Bump up FMC_SIZE - FMC overlfowing in 12 bytes during nightly release tests. #1937
Conversation
rusty1968
requested review from
FerralCoder,
bluegate010,
mhatrevi,
swenson and
vsonims
as code owners
| @@ -38,7 +38,7 @@ pub use fuse::{FuseLogEntry, FuseLogEntryId}; | |||
| pub use pcr::{PcrLogEntry, PcrLogEntryId, RT_FW_CURRENT_PCR, RT_FW_JOURNEY_PCR}; | |||
|
|
|||
| pub const FMC_ORG: u32 = 0x40000000; | |||
| pub const FMC_SIZE: u32 = 21 * 1024; | |||
There was a problem hiding this comment.
This results in a somewhat odd memory layout where the max size of FMC + RT is 117 KiB + 256 bytes. I think I would prefer to keep that at a 1K boundary if possible.
I think the reason we don't use the full 128K of ICCM is that FMC + RT + Manifest must be <= 128KiB to fit in the mailbox. Is the manifest size > 10KiB?
There was a problem hiding this comment.
Also I think we could probably save 12 bytes in some other way fairly easily rather than changing the max size of FMC.
There was a problem hiding this comment.
Hi, Jordan - I did trim strings last time and got feedback to leave them as is, I got away with trimming a few bytes by removing spaces . If we have the constraint of combined mutable size <= 118k , I will express that in the code as a compile assert. I guess we can trim a few bytes but we have to pull in the condiiton that triggered the fault in and Nick will push this shortly.
No description provided.