why clear_buffer_uptodate in end_buffer_write_sync on write error?

why clear_buffer_uptodate in end_buffer_write_sync on write error?

[Robert W. Fuller]

I'm trying to understand the file systems.  I'm making good progress, 
but I don't get this.  If the buffer is marked not up to date because 
the write failed, won't this cause block_prepare_write to try and read 
the buffer using ll_rw_block thereby overwriting the data that couldn't 
be written?  Is this desirable?  Perhaps I'm confused?

Re: why clear_buffer_uptodate in end_buffer_write_sync on write error?

[Robert W. Fuller]

Hmm, ok this makes sense for the sync case, but what about the async 
case?  Should clear_buffer_uptodate be called in end_buffer_async_write? 
  How is the error going to be reported to the user, particularly in the 
case of a memory mapped file?  Shouldn't the system try to write that 
page again?

Robert W. Fuller wrote:
> I'm trying to understand the file systems.  I'm making good progress, 
> but I don't get this.  If the buffer is marked not up to date because 
> the write failed, won't this cause block_prepare_write to try and read 
> the buffer using ll_rw_block thereby overwriting the data that couldn't 
> be written?  Is this desirable?  Perhaps I'm confused?
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>