[Devel] Re: [PATCH] Initialize the sem->sem_pending list after restoring

[Oren Laadan]

Good catch - applied.

Oren.

On 06/16/2010 12:54 PM, Dan Smith wrote:
> Right now, we leave (effectively) garbage in it after restart.  Not only
> must this be re-initialized (since we know it was empty on checkpoint), but
> it must be initialized after being copied into the final resting place.
> 
> Without this, you get lots of fun crashes in various places :)
> 
> Signed-off-by: Dan Smith <danms at us.ibm.com>
> ---
>  ipc/checkpoint_sem.c |    5 ++++-
>  1 files changed, 4 insertions(+), 1 deletions(-)
> 
> diff --git a/ipc/checkpoint_sem.c b/ipc/checkpoint_sem.c
> index ab1830e..43e3cdf 100644
> --- a/ipc/checkpoint_sem.c
> +++ b/ipc/checkpoint_sem.c
> @@ -171,7 +171,7 @@ int restore_ipc_sem(struct ckpt_ctx *ctx, struct ipc_namespace *ns)
>  	struct sem_array *sem;
>  	struct sem *sma = NULL;
>  	struct ipc_ids *sem_ids = &ns->ids[IPC_SEM_IDS];
> -	int semflag, ret;
> +	int semflag, ret, i;
>  
>  	h = ckpt_read_obj_type(ctx, sizeof(*h), CKPT_HDR_IPC_SEM);
>  	if (IS_ERR(h))
> @@ -222,6 +222,9 @@ int restore_ipc_sem(struct ckpt_ctx *ctx, struct ipc_namespace *ns)
>  	sem = container_of(ipc, struct sem_array, sem_perm);
>  	memcpy(sem->sem_base, sma, sem->sem_nsems * sizeof(*sma));
>  
> +	for (i = 0; i < sem->sem_nsems; i++)
> +		INIT_LIST_HEAD(&sem->sem_base[i].sem_pending);
> +
>  	/* this is safe because no unauthorized access is possible */
>  	ipc_unlock(ipc);
>  
_______________________________________________
Containers mailing list
Containers at lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers