EINTR implementation in the kernel sources - statfs & newfstatat

I'm trying to locate where in the kernel source files does statfs and newfstatat syscalls might fail with EINTR (due to a pending signal). The man pages list EINTR as one of the possible errors these functions might encounter and I can't seem to locate where actually in the Linux source code it is implemented. **EDIT**: I realized that if the filesystem is network-based, say NFS, its implementation can block and thus return EINTR. I'm interested if EINTR might be returned when working against the procfs filesystem. An example idea: if a hardlimit on cputime is reached, a SIGXCPU signal should be sent - if it happens while calling statfs, will it return EINTR?