[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Chicken-hackers] [PATCH] fix stack-overflow check
From: |
Felix |
Subject: |
Re: [Chicken-hackers] [PATCH] fix stack-overflow check |
Date: |
Mon, 13 Aug 2012 14:19:31 +0200 (CEST) |
From: Christian Kellermann <address@hidden>
Subject: Re: [Chicken-hackers] [PATCH] fix stack-overflow check
Date: Mon, 13 Aug 2012 13:04:39 +0200
> * felix winkelmann <address@hidden> [120813 12:34]:
>> The attached patch fixes the computation of stack depth beyond
>> C_stack_limit. The old way of calculating the depth could overflow if
>> the stack is located very high in memory. This was causing incorrect
>> stack-overflow errors on some 64-bit systems (sparc64 and x86-64). The
>> new method checks whether the stack-pointer is beyond the limit
>> computed at startup, including a 64k "slack" area to allow (modestly)
>> deep recursions in direct-call-optimized (simple, non-allocating, not
>> externally visible) procedures, if the nursery is already mostly
>> exhausted.
>>
>> This patch fixes ticket #895.
>>
>> I recommend to include this fix in the upcoming 4.8.0 release (this will
>> require a new release candidate).
>
> A minor nitpick from my side....
>
>
>> +#define C_stack_overflow_check C_stack_check1(C_stack_overflow())
>> +
>> +/*XXX OBSOLETE */
>> +#define C_stack_check C_stack_overflow_check
>> +
>
> According to "git grep C_stack_check" the only places where that
> string is used is in the places where you replace it. Why not remove
> it? Is it part of the API?
It is part of the API used by already existing C files, generated by
earlier versions of "chicken". I don't know whether this is a real
problem in practice, but it avoids requiring a recompile.
cheers,
felix