Re: [PATCH][RESEND] SCSI, Brocade FC HBA: Remember to always release_firmware() so we don't leak memory.

[Rolf Eike Beer <eike@sf-mail.de>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="US-ASCII", Size: 1335 bytes --]

Am Donnerstag, 24. März 2011, 00:00:39 schrieb Jesper Juhl:
> Once we've called request_firmware() we must remember to call
> release_firmware() to free memory. We don't currently do this in
> bfad_read_firmware(); causing a memory leak.
> 
> Signed-off-by: Jesper Juhl <jj@chaosbits.net>
> Acked-by: Jing Huang <huangj@brocade.com>
> ---
>  bfad.c |   11 +++++------
>  1 file changed, 5 insertions(+), 6 deletions(-)
> 
>  Could someone merge this please?
> 
> diff --git a/drivers/scsi/bfa/bfad.c b/drivers/scsi/bfa/bfad.c
> index 44524cf..d7bafeb 100644
> --- a/drivers/scsi/bfa/bfad.c
> +++ b/drivers/scsi/bfa/bfad.c
> @@ -1558,23 +1558,22 @@ bfad_read_firmware(struct pci_dev *pdev, u32
> **bfi_image,
> 
>  	if (request_firmware(&fw, fw_name, &pdev->dev)) {
>  		printk(KERN_ALERT "Can't locate firmware %s\n", fw_name);
> -		goto error;
> +		*bfi_image = NULL;
> +		goto out;
>  	}

A simple "return NULL;" here is enough, there is nothing that could be freed 
later on.

Looking a bit deeper I think the interface of this function is totally b0rked:
-it has a return value that is always the same as one of the arguments 
(+dereference)
-noone ever checks this return value
-at least in my tree it is never called from anywhere outside this file but is 
still exported and not static

Eike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]