[PATCH] SH simulator illegal instructions

[PATCH] SH simulator illegal instructions

[Andrew STUBBS]

[-- Attachment #1: Type: text/plain, Size: 415 bytes --]

Hi,

The SH simulator currently accepts some instructions in delay slots that 
should not be allowed.

Specifically:
	mov.l @(<disp>,PC),<REG_N>
	mov.w @(<disp>,PC),<REG_N>
	mova @(<disp>,PC),R0

I have checked the remaining SH-4 instructions, and don't believe there 
are any more, however I have not checked other architecture variants.

The attached patch add the proper check.

:ADDPATCH sh sim:

Andrew Stubbs

[-- Attachment #2: illslot.patch --]
[-- Type: text/plain, Size: 937 bytes --]

2007-09-24  Andrew Stubbs  <andrew.stubbs@st.com>

	* gencode.c (tab): Add RAISE_EXCEPTION_IF_IN_DELAY_SLOT to the
	definition of PC relative 'mov.l'/'mov.w' and also 'mova'.

--- orig/sim/sh/gencode.c	2007-03-06 13:58:46.000000000 +0000
+++ src/sim/sh/gencode.c	2007-09-24 15:02:18.000000000 +0100
@@ -993,6 +993,7 @@
     "L (0);",
   },
   { "n", "", "mov.l @(<disp>,PC),<REG_N>", "1101nnnni8p4....",
+    "RAISE_EXCEPTION_IF_IN_DELAY_SLOT ();",
     "MA (1);",
     "R[n] = RLAT ((PH2T (PC) & ~3) + 4 + i);",
     "L (n);",
@@ -1059,6 +1060,7 @@
     "L (0);",
   },
   { "n", "", "mov.w @(<disp>,PC),<REG_N>", "1001nnnni8p2....",
+    "RAISE_EXCEPTION_IF_IN_DELAY_SLOT ();",
     "MA (1);",
     "R[n] = RSWAT (PH2T (PC + 4 + i));",
     "L (n);",
@@ -1120,6 +1122,7 @@
   },
 
   { "0", "", "mova @(<disp>,PC),R0", "11000111i8p4....",
+    "RAISE_EXCEPTION_IF_IN_DELAY_SLOT ();",
     "R0 = ((i + 4 + PH2T (PC)) & ~0x3);",
   },
 

Re: [PATCH] SH simulator illegal instructions

[Daniel Jacobowitz]

On Mon, Sep 24, 2007 at 03:44:35PM +0100, Andrew STUBBS wrote:
> 2007-09-24  Andrew Stubbs  <andrew.stubbs@st.com>
> 
> 	* gencode.c (tab): Add RAISE_EXCEPTION_IF_IN_DELAY_SLOT to the
> 	definition of PC relative 'mov.l'/'mov.w' and also 'mova'.

:REVIEWMAIL:

Since no one has raised a technical objection, I'll trust your
judgement :-)  Ok to commit, thanks.

-- 
Daniel Jacobowitz
CodeSourcery

Re: [PATCH] SH simulator illegal instructions

[Andrew STUBBS]

Daniel Jacobowitz wrote:
> On Mon, Sep 24, 2007 at 03:44:35PM +0100, Andrew STUBBS wrote:
>> 2007-09-24  Andrew Stubbs  <andrew.stubbs@st.com>
>>
>> 	* gencode.c (tab): Add RAISE_EXCEPTION_IF_IN_DELAY_SLOT to the
>> 	definition of PC relative 'mov.l'/'mov.w' and also 'mova'.
> 
> :REVIEWMAIL:
> 
> Since no one has raised a technical objection, I'll trust your
> judgement :-)  Ok to commit, thanks.
> 

Thanks, but I can't! :(

Assuming there's no problem with sourceware.org, I'm probably having 
trouble with the corporate firewall (again).

Would somebody be so good as to check in this patch on my behalf?

Thanks

Andrew

Re: [PATCH] SH simulator illegal instructions

[Denis PILAT]

Andrew STUBBS wrote:
> Daniel Jacobowitz wrote:
>> On Mon, Sep 24, 2007 at 03:44:35PM +0100, Andrew STUBBS wrote:
>>> 2007-09-24  Andrew Stubbs  <andrew.stubbs@st.com>
>>>
>>>     * gencode.c (tab): Add RAISE_EXCEPTION_IF_IN_DELAY_SLOT to the
>>>     definition of PC relative 'mov.l'/'mov.w' and also 'mova'.
>>
>> :REVIEWMAIL:
>>
>> Since no one has raised a technical objection, I'll trust your
>> judgement :-)  Ok to commit, thanks.
>>
>
> Thanks, but I can't! :(
>
> Assuming there's no problem with sourceware.org, I'm probably having 
> trouble with the corporate firewall (again).
>
> Would somebody be so good as to check in this patch on my behalf?
>
> Thanks
>
> Andrew
>
That's done !
That means it's not the corporate firewall.
-- 
Denis

Re: [PATCH] SH simulator illegal instructions

[Andrew STUBBS]

Denis PILAT wrote:
> Andrew STUBBS wrote:
>> Would somebody be so good as to check in this patch on my behalf?
>>
> That's done !
> That means it's not the corporate firewall.

Thanks Denis.

After a little investigation, it would appear to be just Bristol that is 
cut off! :(