> I pass in an uninitialized pointer. The code crashes in memcpy as
> called from io_read. I will send a full bug report my next mail run
That is your bug. There is no need to report anything about it.
___
Bug-hurd mailing list
[EMAIL PROTECTED]
htt
On Wed, Mar 07, 2001 at 08:39:39AM -0500, Roland McGrath wrote:
> > Alright, then way does it crash when I use a store backed by a file
> > and not when I use a store back by a device? Further more, why does
> > it work when I use my patch?
>
> Are you talking about your own code that calls stor
> Alright, then way does it crash when I use a store backed by a file
> and not when I use a store back by a device? Further more, why does
> it work when I use my patch?
Are you talking about your own code that calls store_read?
It might be that in the two particular things you are trying, one
On Wed, Mar 07, 2001 at 08:23:55AM -0500, Roland McGrath wrote:
> You have misunderstood the calling conventions.
>
> io_read is an RPC stub. The store_read_meth_t functions, and store_read,
> obey the usual buffer-passing conventions for data_t arguments in RPCs.
Alright, then way does it cr
You have misunderstood the calling conventions.
io_read is an RPC stub. The store_read_meth_t functions, and store_read,
obey the usual buffer-passing conventions for data_t arguments in RPCs.
___
Bug-hurd mailing list
[EMAIL PROTECTED]
http://mail.
It appears to me that io_read requires a malloced buffer. According
to store semantics, only a pointer to a pointer to a buffer is passed.
Thus, those programs using libstore and the file class crash violently.
The attached patch fixes this problem, however, I do not know if this
is the correct p
