Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Pedro Alves <palves@redhat.com>
To: Gary Benson <gbenson@redhat.com>, gdb-patches@sourceware.org
Cc: Tom Tromey <tromey@redhat.com>, Doug Evans <dje@google.com>
Subject: Re: [PATCH 02/15 v2] Remove some GDBSERVER checks from linux-ptrace
Date: Thu, 17 Jul 2014 16:40:00 -0000	[thread overview]
Message-ID: <53C7FB3E.5010402@redhat.com> (raw)
In-Reply-To: <1405520243-17282-3-git-send-email-gbenson@redhat.com>

On 07/16/2014 03:17 PM, Gary Benson wrote:
> This patch removes some GDBSERVER checks from nat/linux-ptrace.c.
> Currently the code uses a compile-time check to decide whether some
> flags should be used.  This changes the code to instead let users of
> the module specify an additional set of flags; and then changes gdb's
> linux-nat.c to call this function. 

> At some later date, when the back
> ends are fully merged, we will be able to remove this function again.


> @@ -37,6 +37,10 @@
>     there are no supported features.  */
>  static int current_ptrace_options = -1;
>  
> +/* Additional flags to test.  */
> +
> +static int additional_flags;
> +
>  /* Find all possible reasons we could fail to attach PID and append
>     these as strings to the already initialized BUFFER.  '\0'
>     termination of BUFFER must be done by the caller.  */
> @@ -359,16 +363,15 @@ linux_check_ptrace_features (void)
>  static void
>  linux_test_for_tracesysgood (int child_pid)
>  {
> -#ifdef GDBSERVER
> -  /* gdbserver does not support PTRACE_O_TRACESYSGOOD.  */
> -#else
> -  int ret;
> +  if ((additional_flags & PTRACE_O_TRACESYSGOOD) != 0)
> +    {

Please avoid this indenting unless necessary, by
reversing the logic:

  if ((additional_flags & PTRACE_O_TRACESYSGOOD) == 0)
     return;

As this will end up being undone once additional_flags is
removed again later on, this way avoid re-indenting back,
and complicating history / git blame.

> +      int ret;
>  
> -  ret = ptrace (PTRACE_SETOPTIONS, child_pid, (PTRACE_TYPE_ARG3) 0,
> -		(PTRACE_TYPE_ARG4) PTRACE_O_TRACESYSGOOD);
> -  if (ret == 0)
> -    current_ptrace_options |= PTRACE_O_TRACESYSGOOD;
> -#endif
> +      ret = ptrace (PTRACE_SETOPTIONS, child_pid, (PTRACE_TYPE_ARG3) 0,
> +		    (PTRACE_TYPE_ARG4) PTRACE_O_TRACESYSGOOD);
> +      if (ret != 0)
> +	additional_flags &= ~PTRACE_O_TRACESYSGOOD;

This isn't right.  The function should still update
current_ptrace_options (the effectively enabled options).

Otherwise ...

> +    }
>  }
>  
>  /* Determine if PTRACE_O_TRACEFORK can be used to follow fork
> @@ -388,16 +391,15 @@ linux_test_for_tracefork (int child_pid)
>    if (ret != 0)
>      return;
>  
> -#ifdef GDBSERVER
> -  /* gdbserver does not support PTRACE_O_TRACEVFORKDONE yet.  */
> -#else
> -  /* Check if the target supports PTRACE_O_TRACEVFORKDONE.  */
> -  ret = ptrace (PTRACE_SETOPTIONS, child_pid, (PTRACE_TYPE_ARG3) 0,
> -		(PTRACE_TYPE_ARG4) (PTRACE_O_TRACEFORK
> -				    | PTRACE_O_TRACEVFORKDONE));
> -  if (ret == 0)
> -    current_ptrace_options |= PTRACE_O_TRACEVFORKDONE;
> -#endif
> +  if ((additional_flags & PTRACE_O_TRACEVFORKDONE) != 0)
> +    {
> +      /* Check if the target supports PTRACE_O_TRACEVFORKDONE.  */
> +      ret = ptrace (PTRACE_SETOPTIONS, child_pid, (PTRACE_TYPE_ARG3) 0,
> +		    (PTRACE_TYPE_ARG4) (PTRACE_O_TRACEFORK
> +					| PTRACE_O_TRACEVFORKDONE));
> +      if (ret != 0)
> +	additional_flags &= ~PTRACE_O_TRACEVFORKDONE;
> +    }
>  
>    /* Setting PTRACE_O_TRACEFORK did not cause an error, however we
>       don't know for sure that the feature is available; old
> @@ -433,18 +435,7 @@ linux_test_for_tracefork (int child_pid)
>  
>  	  /* We got the PID from the grandchild, which means fork
>  	     tracing is supported.  */
> -#ifdef GDBSERVER
> -	  /* Do not enable all the options for now since gdbserver does not
> -	     properly support them.  This restriction will be lifted when
> -	     gdbserver is augmented to support them.  */
> -	  current_ptrace_options |= PTRACE_O_TRACECLONE;
> -#else
> -	  current_ptrace_options |= PTRACE_O_TRACEFORK | PTRACE_O_TRACEVFORK
> -	    | PTRACE_O_TRACECLONE | PTRACE_O_TRACEEXEC;
> -
> -	  /* Do not enable PTRACE_O_TRACEEXIT until GDB is more prepared to
> -	     support read-only process state.  */
> -#endif
> +	  current_ptrace_options |= PTRACE_O_TRACECLONE | additional_flags;

