Re: random crashes on -HEAD for a few days now - Mailing list pgsql-hackers

Stefan Kaltenbrunner <stefan@kaltenbrunner.cc> writes:
> as for len it seems to be 0:

> #0  varbit_out (fcinfo=0x88c75000) at varbit.c:549
>          s = (VarBit *) 0x88c75000
>          result = 0x84d33128 ""
>          r = 0x84d33128 ""
>          sp = (bits8 *) 0x88c75000 <Address 0x88c75000 out of bounds>
>          x = 0 '\0'
>          i = 0
>          k = 0
>          len = 0

Hmm ... s and sp really shouldn't be equal, nor equal to fcinfo, but
it's likely that the compiler optimized them into the same register.
We need to confirm what actually got passed as the argument.  Could you
go to frame 1 and see what is in its local fcinfo var, in particular see
what args[0] is?  I'm betting it's 0x88c75000 minus 8 ... if so, look at
what is in those last 8 bytes.  If that's int32 8 followed by int32 0,
then indeed we have a zero-length bitstring at the end of memory.
        regards, tom lane