In D37127 the option TCP_HHOOK was moved from opt_inet.h to opt_global.h. The corresponding change in sys/modules/cc/Makefile and sys/modules/khelp/Makefile was missed.
Details
Details
- Reviewers
glebius rrs cc rscheff - Commits
- rGbe78a31188c5: tcp: fix build issue for some cc modules
Check that cc_vegas module is build, when the TCP_HHOOK kernel configuration is enabled, which is the default.
Diff Detail
Diff Detail
- Repository
- rG FreeBSD src repository
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
Comment Actions
Yea, this likely isn't the right fix here at all...
| sys/modules/cc/Makefile | ||
|---|---|---|
| 13 | I don't understand this bit at all... | |
Comment Actions
I'm willing to revert it, of course. But my understanding is that the 'old' condition checks for TCP_HHOOK in opt_inet.h. This was correct before e68b3792440cac248347afe08ba5881a00ba6523, in which`TCP_HHOOK` was moved from opt_inet.h to opt_global.h. So shouldn't this condition reflect this change?
The result was, that cc_cdg, cc_chd, cc_hd, and cc_vegas were not build anymore on systems with TCP_HHOOK in the kernel config. I considered that a bug and this change addresses it.
What do you consider to be the right fix?
| sys/modules/cc/Makefile | ||
|---|---|---|
| 13 |
Old version: Build the CC module if you want to build all modules or TCP_HHOOK is defined in opt_inet.h. New version: Build the CC module if you want to build all modules or TCP_HHOOK is defined in opt_global.h. That is my understanding, however I did not write the original code. I assumed that it is acceptable and concluded that it should change, since TCP_HHOOK was moved from opt_inet.h to opt_global.h. If that is wrong, please let me know. | |