... note how PTRACE_O_TRACESYSGOOD is lost if PTRACE_O_TRACECLONE
isn't supported.

So don't clear additional_flags.  Instead add to current_ptrace_options
as options required in additional_flags are detected to be supported.

-- 
Pedro Alves


  parent reply	other threads:[~2014-07-17 16:35 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-07-16 16:19 [PATCH 00/15 v2] Common code cleanups Gary Benson
2014-07-16 16:19 ` [PATCH 15/15 v2] Finally remove GDBSERVER (mostly) from linux-btrace.c Gary Benson
2014-07-16 16:19 ` [PATCH 04/15 v2] Introduce common-types.h Gary Benson
2014-07-17 11:21   ` Doug Evans
2014-07-16 16:32 ` [PATCH 09/15 v2] Mostly remove GDBSERVER from linux-waitpid.c Gary Benson
2014-07-16 16:33 ` [PATCH 14/15 v2] Introduce get_thread_regcache_for_ptid Gary Benson
2014-07-16 16:45 ` [PATCH 02/15 v2] Remove some GDBSERVER checks from linux-ptrace Gary Benson
2014-07-17  8:37   ` Doug Evans
2014-07-17 16:40   ` Pedro Alves [this message]
2014-07-16 16:45 ` [PATCH 05/15 v2] Introduce and use debug_printf and debug_vprintf Gary Benson
2014-07-16 16:45 ` [PATCH 01/15 v2] Introduce common/errors.h Gary Benson
2014-07-16 18:36   ` Doug Evans
2014-07-17 13:41     ` [PATCH] " Gary Benson
2014-07-17 13:47       ` Gary Benson
2014-07-17 14:05     ` [PATCH 01/15 v3] " Gary Benson
2014-07-17 15:40       ` Pedro Alves
2014-07-17 16:03         ` Gary Benson
2014-07-17 16:19           ` Pedro Alves
2014-07-18  9:20             ` Gary Benson
2014-07-18 10:42               ` Doug Evans
2014-07-18 11:23                 ` Gary Benson
2014-07-18 12:31                   ` Doug Evans
2014-07-18 10:44               ` Pedro Alves
2014-07-16 16:48 ` [PATCH 06/15 v2] Remove simple GDBSERVER uses from common, nat and target Gary Benson
2014-07-16 16:48 ` [PATCH 03/15 v2] Make gdbserver CORE_ADDR unsigned Gary Benson
2014-07-17  9:02   ` Doug Evans
2014-07-17 16:42   ` Pedro Alves
2014-07-18  8:07     ` Maciej W. Rozycki
2014-07-16 17:03 ` [PATCH 11/15 v2] More target unification Gary Benson
2014-07-16 17:03 ` [PATCH 13/15 v2] Finally remove GDBSERVER (mostly) from agent.c Gary Benson
2014-07-16 17:03 ` [PATCH 08/15 v2] Make btrace-common.h not use GDBSERVER Gary Benson
2014-07-16 17:04 ` [PATCH 10/15 v2] Add target/target.h Gary Benson
2014-07-16 17:20 ` [PATCH 12/15 v2] Add target/symbol.h, update users Gary Benson
2014-07-16 17:24 ` [PATCH 07/15 v2] Remove GDBSERVER use from nat/i386-dregs.c Gary Benson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=53C7FB3E.5010402@redhat.com \
    --to=palves@redhat.com \
    --cc=dje@google.com \
    --cc=gbenson@redhat.com \
    --cc=gdb-patches@sourceware.org \
    --cc=tromey@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox