Implement basic support for allocating memory from a specific numa node in the LinuxKPI.
MFC after: 1 week
Sponsored by: Mellanox Technologies // NVIDIA Networking
Details
Details
Diff Detail
Diff Detail
- Repository
- rG FreeBSD src repository
- Lint
Lint Skipped - Unit
Tests Skipped
Event Timeline
| sys/compat/linuxkpi/common/src/linux_domain.c | ||
|---|---|---|
| 47 | Are you sure that FIXED policy is right there? What is the Linux' semantic? From my experience, it almost never is. You typically want a graceful fallback to to other domains if requested domain is full. AFAIR in our terminology this means _PREF(). | |
| sys/kern/kern_malloc.c | ||
| 808 ↗ | (On Diff #85195) | This line looks too long. kern_malloc.c/malloc.h bits should be committed separately. |
Comment Actions
Looks fine to me. I did not looked at the Linux KPI to see if this is a proper implementation.
I suggest to wait some time for Mark' comments.
Comment Actions
Looks ok to me.
| sys/compat/linuxkpi/common/include/linux/device.h | ||
|---|---|---|
| 557–559 | Why is it variadic? | |
| sys/compat/linuxkpi/common/include/linux/device.h | ||
|---|---|---|
| 557–559 | Just to get a better compile error, if the argument passed is bad. | |
| sys/compat/linuxkpi/common/include/linux/device.h | ||
|---|---|---|
| 557–559 | I think there should at least be a comment explaining it, otherwise it looks strange. And, why is it especially likely to get compile errors with this function? | |
| sys/compat/linuxkpi/common/include/linux/device.h | ||
|---|---|---|
| 557–559 | OK, I'll fix. | |