| Summary: | sys-fs/zfs with reiser4-patched kernel linux-4.* | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | piotr5 |
| Component: | Current packages | Assignee: | Sam James <sam> |
| Status: | UNCONFIRMED --- | ||
| Severity: | normal | CC: | fearedbliss, sam |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: |
add this into /etc/portage/patches/sys-fs/zfs-kmod/ to use zfs and reiser4
fixes configure.am instead of configure -- using both patches works too... |
||
Created attachment 496296 [details, diff]
fixes configure.am instead of configure -- using both patches works too...
|
Created attachment 496292 [details, diff] add this into /etc/portage/patches/sys-fs/zfs-kmod/ to use zfs and reiser4 it's an old issue and is marked as closed by the devs: problem: patch a kernel with a fitting reiser4, compile the kernel and modules, then "emerge spl zfs-kmod" and zfs (if version is high enough for the kernel you chose) will fail compiling with the error of gcc saying it cannot find any function called "new_sync_write". upstream observed the issue goes away when not using reiser4 so the solution is to not use that patch. solution: the problem is actually caused by zfs configuration process, it wrongly assumed that since there is a "new_sync_read" function (which has been put into "linux/fs.h" by reiser4), there also must be a new_sync_write function too and neglected to test for such a function explicitly. my first patch adds that test into the configure script. haven't tested if a system offering new_sync_read and new_sync_write will pass the test, but at least for me the reiser4 patched kernel will fail that test. for the version-controlled release my 2nd patch will add the same test to the m4 files used for creating configure out of configure.am I hope. haven't tested that either.