On 6/29/15 4:43 PM, Konstantin Belousov wrote:
On Mon, Jun 29, 2015 at 04:33:31PM +0800, Julian Elischer wrote:
Hi all,

At $JOB we have a few extra syscalls that we have added to our kernel.

After generating the new sysent files in /sys/kern, libc fails to
compile with:

===> lib/libc (obj,depend,all,install)
building shared library libc.so.7
[...]
/usr/bin/ld: rlk_check_offline.So: relocation R_X86_64_32 against
`SYS_rlk_check_offline' can not be used when making a shared object;
recompile with -fPIC
rlk_check_offline.So: could not read symbols: Bad value
*** [libc.so.7] Error code 1

this suggests that the code that generates the libc syscall stubs is
generating something the linker doesn't like.
No, this suggests that the symbol SYS_rlk_check_offline was undefined
when assembling your rlk_check_offline.S.

Check the way you generated the stuff from syscalls.master.  Most likely,
sys/sys/syscall.h update was lost.

/sys/sys/syscall.h:
#define SYS_pz_openat_ex        587
#define SYS_rlk_check_offline   588  <----
#define SYS_MAXSYSCALL  589

as shown below..



the definition of the syscall is:

588     AUE_NULL        NOSTD   { int rlk_check_offline(char *localfs,
char *path, \
                                      int *is_offline, int rlk_flags, \
                                      int *cache_status); }
------
which generates (in various files):
          { AS(rlk_check_offline_args), (sy_call_t *)lkmressys,
AUE_NULL, NULL, 0, 0, 0, SY_THR_ABSENT }, /* 588 = rlk_check_offline */
------
          "rlk_check_offline",                    /* 588 =
rlk_check_offline */
------
          /* rlk_check_offline */
          case 588: {
                  struct rlk_check_offline_args *p = params;
                  uarg[0] = (intptr_t) p->localfs; /* char * */
                  uarg[1] = (intptr_t) p->path; /* char * */
                  uarg[2] = (intptr_t) p->is_offline; /* int * */
                  iarg[3] = p->rlk_flags; /* int */
                  uarg[4] = (intptr_t) p->cache_status; /* int * */
                  *n_args = 5;
                  break;
          }
-------
#define SYS_rlk_check_offline   588
^^^^^ This must be seen by asm.

Note that in the stock sources, lib/libc/amd64/SYS.h includes syscall.h.

and it does here too..
otherwise all the other syscalls would fail.... hmmm wonder if I need to install the new one into /usr/include/sys/first.

ah that's it I'm doing "make libraries" and it requires the new syscall.h to be in /usr/include.
thanks for that.. now for the next problem...


-------
struct rlk_check_offline_args {
          char localfs_l_[PADL_(char *)]; char * localfs; char
localfs_r_[PADR_(char *)];
          char path_l_[PADL_(char *)]; char * path; char
path_r_[PADR_(char *)];
          char is_offline_l_[PADL_(int *)]; int * is_offline; char
is_offline_r_[PADR_(int *)];
          char rlk_flags_l_[PADL_(int)]; int rlk_flags; char
rlk_flags_r_[PADR_(int)];
          char cache_status_l_[PADL_(int *)]; int * cache_status; char
cache_status_r_[PADR_(int *)];
};
int     sys_rlk_check_offline(struct thread *, struct
rlk_check_offline_args *);
#define SYS_AUE_rlk_check_offline       AUE_NULL

-------



the generated stub looks like:
$ cat /usr/obj/usr/src/lib/libc/rlk_check_offline.S

#include "compat.h"
#include "SYS.h"
RSYSCALL(rlk_check_offline)
          .section .note.GNU-stack,"",%progbits

--------


nothing in this definition looks special, So I'm surprised that the
libc build doesn't like it.
This is a (just) post 10.1 10-stable. But we plan to move to 11 soon too.

Any suggestions as to what I should change would be greatly
appreciated.. I'm running out of ideas.







_______________________________________________
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"


_______________________________________________
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"

Reply via email to