[CRIU] [PATCH] page-xfer: receive information about presence of a parent from a page server

[Pavel Emelyanov]

On 11/07/2014 02:18 AM, Andrey Vagin wrote:
> We need to know this to insert holes. Currently xfer->parent isn't
> initialized for remote sessions. In most cases it has a non-zero value,
> so generate_iovs() is called with has_parent = true.

This worth being fixed, yes :)

> 
> bash test/zdtm.sh -p -P -i 3 ns/transition/fork
> (00.000106) Error (sysctl.c:194): Can't open sysctl net/ipv4/tcp_wmem: No such file or directory
> (00.017048)    420: Error (image.c:231): Unable to open pagemap-420.img: No such file or directory
> (00.017065)    420: Error (image.c:231): Unable to open pages-420.img: No such file or directory
> (00.017090)    420: Error (page-read.c:73): No parent for snapshot pagemap
> (00.017290)     86: Error (cr-restore.c:1185): 420 exited, status=1

Wait a second, this is restore-time error, not dump time. Why do we fix
it in dump code? The way I see it now -- we should set has_parent in dump
to true, provided we do page-server dump and the parent link exists. If
we meet a "hole" on dump side and send the hole to server, it's the page
server who should emit an error if it doesn't have the parent images.

> (00.017317) Error (cr-restore.c:1831): Restoring FAILED.
> 
> Reported-by: Mr Jenkins
> Signed-off-by: Andrey Vagin <avagin at openvz.org>
> ---
>  page-xfer.c | 28 +++++++++++++++++++++++++---
>  1 file changed, 25 insertions(+), 3 deletions(-)
> 
> diff --git a/page-xfer.c b/page-xfer.c
> index 86e26d8..c3af4a5 100644
> --- a/page-xfer.c
> +++ b/page-xfer.c
> @@ -79,7 +79,8 @@ static void page_server_close(void)
>  		cxfer.loc_xfer.close(&cxfer.loc_xfer);
>  }
>  
> -static int page_server_open(struct page_server_iov *pi)
> +static void close_page_xfer(struct page_xfer *xfer);
> +static int page_server_open(int sk, struct page_server_iov *pi)
>  {
>  	int type;
>  	long id;
> @@ -94,6 +95,17 @@ static int page_server_open(struct page_server_iov *pi)
>  		return -1;
>  
>  	cxfer.dst_id = pi->dst_id;
> +
> +	if (sk >= 0) {
> +		char has_parent = !!cxfer.loc_xfer.parent;
> +
> +		if (write(sk, &has_parent, 1) != 1) {
> +			pr_perror("Unable to send reponse");
> +			close_page_xfer(&cxfer.loc_xfer);
> +			return -1;
> +		}
> +	}
> +
>  	return 0;
>  }
>  
> @@ -101,7 +113,7 @@ static int prep_loc_xfer(struct page_server_iov *pi)
>  {
>  	if (cxfer.dst_id != pi->dst_id) {
>  		pr_warn("Deprecated IO w/o open\n");
> -		return page_server_open(pi);
> +		return page_server_open(-1, pi);
>  	} else
>  		return 0;
>  }
> @@ -192,7 +204,7 @@ static int page_server_serve(int sk)
>  
>  		switch (pi.cmd) {
>  		case PS_IOV_OPEN:
> -			ret = page_server_open(&pi);
> +			ret = page_server_open(sk, &pi);
>  			break;
>  		case PS_IOV_ADD:
>  			ret = page_server_add(sk, &pi);
> @@ -482,6 +494,7 @@ static void close_server_xfer(struct page_xfer *xfer)
>  static int open_page_server_xfer(struct page_xfer *xfer, int fd_type, long id)
>  {
>  	struct page_server_iov pi;
> +	char has_parent;
>  
>  	xfer->sk = page_server_sk;
>  	xfer->write_pagemap = write_pagemap_to_server;
> @@ -489,6 +502,7 @@ static int open_page_server_xfer(struct page_xfer *xfer, int fd_type, long id)
>  	xfer->write_hole = write_hole_to_server;
>  	xfer->close = close_server_xfer;
>  	xfer->dst_id = encode_pm_id(fd_type, id);
> +	xfer->parent = NULL;
>  
>  	pi.cmd = PS_IOV_OPEN;
>  	pi.dst_id = xfer->dst_id;
> @@ -500,6 +514,14 @@ static int open_page_server_xfer(struct page_xfer *xfer, int fd_type, long id)
>  		return -1;
>  	}
>  
> +	if (read(xfer->sk, &has_parent, 1) != 1) {
> +		pr_perror("The page server doesn't answer");
> +		return -1;
> +	}
> +
> +	if (has_parent)
> +		xfer->parent = (void *) 1; /* This is required for generate_iovs() */
> +
>  	return 0;
>  }
>  
>