On Mon, Aug 06, 2018 at 12:01:31PM +1000, Stephen Rothwell wrote:
> Hi all,
>
> Today's linux-next merge of the rdma tree got a conflict in:
>
> drivers/infiniband/core/uverbs_main.c
>
> between commit:
>
> 1eb9364ce81d ("IB/uverbs: Fix ordering of ucontext check in
> ib_uverbs_write")
>
> from Linus' tree and commit:
>
> a9b66d6453d7 ("IB/uverbs: Do not block disassociate during write()")
>
> from the rdma tree.
>
> I fixed it up (I am not entirely sure this is correct, but see below) and
> can carry the fix as necessary. This is now fixed as far as linux-next
> is concerned, but any non trivial conflicts should be mentioned to your
> upstream maintainer when your tree is submitted for merging. You may
> also want to consider cooperating with the maintainer of the conflicting
> tree to minimise any particularly complex conflicts.
> diff --cc drivers/infiniband/core/uverbs_main.c
> index 2094d136513d,6f62146e9738..000000000000
> --- a/drivers/infiniband/core/uverbs_main.c
> +++ b/drivers/infiniband/core/uverbs_main.c
> @@@ -748,24 -758,8 +758,18 @@@ static ssize_t ib_uverbs_write(struct f
> return ret;
>
> srcu_key = srcu_read_lock(&file->device->disassociate_srcu);
> - ib_dev = srcu_dereference(file->device->ib_dev,
> - &file->device->disassociate_srcu);
> - if (!ib_dev) {
> - ret = -EIO;
> - goto out;
> - }
>
> + /*
> + * Must be after the ib_dev check, as once the RCU clears ib_dev ==
> + * NULL means ucontext == NULL
> + */
> + if (!file->ucontext &&
> + (command != IB_USER_VERBS_CMD_GET_CONTEXT || extended)) {
> + ret = -EINVAL;
> + goto out;
> + }
> +
> - if (!verify_command_mask(ib_dev, command, extended)) {
> + if (!verify_command_mask(file, command, extended)) {
> ret = -EOPNOTSUPP;
> goto out;
> }
The resolution should end up like this:
srcu_key = srcu_read_lock(&file->device->disassociate_srcu);
if (!verify_command_mask(file, command, extended)) {
ret = -EOPNOTSUPP;
goto out;
}
(ie take the hunk from for-next, discard the for-rc stuff)
As commit 22fa27fbc64d01cbbe1e4da751e64cc22d24a6e4 also removed the
'file->ucontext' stuff
Thanks,
Jason
