|
Bugzilla – Full Text Bug Listing |
| Summary: | Building cilium on x86_64 fails after LLVM 11 update | ||
|---|---|---|---|
| Product: | [openSUSE] openSUSE Tumbleweed | Reporter: | Aaron Puchert <aaronpuchert> |
| Component: | Kubic | Assignee: | Michał Rostecki <mrostecki> |
| Status: | RESOLVED FIXED | QA Contact: | E-mail List <qa-bugs> |
| Severity: | Normal | ||
| Priority: | P5 - None | CC: | dmueller, mrostecki |
| Version: | Current | ||
| Target Milestone: | --- | ||
| Hardware: | Other | ||
| OS: | Other | ||
| Whiteboard: | |||
| Found By: | --- | Services Priority: | |
| Business Priority: | Blocker: | --- | |
| Marketing QA Status: | --- | IT Deployment: | --- |
| Bug Depends on: | |||
| Bug Blocks: | 1184920 | ||
To quote C11 [1], 6.3.2.3.6: > Any pointer type may be converted to an integer type. Except as previously > specified, the result is implementation-defined. If the result cannot be > represented in the integer type, the behavior is undefined. The result need > not be in the range of values of any integer type. So this is warning about undefined behavior. However, 7.20.1.4 comes to the rescue: > The following type designates a signed integer type with the property that any > valid pointer to void can be converted to this type, then converted back to > pointer to void, and the result will compare equal to the original pointer: > intptr_t > The following type designates an unsigned integer type with the property that > any valid pointer to void can be converted to this type, then converted back > to pointer to void, and the result will compare equal to the original pointer: > uintptr_t > These types are optional. [1] http://www.open-std.org/jtc1/sc22/wg14/www/docs/n1570.pdf Cilium doesn't use stdint.h on purpose - we don't want to have a dependency on glibc. But there is already a fix which does: # define __fetch(X) (__u32)(__u64)(&(X)) https://github.com/cilium/cilium/commit/e43c2fff3749614a0d5908df6f15db6ea96aed94 Let me push an update. Fix landed. (Quite some time ago in fact.) Could you bring this fix to openSUSE:Backports:SLE-15-SP3/cilium as well? It's currently broken because we updated to LLVM 11 by default (bug 1184920). (In reply to Aaron Puchert from comment #5) > Could you bring this fix to openSUSE:Backports:SLE-15-SP3/cilium as well? > > It's currently broken because we updated to LLVM 11 by default (bug 1184920). Fixed by Max Lin via https://build.opensuse.org/request/show/887868. |
There is a new default warning, and cilium builds with -Werror: /home/abuild/rpmbuild/BUILD/go/src/github.com/cilium/cilium/bpf/lib/dbg.h:168:13: error: cast to smaller integer type '__u32' (aka 'unsigned int') from '__u32 *' (aka 'unsigned int *') [-Werror,-Wpointer-to-int-cast] .source = EVENT_SOURCE, ^~~~~~~~~~~~ bpf_lxc.c:21:22: note: expanded from macro 'EVENT_SOURCE' #define EVENT_SOURCE LXC_ID ^~~~~~ /home/abuild/rpmbuild/BUILD/go/src/github.com/cilium/cilium/bpf/lxc_config.h:28:16: note: expanded from macro 'LXC_ID' #define LXC_ID fetch_u32(LXC_ID) ^~~~~~~~~~~~~~~~~ /home/abuild/rpmbuild/BUILD/go/src/github.com/cilium/cilium/bpf/lib/utils.h:130:22: note: expanded from macro 'fetch_u32' #define fetch_u32(x) __fetch(x) ^~~~~~~~~~ /home/abuild/rpmbuild/BUILD/go/src/github.com/cilium/cilium/bpf/lib/utils.h:122:21: note: expanded from macro '__fetch' # define __fetch(x) (__u32)(&(x)) ^~~~~~~~~~~~~ There are more occurrences. The warning seems valid to me: pointers on x86_64 are 64-bit wide, so the conversion to a 32-bit integer is lossy. If this is intended, convert to uintptr_t first, then to __u32: # define __fetch(x) (__u32)(uintptr_t)(&(x)) That might require #include <stdint.h>. Generally only conversions to uintptr_t are guaranteed to have no information loss, i.e. you can convert them back to a pointer and get the original pointer back. LLVM 11 is currently staged in K, so you can see the full log here: https://build.opensuse.org/package/live_build_log/openSUSE:Factory:Staging:K/cilium/standard/x86_64