Jiri Kosina wrote:
> On Sun, 2 Mar 2008, Jiri Kosina wrote:
> 
>> Could you please try what is the output with the attached patch below, so 
>> that we know which spinlock buggers?
> 
> Uhm, and hereby I attach the patch.
> 
> diff --git a/ipc/shm.c b/ipc/shm.c
> index c47e872..2abfb70 100644
> --- a/ipc/shm.c
> +++ b/ipc/shm.c
> @@ -615,6 +615,7 @@ static void shm_get_stat(struct ipc_namespace *ns, unsigned long *rss,
>  			*rss += (HPAGE_SIZE/PAGE_SIZE)*mapping->nrpages;
>  		} else {
>  			struct shmem_inode_info *info = SHMEM_I(inode);
> +			printk(KERN_DEBUG "info: %p\n", info);
>  			spin_lock(&info->lock);
>  			*rss += inode->i_mapping->nrpages;
>  			*swp += info->swapped;
> diff --git a/ipc/util.c b/ipc/util.c
> index fd1b50d..5a0d4d2 100644
> --- a/ipc/util.c
> +++ b/ipc/util.c
> @@ -662,6 +662,7 @@ struct kern_ipc_perm *ipc_lock(struct ipc_ids *ids, int id)
>  
>  	up_read(&ids->rw_mutex);
>  
> +	printk(KERN_DEBUG "out: %p\n", out);
>  	spin_lock(&out->lock);
>  	
>  	/* ipc_rmid() may have already freed the ID while ipc_lock

I'll try the patch, so far I attached the ipcs strace and adjacent kernel logs.