On 1-11-2018 20:59, Beierl, Mark wrote:
On 2018-10-29, 20:35, "ceph-devel-owner@xxxxxxxxxxxxxxx on behalf of Beierl, Mark" <ceph-devel-owner@xxxxxxxxxxxxxxx on behalf of Mark.Beierl@xxxxxxxx> wrote:
On 2018-10-29, 17:58, "ceph-devel-owner@xxxxxxxxxxxxxxx on behalf of Willem Jan Withagen" <ceph-devel-owner@xxxxxxxxxxxxxxx on behalf of wjw@xxxxxxxxxxx> wrote:
>
>> On Oct 29, 2018, at 17:07, Willem Jan Withagen <wjw@xxxxxxxxxxx> wrote:
>>
>>
>>> On 29-10-2018 22:01, Beierl, Mark wrote:
>>> On 2018-10-29, 16:57, "ceph-devel-owner@xxxxxxxxxxxxxxx on behalf of Willem Jan Withagen" <ceph-devel-owner@xxxxxxxxxxxxxxx on behalf of wjw@xxxxxxxxxxx> wrote:
>>> On 29-10-2018 21:45, Noah Watkins wrote:
>>> >>
>>> >> TEMP_FAILURE_RETRY is included in src/include/compat.h, but that isn’t included in blkdev.cc, so I am not sure how this is supposed to compile? What am I missing?
>>> >
>>> > This is probably because musl c doesn't provide that macro (found in
>>> > unistd.h on glibc). I suspect then that blkdev.cc should include the
>>> > compat header.
>>> >
>>> This is indeed fixable by including compat.h
>>> ACCESSPERMS is a bit more complicated, unless we lift the definition for
>>> missing that also compat.h
>>> ./src/client/posix_acl.cc:6:#ifndef ACCESSPERMS
>>> ./src/client/posix_acl.cc:7:#define ACCESSPERMS (S_IRWXU|S_IRWXG|S_IRWXO)
>>> ./src/client/posix_acl.cc:115: *mode_p = (*mode_p & ~ACCESSPERMS) | mode;
>>> ./src/client/posix_acl.cc:177: *mode_p = (*mode_p & ~ACCESSPERMS) | mode;
>>> ./src/common/common_init.cc:129: if (!(ret & (~ACCESSPERMS))) {
>>> Let me know if you don't want to make a PR for this yourself, then I'll
>>> put one in.
>>> --WjW
>>> Sorry, is the suggestion to put ACCESSPERMS into compat.h instead of posix_acl.cc?
>>> I’m not super familiar with the PR process (coming from a Gerrit background), so I would appreciate if you could open one for this.
>> I would think that that would be the easiest fix to do.
>> Got to find an hour to get this done.
>>
>> --WjW
>>
>>
> Thanks, Willem.
>
> I can take a stab at creating a PR on Wednesday if you don’t get the time.
>
> Regards,
> Mark
https://github.com/ceph/ceph/pull/24813
--WjW
Hello, again Willem.
I did the following to test your PR:
git clone https://github.com/wjwithagen/ceph.git
cd ceph
git checkout ed634682263a27dddd2c1525b9720a0718382757
./do_cmake.sh -DWITH_SYSTEM_BOOST=ON -DWITH_LTTNG=OFF -DWITH_REENTRANT_STRSIGNAL=ON -DWITH_THREAD_SAFE_RES_QUERY=ON -DWITH_OPENLDAP=OFF -DWITH_LEVELDB=OFF -DWITH_BABELTRACE=OFF -DWITH_DPDK=OFF -DWITH_SPDK=OFF
Unfortunately, I get the following error now. Did I fetch the PR incorrectly?
CMake Error at /usr/share/cmake/Modules/FindPackageHandleStandardArgs.cmake:137 (message):
Could NOT find verbs (missing: VERBS_LIBRARIES VERBS_INCLUDE_DIR)
Hi Mark,
I guess that that perhaps is not a thing to do, as I usually do not put
my PR's in the head of my github. (I'm not really into git).
Wh I usually do with other peoples PR's is:
fetch https://github.com/ceph/ceph/pull/24813.diff
and then patch that on my sources.
I have a script that does that, pre `cmake`
Then in due time the PR gets in ceph:master, and I get ride of the pathc.
--WjW
