Redefine asm_inline for Linux 5.4+, Fixes #2546 (#2547)

[prakashsurya]

Here's the upstream commit that introduced asm_inline:

• torvalds/linux@eb11186

Without this patch BCC fails with the following:

$ sudo /usr/share/bcc/tools/opensnoop
In file included from /virtual/main.c:2:
In file included from include/uapi/linux/ptrace.h:142:
In file included from ./arch/x86/include/asm/ptrace.h:5:
./arch/x86/include/asm/segment.h:254:2: error: expected '(' after 'asm'
        alternative_io ("lsl %[seg],%[p]",
        ^
./arch/x86/include/asm/alternative.h:240:2: note: expanded from macro 'alternative_io'
        asm_inline volatile (ALTERNATIVE(oldinstr, newinstr, feature)   \
        ^
include/linux/compiler_types.h:210:24: note: expanded from macro 'asm_inline'
                       ^

[prakashsurya]

git-ab-pre-push is here

This is a clean cherry-pick of the upstream commit (2d1497c) to our 6.0/stage branch.

I've tested this with git-linux-pkg to build and deploy this change to a 6.0.5.0 based VM, and verified the profile command now works:

$ sudo /usr/share/bcc/tools/profile -adfI --stack-storage-size 1024 5
postgres;[unknown];__libc_start_main;main;PostmasterMain;ServerLoop;PostgresMain;ReadyForQuery;socket_flush;internal_flush;send;-;entry_SYSCALL_64_after_hwframe_[k];do_syscall_64_[k];__x64_sys_sendto_[k];__sys_sendto_[k];sock_sendmsg_[k];inet_sendmsg_[k];tcp_sendmsg_[k];tcp_sendmsg_locked_[k];tcp_push_[k];__tcp_push_pending_frames_[k];tcp_write_xmit_[k];tcp_event_new_data_sent_[k];tcp_rearm_rto_[k];sk_reset_timer_[k];mod_timer_[k];__lock_text_start_[k];__lock_text_start_[k] 1
z_wr_int;ret_from_fork_[k];kthread_[k];taskq_thread_[k];zio_execute_[k];zio_vdev_io_done_[k];vdev_queue_io_done_[k];zio_nowait_[k];zio_vdev_io_start_[k];vdev_disk_io_start_[k];__vdev_disk_physio.constprop.0_[k];vdev_submit_bio_[k];vdev_submit_bio_impl_[k];submit_bio_[k];generic_make_request_[k];generic_make_request_checks_[k];percpu_counter_add_batch_[k];percpu_counter_add_batch_[k] 1
swapper/1;secondary_startup_64_[k];start_secondary_[k];cpu_startup_entry_[k];do_idle_[k];default_idle_call_[k];arch_cpu_idle_[k];__sched_text_end_[k];native_safe_halt_[k];ret_from_intr_[k];do_IRQ_[k];irq_exit_[k];__softirqentry_text_start_[k];net_rx_action_[k];vmxnet3_poll_rx_only_[k];vmxnet3_poll_rx_only_[k] 1
z_ioctl_int;ret_from_fork_[k];kthread_[k];taskq_thread_[k];schedule_[k];__sched_text_start_[k];finish_task_switch_[k];finish_task_switch_[k] 1
z_wr_iss;ret_from_fork_[k];kthread_[k];taskq_thread_[k];zio_execute_[k];zio_write_compress_[k];zio_compress_data_[k];lzjb_compress_[k];lzjb_compress_[k] 1
swapper/1;secondary_startup_64_[k];start_secondary_[k];cpu_startup_entry_[k];do_idle_[k];default_idle_call_[k];arch_cpu_idle_[k];__sched_text_end_[k];native_safe_halt_[k];native_safe_halt_[k] 242
swapper/0;secondary_startup_64_[k];x86_64_start_kernel_[k];x86_64_start_reservations_[k];start_kernel_[k];arch_call_rest_init_[k];rest_init_[k];cpu_startup_entry_[k];do_idle_[k];default_idle_call_[k];arch_cpu_idle_[k];__sched_text_end_[k];native_safe_halt_[k];native_safe_halt_[k] 243

[pzakha]

@prakashsurya I think instead of cherry-picking a particular commit, let's fast-forward bcc on 6.0/stage to point to the same commit as master, which has that change.

[prakashsurya]

I'll leave that decision up to you and @sebroy. I'm fine with that, but it does mean more changes will be going into the release branch, and I can't comment on if that's safe (or not) to do.

I do feel relatively confident in landing this PR in isolation for the upcoming release branch, though.

Feel free to close this, and do the bulk merge if that's the the approach you want to take.

[pzakha]

I've opened a PR for the bulk merge here: #6

[prakashsurya]

@pzakha Keep in mind, 6.0/stage is 6.0.6.0. So unless we re-target that PR for 6.0/release, it wouldn't apply for 6.0.5.0.