Re: linux-next failing build due to missing cubictcp_state symbol

From: Jiri Slaby
Date: Mon May 03 2021 - 04:59:51 EST


CCing pahole people.

On 03. 05. 21, 9:59, Jiri Slaby wrote:
On 03. 05. 21, 8:11, Jiri Slaby wrote:
looks like vfs_truncate did not get into BTF data,
I'll try to reproduce

_None_ of the functions are generated by pahole -J from debuginfo on ppc64. debuginfo appears to be correct. Neither pahole -J fs/open.o works correctly. collect_functions in dwarves seems to be defunct on ppc64... "functions" array is bogus (so find_function -- the bsearch -- fails).

It's not that bogus. I forgot an asterisk:
#0  find_function (btfe=0x100269f80, name=0x10024631c "stream_open") at /usr/src/debug/dwarves-1.21-1.1.ppc64/btf_encoder.c:350
(gdb) p (*functions)@84
$5 = {{name = 0x7ffff68e0922 ".__se_compat_sys_ftruncate", addr = 75232, size = 72, sh_addr = 65536, generated = false}, {
    name = 0x7ffff68e019e ".__se_compat_sys_open", addr = 80592, size = 216, sh_addr = 65536, generated = false}, {
    name = 0x7ffff68e0076 ".__se_compat_sys_openat", addr = 80816, size = 232, sh_addr = 65536, generated = false}, {
    name = 0x7ffff68e0908 ".__se_compat_sys_truncate", addr = 74304, size = 100, sh_addr = 65536, generated = false}, {
...
    name = 0x7ffff68e0808 ".stream_open", addr = 65824, size = 72, sh_addr = 65536, generated = false}, {
...
    name = 0x7ffff68e0751 ".vfs_truncate", addr = 73392, size = 544, sh_addr = 65536, generated = false}}

The dot makes the difference, of course. The question is why is it there? I keep looking into it. Only if someone has an immediate idea...

Well, .vfs_truncate is in .text (and contains an ._mcount call). And vfs_truncate is in .opd (w/o an ._mcount call). Since setup_functions excludes all functions without the ._mcount call, is_ftrace_func later returns false for such functions and they are filtered before the BTF processing.

Technically, get_vmlinux_addrs looks at a list of functions between __start_mcount_loc and __stop_mcount_loc and considers only the listed.

I don't know what the correct fix is (exclude .opd functions from the filter?). Neither why cross compiler doesn't fail, nor why ebi v2 avoids this too.

Attaching a patch for pahole which fixes the issue, but I have no idea whether it is the right fix at all.

regards,--
js
suse labs
From: Jiri Slaby <jslaby@xxxxxxx>
Subject: ppc64: .opd section fix
Patch-mainline: submitted 2021/05/03

Functions in the .opd section should be considered valid too. Otherwise,
pahole cannot produce a .BTF section from vmlinux and kernel build
fails on ppc64.
---
btf_encoder.c | 18 +++++++++++++++++-
1 file changed, 17 insertions(+), 1 deletion(-)

--- a/btf_encoder.c
+++ b/btf_encoder.c
@@ -31,6 +31,8 @@ struct funcs_layout {
unsigned long mcount_start;
unsigned long mcount_stop;
unsigned long mcount_sec_idx;
+ unsigned long opd_start;
+ unsigned long opd_stop;
};

struct elf_function {
@@ -271,11 +273,24 @@ static int is_ftrace_func(struct elf_fun
return start <= addrs[r] && addrs[r] < end;
}

+static int is_opd_func(struct elf_function *func, struct funcs_layout *fl)
+{
+ return fl->opd_start <= func->addr && func->addr < fl->opd_stop;
+}
+
static int setup_functions(struct btf_elf *btfe, struct funcs_layout *fl)
{
__u64 *addrs, count, i;
int functions_valid = 0;
bool kmod = false;
+ GElf_Shdr shdr;
+ Elf_Scn *sec;
+
+ sec = elf_section_by_name(btfe->elf, &btfe->ehdr, &shdr, ".opd", NULL);
+ if (sec) {
+ fl->opd_start = shdr.sh_addr;
+ fl->opd_stop = shdr.sh_addr + shdr.sh_size;
+ }

/*
* Check if we are processing vmlinux image and
@@ -322,7 +337,8 @@ static int setup_functions(struct btf_el
func->addr += func->sh_addr;

/* Make sure function is within ftrace addresses. */
- if (is_ftrace_func(func, addrs, count)) {
+ if (is_opd_func(func, fl) ||
+ is_ftrace_func(func, addrs, count)) {
/*
* We iterate over sorted array, so we can easily skip
* not valid item and move following valid field into