lkl: Remove the stat symbol hijack #583
Conversation
|
@thehajime Got any recollection on why the |
|
Erk, not sure what Ubuntu has done to their glibc... |
|
Turns out aarch64 apparently uses different constants, marking as a draft again. |
timschumi
changed the title
|
Removing the symbol fully passes CI, let me know if there is some user I am missing. The only alternative to this approach (as far as I can see) is to update the minimum environment to have at least glibc 2.33, where a non-internal |
tavip
left a comment
tavip
left a comment
There was a problem hiding this comment.
LGTM, but lets wait for @thehajime to review this as we don't run DPDK tests in CI.
We are hijacking `stat`, where the size of the target struct is dependent on the wordsize. Delegating to `__xstat64` is incorrect, as it unconditionally uses the stat64 layout and will result in an out-of-bounds write on 32-bit architectures. The replacement would be to use `__xstat` as the implementation instead, but the requirements on the `vers` parameter make it hard to portably call. Since the symbol just delegates to the real implementation unconditionally and therefore does not seem to have a deeper purpose, just remove it. Fixes: d4b9b65 ("lkl: update dpdk version to 17.02 from 2.2") Signed-off-by: Tim Schumacher <[email protected]>
|
No changes made in the force push, had an accidental leftover line in the commit message from squashing. |
|
thanks @timschumi for the patch. the changes looks good to me. |
3 participants
We are hijacking
stat, where the size of the target struct is dependent on the wordsize. Delegating to__xstat64is incorrect, as it unconditionally uses the stat64 layout and will result in an out-of-bounds write on 32-bit architectures.The replacement would be to use
__xstatas the implementation instead, but the requirements on theversparameter make it hard to portably call.Since the symbol just delegates to the real implementation unconditionally and therefore does not seem to have a deeper purpose, just remove it.
Fixes: #361