SuccessWithValue

a89173c

hudson-ayers force-pushed the 2.0-no-successwithvalue branch from 2edf62d to a89173c Compare February 18, 2021 20:46

hudson-ayers approved these changes Feb 18, 2021

View reviewed changes

mark ReturnCode as repr(isize) to reduce code size

75098fb

alevy approved these changes Feb 19, 2021

View reviewed changes

ppannuto approved these changes Feb 19, 2021

View reviewed changes

lschuermann approved these changes Feb 19, 2021

View reviewed changes

bors bot merged commit cf21e7f into tock-2.0-dev Feb 19, 2021

bors bot deleted the 2.0-no-successwithvalue branch February 19, 2021 17:35

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

2.0: Remove SuccessWithValue #2430

2.0: Remove SuccessWithValue #2430

Uh oh!

bradjc commented Feb 12, 2021 •

edited

Loading

Uh oh!

hudson-ayers left a comment

Uh oh!

phil-levis Feb 16, 2021

Uh oh!

bradjc commented Feb 17, 2021

Uh oh!

bradjc commented Feb 17, 2021

Uh oh!

phil-levis commented Feb 18, 2021

Uh oh!

hudson-ayers commented Feb 18, 2021

Uh oh!

hudson-ayers commented Feb 18, 2021

Uh oh!

hudson-ayers commented Feb 18, 2021 •

edited

Loading

Uh oh!

alevy commented Feb 19, 2021

Uh oh!

ppannuto left a comment

Uh oh!

alevy commented Feb 19, 2021

Uh oh!

bors bot commented Feb 19, 2021

Uh oh!

bors bot commented Feb 19, 2021

Uh oh!

phil-levis commented Feb 19, 2021

Uh oh!

Uh oh!

Uh oh!

2.0: Remove SuccessWithValue #2430

2.0: Remove SuccessWithValue #2430

Uh oh!

Conversation

bradjc commented Feb 12, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Pull Request Overview

Testing Strategy

TODO or Help Wanted

Documentation Updated

Formatting

Uh oh!

hudson-ayers left a comment

Choose a reason for hiding this comment

Uh oh!

phil-levis Feb 16, 2021

Choose a reason for hiding this comment

Uh oh!

bradjc commented Feb 17, 2021

Uh oh!

bradjc commented Feb 17, 2021

Uh oh!

phil-levis commented Feb 18, 2021

Uh oh!

hudson-ayers commented Feb 18, 2021

Uh oh!

hudson-ayers commented Feb 18, 2021

Uh oh!

hudson-ayers commented Feb 18, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

alevy commented Feb 19, 2021

Uh oh!

ppannuto left a comment

Choose a reason for hiding this comment

Uh oh!

alevy commented Feb 19, 2021

Uh oh!

bors bot commented Feb 19, 2021

Uh oh!

bors bot commented Feb 19, 2021

Uh oh!

phil-levis commented Feb 19, 2021

Uh oh!

Uh oh!

bradjc commented Feb 12, 2021 •

edited

Loading

hudson-ayers commented Feb 18, 2021 •

edited

Loading