Remove generic type parameter M from QueueState
#104
Conversation
alexandruag
force-pushed
the
bye_mem2
branch
2 times, most recently
from
7536050 to
1038f7b
Compare
|
Hmm not sure why the coverage drops so much (why it drops at all actually). Will have a look into that. |
| fn check_status_desc<M: GuestMemory>(mem: &M, desc: Descriptor) -> Result<()> { | ||
| fn check_status_desc<M>(mem: &M, desc: Descriptor) -> Result<()> | ||
| where | ||
| M: GuestMemory + ?Sized, |
There was a problem hiding this comment.
Why special "?Sized" marker here?
Comparing the kcov output between this branch and main, looks like kcov is counting this time all the lines that no longer have the |
so let's update the coverage file:) |
|
Can't the drop be covered by unit test(s)? 6.9% is pretty much. |
There are no new lines of code, so it seems like an artifact of how |
|
@alexandruag this change also allows the removal of |
There were new lines of code in this PR that were not covered by unit tests. I wasn't suggesting to add tests in the current PR, but hopefully at some point before publishing the crate. |
|
The pull request in #111 brings coverage up by 2.5% compared to this one. @alexandruag, if you rebase and fix the coverage I think this can be merged. Thanks! |
|
I have included rebased patches in #111. |
|
Any updates about this MR? |
We are using a `: GuestAddressSpace` trait bound for `AvailIter` and `DescriptorChain`, but the objects are actually only interested in the fact that `M` can dereference to a `GuestMemory` implementation. Signed-off-by: Alexandru Agache <aagch@amazon.com>
Signed-off-by: Alexandru Agache <aagch@amazon.com>
There were a couple of changes in rust-vmm#76 that are not yet tested. PR that covers that functionality is in review here: rust-vmm#111. Until that one is merged, the coverage needs to be decreased. Signed-off-by: Laura Loghin <lauralg@amazon.com>
This PR simplifies the
QueueStateabstraction by removing the generic type parameterM, and also relaxes the trait bounds forAvailIterandDescriptorchain to not depend onGuestAddressSpaceanymore, as the logic within is only interested inMdereferencing to something that implementsGuestMemory(and not also in the part that it might be one of the associated type of aGuestAddressSpaceimplementation).