strace 4.10 is FTBFS on F-23/22/21: F-23 http://arm.koji.fedoraproject.org/koji/buildinfo?buildID=265919 F-22 http://arm.koji.fedoraproject.org/koji/buildinfo?buildID=266426 F-21 http://arm.koji.fedoraproject.org/koji/buildinfo?buildID=266164 gcc -DHAVE_CONFIG_H -I. -I./linux/aarch64 -I./linux/aarch64 -I./linux -I./linux -Wall -Wwrite-strings -Wsign-compare -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong --param=ssp-buffer-size=4 -grecord-gcc-switches -c -o strace-sigreturn.o `test -f 'sigreturn.c' || echo './'`sigreturn.c gcc -DHAVE_CONFIG_H -I. -I./linux/aarch64 -I./linux/aarch64 -I./linux -I./linux -Wall -Wwrite-strings -Wsign-compare -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong --param=ssp-buffer-size=4 -grecord-gcc-switches -c -o strace-sock.o `test -f 'sock.c' || echo './'`sock.c gcc -DHAVE_CONFIG_H -I. -I./linux/aarch64 -I./linux/aarch64 -I./linux -I./linux -Wall -Wwrite-strings -Wsign-compare -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong --param=ssp-buffer-size=4 -grecord-gcc-switches -c -o strace-socketutils.o `test -f 'socketutils.c' || echo './'`socketutils.c sigreturn.c: In function 'sys_sigreturn': sigreturn.c:27:7: error: 'aarch64_sp_ptr' undeclared (first use in this function) (*aarch64_sp_ptr + SIZEOF_STRUCT_SIGINFO + ^ sigreturn.c:27:7: note: each undeclared identifier is reported only once for each function it appears in sigreturn.c:30:7: error: 'arm_sp_ptr' undeclared (first use in this function) (*arm_sp_ptr + ^ Makefile:2049: recipe for target 'strace-sigreturn.o' failed
Strace needs several fixes which were done after 4.10 release and some which are still discussed on strace-devel ML (syscall table one which breaks getrandom test).
commit 71d6ffc6ed0a539257f0ec1a28d34c8d88bebce8 Author: Andreas Schwab <schwab> Date: Thu Mar 12 16:17:45 2015 +0100 aarch64: properly decode generic syscalls * linux/aarch64/syscallent1.h: Don't override entries 277 to 1023. On this commit there is only stat64-v test failing on aarch64.
stat64-v test was fixed by commit v4.10-15-gf79252f.
Dmitry: cool! (I did not checked your branch before)
There is another aarch64-related commit in HEAD: v4.10-16-g68804b3 enables uid.test on architectures like aarch64. Scratch build reports no problems with v4.10-16-g68804b3 on aarch64: http://arm.koji.fedoraproject.org/koji/taskinfo?taskID=2923097
Dmitry: how about update of package in Fedora to get it built on aarch64?
(In reply to Marcin Juszkiewicz from comment #6) > Dmitry: how about update of package in Fedora to get it built on aarch64? That is, to release a new version of strace, and update the package in Fedora?
Would be great :)
ping?
Ping. Can you please update the Fedora 22 package?
Can we have this fixed for aarch64 and updates pushed. This has been broken/regressed for 2 months and is now causing us quite a number of issues due to dependency chains on builds.
strace-4.10-2.fc23 uploaded into rawhide with set of backports added to get it back into buildable state on aarch64. Package builds on all primary and secondary architectures.