[PATCH v3] Retry Large Buffer Allocations

[Artem Bityutskiy]

On Wed, 2011-04-06 at 08:56 -0700, Grant Erickson wrote:
> When handling user space read or write requests via mtd_{read,write}
> or JFFS2 medium scan requests, exponentially back off on the size of
> the requested kernel transfer buffer until it succeeds or until the
> requested transfer buffer size falls below the page size.
> 
> This helps ensure the operation can succeed under low-memory,
> highly-fragmented situations albeit somewhat more slowly.
> 
>   v2: Incorporated coding style and comment feedback from Artem.
>   v3: Incorporated more feedback from Artem. Retargeted patch against
>       l2-mtd-2.6.
> 
> Signed-off-by: Grant Erickson <marathon96 at gmail.com>

Looks good, but I have 2 requests still.

> ---
>  drivers/mtd/mtdchar.c   |   50 +++++++++++++++++++++-------------------------
>  drivers/mtd/mtdcore.c   |   41 ++++++++++++++++++++++++++++++++++++++
>  fs/jffs2/scan.c         |   11 +++++----
>  include/linux/mtd/mtd.h |    2 +
>  4 files changed, 72 insertions(+), 32 deletions(-)
> 
> diff --git a/drivers/mtd/mtdchar.c b/drivers/mtd/mtdchar.c
> index 145b3d0d..8651a79 100644
> --- a/drivers/mtd/mtdchar.c
> +++ b/drivers/mtd/mtdchar.c
> @@ -166,10 +166,23 @@ static int mtd_close(struct inode *inode, struct file *file)
>  	return 0;
>  } /* mtd_close */
>  
> -/* FIXME: This _really_ needs to die. In 2.5, we should lock the
> -   userspace buffer down and use it directly with readv/writev.
> -*/
> -#define MAX_KMALLOC_SIZE 0x20000
> +/* Back in April 2005, Linus wrote:
> + *
> + *   FIXME: This _really_ needs to die. In 2.5, we should lock the
> + *   userspace buffer down and use it directly with readv/writev.
> + *
> + * The implementation below, using mtd_alloc_up_to, mitigates

1. s/mtd_alloc_up_to/mtd_kmalloc_up_to/ here as well.

BTW, It still does not apply on top of l2-mtd-2.6.git, the conflict is
in include/linux/mtd/mtd.h - but well, I can fix that conflict.

2. Please, fix compilation warning:

fs/jffs2/scan.c: In function ‘jffs2_scan_medium’:
fs/jffs2/scan.c:123: warning: passing argument 1 of ‘mtd_kmalloc_up_to’ from incompatible pointer type
include/linux/mtd/mtd.h:351: note: expected ‘size_t *’ but argument is of type ‘uint32_t *’

This is a bug in 64-bit machines because sizeof(size_t) is 8 there.

Could you please fix that, test, and then send the final version.

Thanks!

-- 
Best Regards,
Artem Bityutskiy (Артём Битюцкий)