Can you please change the walk function to check if fs->root is NULL.
I don't want to spill the knowledge that fs has a field named fd
outside fsys.c. Some day (hopefully soon) we'll have transports other
than file descriptors/sockets and changing only fsys.c will make the
redesign easier.
Thanks,
Lucho
On Sep 24, 2008, at 1:53 PM, Hugh Greenberg wrote:
Yes, it does. However, in walk the line:
nfid = fs->root->fid
takes place before spc_rpc is called and this causes a segfault.
On Wed, 2008-09-24 at 13:41 -0600, Latchesar Ionkov wrote:
But walk calls spc_rpc too, doesn't it?
On Sep 24, 2008, at 1:43 PM, Hugh Greenberg wrote:
Ok, here is a patch that puts the fix in spc_rpc and spc_walk. I
need
the fix in spc_walk also since spc_open calls spc_walk before it
calls
spc_rpc.
Signed-off-by: Hugh Greenberg <[EMAIL PROTECTED]>
On Wed, 2008-09-24 at 12:18 -0600, Latchesar Ionkov wrote:
It doesn't seem like the right solution to me. I think it the fix
should be in spc_rpc.
Thanks,
Lucho
On Sep 24, 2008, at 12:18 PM, Hugh Greenberg wrote:
This patch prevents a segfault caused by spc_open and spc_clunk if
the
filesystem was previously disconnected by spc_disconnect_fsys .
Signed-off-by: Hugh Greenberg <[EMAIL PROTECTED]>
<open.patch>
<all.patch>
