select: fix overflow protection - #5286
Conversation
| #if INT_MAX > TIME_T_MAX | ||
| if(timeout_ms > (int)TIME_T_MAX) | ||
| timeout_ms = (int)TIME_T_MAX; | ||
| #endif |
There was a problem hiding this comment.
Can you mention a platform where this change is necessary?
There was a problem hiding this comment.
I'm not aware of any platform where INT_MAX > TIME_T_MAX. It's correct to check it though. typecast is because time_t may be unsigned.
There was a problem hiding this comment.
While it might be correct to check, curl doesn't build on any platform where int is not 32 bit (and time_t is never smaller than 32 bits either). I'm sure such a theoretical platform would break the curl build in numerous other places. And @mback2k has a larger take on this code in progress anyway. But I won't object if you insist.
There was a problem hiding this comment.
Yes I can't think of one. For posterity. #5262 may cover this PR, and if not it can be incorporated. I'll wait to see what happens there.
There was a problem hiding this comment.
@jay I would be fine with merging this PR before the release, as my changes will be part of the next feature window.
There was a problem hiding this comment.
Ok. I will put this in for now and then you can replace it when you're ready.
Make all functions in select.[ch] take timeout_ms as timediff_t which should always be large enough and signed on all platforms to take all possible timeout values and avoid type conversions. Reviewed-by: Jay Satiro Reviewed-by: Daniel Stenberg Replaces #5107 and partially #5262 Related to #5240 and #5286 Closes #5343
3 participants
Follow-up to a96c752 which changed the timeout_ms type from time_t to
timediff_t.
Ref: #5240
Closes #xxxx
5240 added a change after the review to typecast timeout_ms to time_t but the problem with that is if time_t max is less than timediff_t max then possible signed overflow