Reduce fudge factor from shoehorn and strip elf files before loading #158
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
kent-mcleod
added
hw-build
|
There are some errors that I'll try and track down. The Elfloader is reporting that the loaded regions are all exclusive, however there appears to be consistent hanging before executing the kernel in some cases. |
kent-mcleod
force-pushed
the
kent/fudge
branch
8 times, most recently
from
e396746 to
f438dbf
Compare
kent-mcleod
changed the title
kent-mcleod
changed the title
lsf37
reviewed
Feb 28, 2025
| @@ -57,6 +57,13 @@ function(ApplyData61ElfLoaderSettings kernel_platform kernel_sel4_arch) | |||
| endif() | |||
| set(IMAGE_START_ADDR 0x10000000 CACHE INTERNAL "" FORCE) | |||
| endif() | |||
| if(KernelPlatformSpike OR KernelPlatformQEMURiscVVirt) | |||
| # Spike loads elfloader to either 0x80200000 or 0x80400000 | |||
| # But the RISC-V kernel is currently loaded to a differen't | |||
There was a problem hiding this comment.
Suggested change
| # But the RISC-V kernel is currently loaded to a differen't | |
| # But the RISC-V kernel is currently loaded to a different |
lsf37
approved these changes
Feb 28, 2025
kent-mcleod
force-pushed
the
kent/fudge
branch
from
f438dbf to
347b969
Compare
Correct the calculated memory needed to load an elf image. Using only the p_memsz value ignores gaps between segments; instead calculate the first+last addresses using p_vaddr & p_memsz. Signed-off-by: Kent McLeod <kent@kry10.com>
Strip out sections from kernel and rootserver elffiles before loading them into the cpio archive. Nothing makes use of this information at runtime and it causes the elf binary to be larger than it needs to be. Signed-off-by: Kent McLeod <kent@kry10.com>
As get_memory_usage in elf_sift.py now can calculate correct memory usage from a list of ELF segments, a more accurate upper bound on memory usage for the loaded apps is possible and the old fudge factor is no longer needed. We still need some gap to account for the size of the elfloader program so that the elfloader ends up above itself in memory if loaded to the lowest available address. Signed-off-by: Kent McLeod <kent@kry10.com>
Shoehorn currently can't manage to pick a good load address. Signed-off-by: Kent McLeod <kent@kry10.com>
kent-mcleod
force-pushed
the
kent/fudge
branch
from
347b969 to
7726e9b
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Correct the calculated memory needed to load an elf image. Using only
the p_memsz value ignores gaps between segments; instead calculate the
first+last addresses using p_vaddr & p_memsz.
As get_memory_usage in elf_sift.py now can calculate correct memory
usage from a list of ELF segments, a more accurate upper bound on
memory usage for the loaded apps is possible and the old fudge
factor is no longer needed.
Thanks to @sleffler for the patch.