Definition conflict in 2.4.19-pre?? code

Definition conflict in 2.4.19-pre?? code

[Roger W. Brown]

  Hi,

      I am not able to compile 2.4 series kernels after
  linux-2.4.19-pre1 and I don't understand how others can !

  Consider the "struct request_queue" definition in blkdev.h

  In the 2.4.19-pre1 version, the last few lines read:
        /*
         * Tasks wait here for free read and write requests
         */
        wait_queue_head_t       wait_for_requests[2];
  };
  and for later versions this is changed to:
        /*
         * Tasks wait here for free request
         */
        wait_queue_head_t       wait_for_request;
  };

  yet drivers/block/ll_rw_blk.c still makes references to
  wait_for_requests[?]  in the void blk_init_free_list()
  and blkdev_release_request() functions and elsewhere.

  Roger

-- 

Re: Definition conflict in 2.4.19-pre?? code

[Jens Axboe]

On Wed, Jun 05 2002, Roger W. Brown wrote:
> 
>   Hi,
> 
>       I am not able to compile 2.4 series kernels after
>   linux-2.4.19-pre1 and I don't understand how others can !
> 
>   Consider the "struct request_queue" definition in blkdev.h
> 
>   In the 2.4.19-pre1 version, the last few lines read:
>         /*
>          * Tasks wait here for free read and write requests
>          */
>         wait_queue_head_t       wait_for_requests[2];
>   };
>   and for later versions this is changed to:
>         /*
>          * Tasks wait here for free request
>          */
>         wait_queue_head_t       wait_for_request;
>   };
> 
>   yet drivers/block/ll_rw_blk.c still makes references to
>   wait_for_requests[?]  in the void blk_init_free_list()
>   and blkdev_release_request() functions and elsewhere.

Your kernel tree must be corrupted, there's no such change in later
2.4.19-pre. pre10 still uses two request wait queues.

-- 
Jens Axboe

Re: Definition conflict in 2.4.19-pre?? code

[Roger W. Brown]

                  (  Original message deleted )
>
> Your kernel tree must be corrupted, there's no such change in later
> 2.4.19-pre. pre10 still uses two request wait queues.

  Yes indeed !!   My patch program trips and falls whenever it encounters the    
line: "\ No newline at end of file"   I hadn't noticed.  Sorry.

-- 
Roger Brown