This is the mail archive of the mailing list for the glibc project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] S390: Sync ptrace.h with kernel. [BZ #21539]

On 06/19/2017 03:26 PM, Dmitry V. Levin wrote:
On Mon, Jun 19, 2017 at 03:10:57PM +0200, Stefan Liebler wrote:
On 06/13/2017 10:05 PM, Dmitry V. Levin wrote:
On Tue, Jun 06, 2017 at 12:17:33PM +0200, Stefan Liebler wrote:
diff --git a/sysdeps/unix/sysv/linux/s390/tst-ptrace-singleblock.c b/sysdeps/unix/sysv/linux/s390/tst-ptrace-singleblock.c
new file mode 100644
  >> [...]
+      /* Get information about tracee: gprs, last breaking address.  */
+      parea.len = sizeof (regs);
+      parea.process_addr = (unsigned long) &regs;
+      parea.kernel_addr = 0;
+      ptrace (PTRACE_PEEKUSR_AREA, pid, &parea);

Note that you can verify whether PTRACE_PEEKUSR_AREA has returned
the expected result by comparing registers with those returned
by PTRACE_GETREGSET.  The latter is implemented on s390 since
linux 2.6.27 so its use in glibc is safe.

Okay. Now the gprs are obtained by PTRACE_PEEKUSR_AREA and
PTRACE_GETREGSET. Afterwards I use memcmp to check whether the values
are the same.
+      ptrace (PTRACE_GET_LAST_BREAK, pid, NULL, &last_break);

As these ptrace calls are expected to succeed,
you might want to check their return code.

Done with several usages of TEST_VERIFY_EXIT.

+      printf ("child IA: %p last_break: %p\n",
+	      (void *) regs[1], (void *) last_break);
+      /* Execute tracee until next taken branch.
+	 Note:
+	 Before the commit which introduced this testcase,
+	 <glibc>/sysdeps/unix/sysv/linux/s390/sys/ptrace.h
+	 uses ptrace-request 12 for PTRACE_GETREGS,
+	 but <kernel>/include/uapi/linux/ptrace.h
+	 uses 12 for PTRACE_SINGLEBLOCK.
+	 The s390 kernel has no support for PTRACE_GETREGS!
+	 Thus glibc ptrace.h is adjusted to match kernel ptrace.h.
+	 This test ensures, that PTRACE_SINGLEBLOCK defined in glibc
+	 works as expected.  If the kernel would interpret it as
+	 PTRACE_GETREGS, then the tracee will not make any progress
+	 and this testcase will time out.  */
+      ptrace (req_singleblock, pid, NULL, NULL);


I've attached the patch with the mentioned changes and the NEWS entry
requested by Andreas.

Is this okay?

Looks fine, thanks.



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]