linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: mjw@us.ibm.com
Cc: linuxppc-dev@ozlabs.org, mikey@neuling.org, anton@samba.org
Subject: Re: [PATCH] powerpc/ptrace: remove BUG_ON when full register set not available
Date: Wed, 02 Mar 2011 14:08:36 +1100	[thread overview]
Message-ID: <1299035316.8833.806.camel@pasglop> (raw)
In-Reply-To: <1299011204.28753.8.camel@w500>


> --- ptracev2.orig/include/linux/regset.h	2011-02-20 12:15:57.000000000 -0600
> +++ ptracev2/include/linux/regset.h	2011-02-28 13:33:36.685302349 -0600
> @@ -240,6 +240,34 @@ static inline int user_regset_copyout(un
>  	}
>  	return 0;
>  }
> +/* want the count to be the registers 14-31 inclusive */
> +#define POISON_REG_CNT (PT_R31-PT_R14+1)
> +static inline int user_regset_copyout_poison(unsigned int *pos,
> +					   unsigned int *count,
> +					   void **kbuf, void __user **ubuf,
> +					   const int start_pos,
> +					   const int end_pos)
> +{

I wouldn't put that in a generic location... especially something
like POISON_REG_CNT is very specific to powerpc and our ABI.

.../...

>  static inline int user_regset_copyin(unsigned int *pos, unsigned int *count,
>  				     const void **kbuf,
> --- ptracev2.orig/arch/powerpc/kernel/ptrace.c	2011-02-20 12:15:57.000000000 -0600
> +++ ptracev2/arch/powerpc/kernel/ptrace.c	2011-03-01 13:49:12.686354353 -0600
> @@ -234,11 +234,29 @@ static int gpr_get(struct task_struct *t
>  	if (target->thread.regs == NULL)
>  		return -EIO;
>  
> -	CHECK_FULL_REGS(target->thread.regs);

Wouldn't it be a simpler/cleaner approach to use a single function call
here that checks if the reg set is full and if not, put poison values
in the thread.regs structure itself ?

The resulting code would be more palatable...

Cheers,
Ben.
 
> -	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
> -				  target->thread.regs,
> -				  0, offsetof(struct pt_regs, msr));
> +	if (!FULL_REGS(target->thread.regs)) {
> +                /* Don't have the full register set. Copy out register r0-r13 */
> +		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
> +					  target->thread.regs,
> +					  0, sizeof(long)*PT_R14);
> +
> +		/* Dont want to change the actual register values so rather than read the */
> +		/* actual register copy a poison value into the buffer instead		  */
> +		if (!ret)
> +			ret = user_regset_copyout_poison(&pos, &count, &kbuf, &ubuf,
> +						  sizeof(long)*PT_R14, offsetof(struct pt_regs, nip));
> +
> +		/* Copy out the rest of the registers as usual */
> +		if (!ret)
> +			ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
> +						  target->thread.regs,
> +						  offsetof(struct pt_regs, nip), offsetof(struct pt_regs, msr));
> +		
> +	} else 
> +		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
> +					  target->thread.regs,
> +					  0, offsetof(struct pt_regs, msr));
>  	if (!ret) {
>  		unsigned long msr = get_user_msr(target);
>  		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &msr,
> @@ -645,11 +663,29 @@ static int gpr32_get(struct task_struct 
>  	if (target->thread.regs == NULL)
>  		return -EIO;
>  
> -	CHECK_FULL_REGS(target->thread.regs);
> -
>  	pos /= sizeof(reg);
>  	count /= sizeof(reg);
>  
> +	if(!FULL_REGS(target->thread.regs)) {
> +		if (kbuf) {
> +                	/* Don't have the full register set. Copy out register r0-r13 */
> +			for (; count > 0 && pos < PT_R14; --count)
> +				*k++ = regs[pos++];
> +
> +			/* Dont want to change the actual register values so rather than read the */
> +			/* actual register copy a poison value into the buffer instead		  */
> +			for (; count > 0 && pos <= PT_R31; --count,pos++)
> +				*k++ = 0xdeadbeef;
> +
> +		} else { 
> +			for (; count > 0 && pos < PT_R14; --count)
> +				if (__put_user((compat_ulong_t) regs[pos++], u++))
> +					return -EFAULT;
> +			for (; count > 0 && pos <= PT_R31; --count,pos++)
> +				if (__put_user((compat_ulong_t) 0xdeadbeef, u++))
> +					return -EFAULT;
> +		}
> +	}
>  	if (kbuf)
>  		for (; count > 0 && pos < PT_MSR; --count)
>  			*k++ = regs[pos++];
> 
> 
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/linuxppc-dev

  reply	other threads:[~2011-03-02  3:08 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-03-01 20:26 [PATCH] powerpc/ptrace: remove BUG_ON when full register set not available Michael Wolf
2011-03-02  3:08 ` Benjamin Herrenschmidt [this message]
2011-03-02 15:44   ` Michael Wolf
2011-03-21  0:18 [PATCH] powerpc/ptrace: Remove " Benjamin Herrenschmidt

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=1299035316.8833.806.camel@pasglop \
    --to=benh@kernel.crashing.org \
    --cc=anton@samba.org \
    --cc=linuxppc-dev@ozlabs.org \
    --cc=mikey@neuling.org \
    --cc=mjw@us.ibm.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;
as well as URLs for NNTP newsgroup(s).