Re: [PATCH] Fix gdb.arch/aarch64-dbreg-contents.exp build failures

[Luis Machado <luis dot machado at linaro dot org>]

On 2/26/20 8:46 AM, Alan Hayward wrote:
> > On 25 Feb 2020, at 13:15, Luis Machado <luis.machado@linaro.org> wrote:
> >
> > I ran into the following failures when running tests under QEMU:
> >
> > --
> >
> > gdb compile failed, binutils-gdb/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c: In function 'set_watchpoint':
> > binutils-gdb/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c:41:29: error: storage size of 'dreg_state' isn't known
> >    struct user_hwdebug_state dreg_state;
> >                              ^~~~~~~~~~
> > In file included from /usr/include/aarch64-linux-gnu/bits/types/struct_iovec.h:23:0,
> >                  from /usr/include/aarch64-linux-gnu/sys/uio.h:23,
> >                  from binutils-gdb/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c:17:
> > binutils-gdb/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c:69:18: error: invalid use of undefined type 'struct user_hwdebug_state'
> >    iov.iov_len = (offsetof (struct user_hwdebug_state, dbg_regs)
> >                   ^
> > binutils-gdb/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c:74:5: warning: implicit declaration of function 'error'; did you mean 'errno'? [-Wimplicit-function-declaration]
> >      error (1, errno, "PTRACE_SETREGSET: NT_ARM_HW_WATCH");
> >      ^~~~~
> >      errno
> > binutils-gdb/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c: In function 'main':
> > binutils-gdb/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c:87:3: warning: implicit declaration of function 'atexit'; did you mean '_Exit'? [-Wimplicit-function-declaration]
> >    atexit (cleanup);
> >    ^~~~~~
> >    _Exit
> > binutils-gdb/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c:89:11: warning: implicit declaration of function 'fork' [-Wimplicit-function-declaration]
> >    child = fork ();
> >
> >            ^~~~
> >
> > --
> >
> > The following patch fixes those by adding the necessary include files.
> >
> > With that said, the test doesn't pass at present. I'll have to investigate it
> > a bit more.
> >
> > gdb/testsuite/ChangeLog:
> >
> > 2020-02-25  Luis Machado  <luis.machado@linaro.org>
> >
> > 	* gdb.arch/aarch64-dbreg-contents.c: Include stdlib.h, unistd,
> > 	asm/ptrace.h and error.h.
> >
> > Signed-off-by: Luis Machado <luis.machado@linaro.org>
> > ---
> > gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c | 4 ++++
> > 1 file changed, 4 insertions(+)
> >
> > diff --git a/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c b/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c
> > index 66ca76f847..ca690f63de 100644
> > --- a/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c
> > +++ b/gdb/testsuite/gdb.arch/aarch64-dbreg-contents.c
> > @@ -9,13 +9,17 @@
> >    freely.  */
> >
> > #define _GNU_SOURCE 1
> > +#include <stdlib.h>
> > +#include <unistd.h>
> > #include <sys/ptrace.h>
> > +#include <asm/ptrace.h>
>
> I get a little concerned when we have to include multiple sys, asm, bits etc
> version of a header. However, saying that we include both of these in many
> aarch64 gdb files (eg aarch64-linux-hw-point.c).

I was expecting sys/ptrace.h to include asm/ptrace.h or something a bit 
cleaner. But it seems to be the way things are laid out at the moment, 
as some other files are doing the same.

> Ok to push.

Thanks. Pushed now.