AuroraMiddleware/glibc
1
0
Fork 0
mirror of https://sourceware.org/git/glibc.git synced 2024-11-08 14:20:07 +00:00
Code Issues Packages Projects Releases Wiki Activity

nptl: initialize rseq area prior to registration

Browse Source 
Per the rseq syscall documentation, 3 fields are required to be
initialized by userspace prior to registration, they are 'cpu_id',
'rseq_cs' and 'flags'. Since we have no guarantee that 'struct pthread'
is cleared on all architectures, explicitly set those 3 fields prior to
registration.

Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Reviewed-by: Florian Weimer <fweimer@redhat.com>
This commit is contained in:
Michael Jeanson 2024-11-07 22:23:49 +01:00 committed by Florian Weimer
parent c18de3b76a
commit 97f60abd25
2 changed files with 12 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
nptl/descr.h
View File

@ -414,6 +414,8 @@ struct pthread
{
uint32_t cpu_id_start;
uint32_t cpu_id;
uint64_t rseq_cs;
uint32_t flags;
};
char pad[32]; /* Original rseq area size. */
} rseq_area __attribute__ ((aligned (32)));

10
sysdeps/unix/sysv/linux/rseq-internal.h
View File

@ -51,11 +51,21 @@ rseq_register_current_thread (struct pthread *self, bool do_rseq)
/* The initial implementation used only 20 bytes out of 32,
but still expected size 32. */
size = RSEQ_AREA_SIZE_INITIAL;
/* Initialize the rseq fields that are read by the kernel on
registration, there is no guarantee that struct pthread is
cleared on all architectures. */
THREAD_SETMEM (self, rseq_area.cpu_id, RSEQ_CPU_ID_UNINITIALIZED);
THREAD_SETMEM (self, rseq_area.rseq_cs, 0);
THREAD_SETMEM (self, rseq_area.flags, 0);
int ret = INTERNAL_SYSCALL_CALL (rseq, &self->rseq_area,
size, 0, RSEQ_SIG);
if (!INTERNAL_SYSCALL_ERROR_P (ret))
return true;
}
/* When rseq is disabled by tunables or the registration fails, inform
userspace by setting 'cpu_id' to RSEQ_CPU_ID_REGISTRATION_FAILED. */
THREAD_SETMEM (self, rseq_area.cpu_id, RSEQ_CPU_ID_REGISTRATION_FAILED);
return false;
}