mail archive of the barebox mailing list
 help / color / mirror / Atom feed
From: Sascha Hauer <s.hauer@pengutronix.de>
To: Jules Maselbas <jmaselbas@kalray.eu>
Cc: barebox@lists.infradead.org
Subject: Re: [PATCH 2/3] usb: gadget: dfu: Progressive erase if file is a mtd
Date: Thu, 26 Mar 2020 07:03:11 +0100	[thread overview]
Message-ID: <20200326060311.GI27288@pengutronix.de> (raw)
In-Reply-To: <20200324154647.17341-3-jmaselbas@kalray.eu>

Hi Jules,

On Tue, Mar 24, 2020 at 04:46:46PM +0100, Jules Maselbas wrote:
> When downloading a firmware into a big flash partition the erase
> operation can take a long time to be complete from few seconds to
> minutes in extreme cases. During the erase the DFU gadget does not
> respond to any USB setup request, the host only see a stalled USB
> endpoint and cannot get responses from DFU_GETSTATE nor DFU_GETSTATUS.
> 
> After 5 seconds without any updates the host will abort the DFU
> transfer and return an error (when using dfu-util).
> 
> For instance I have a 2MB partition that takes 25 seconds to erase,
> this erase cannot be done in one step as it takes too much time or it
> should be done in the manifestation phase, see previous patch.
> 
> This patch modify the erase behavior when downloading a new firmware.
> If the updated file is a mtd partition then the DFU gadget will do
> a progressive erase by erasing the least amount of required blocks
> before writing a new chunk into the mtd device.
> 
> Signed-off-by: Jules Maselbas <jmaselbas@kalray.eu>
> ---
>  drivers/usb/gadget/dfu.c | 21 ++++++++++++++++++++-
>  1 file changed, 20 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/usb/gadget/dfu.c b/drivers/usb/gadget/dfu.c
> index fef6f3e3a..592586db1 100644
> --- a/drivers/usb/gadget/dfu.c
> +++ b/drivers/usb/gadget/dfu.c
> @@ -55,6 +55,8 @@
>  #include <libbb.h>
>  #include <init.h>
>  #include <fs.h>
> +#include <ioctl.h>
> +#include <linux/mtd/mtd-abi.h>
>  
>  #define USB_DT_DFU			0x21
>  
> @@ -132,6 +134,10 @@ struct file_list_entry *dfu_file_entry;
>  static int dfufd = -EINVAL;
>  static struct file_list *dfu_files;
>  static int dfudetach;
> +static struct mtd_info_user dfu_mtdinfo;
> +static loff_t dfu_written;
> +static loff_t dfu_erased;
> +static int prog_erase;

These are not initialized, this works for the first file updated via
DFU, but what about the following ones?

>  
>  /* USB DFU functional descriptor */
>  static struct usb_dfu_func_descriptor usb_dfu_func = {
> @@ -327,8 +333,16 @@ static void dfu_cleanup(struct f_dfu *dfu)
>  static void dn_complete(struct usb_ep *ep, struct usb_request *req)
>  {
>  	struct f_dfu		*dfu = req->context;
> +	loff_t size;
>  	int ret;
>  
> +	if (prog_erase && (dfu_written + req->length) > dfu_erased) {
> +		size = roundup(req->length, dfu_mtdinfo.erasesize);
> +		erase(dfufd, size, dfu_erased);

You should check the return value here.

> +		dfu_erased += size;
> +	}
> +
> +	dfu_written += req->length;
>  	ret = write(dfufd, req->buf, req->length);
>  	if (ret < (int)req->length) {
>  		perror("write");
> @@ -493,7 +507,12 @@ static int dfu_setup(struct usb_function *f, const struct usb_ctrlrequest *ctrl)
>  			}
>  
>  			if (!(dfu_file_entry->flags & FILE_LIST_FLAG_SAFE)) {
> -				ret = erase(dfufd, ERASE_SIZE_ALL, 0);
> +				ret = ioctl(dfufd, MEMGETINFO, &dfu_mtdinfo);
> +				if (ret) /* not a mtd */
> +					ret = erase(dfufd, ERASE_SIZE_ALL, 0);
> +				else
> +					prog_erase = 1;

I am not aware of any non mtd devices that need erase. I think you can
drop the full erase here.

Sascha

-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

  reply	other threads:[~2020-03-26  6:03 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-24 15:46 [PATCH 0/3] usb: gadget: dfu: Fix timeout on erase when using big partition Jules Maselbas
2020-03-24 15:46 ` [PATCH 1/3] usb: gadget: dfu: Add manifestation phase Jules Maselbas
2020-03-24 15:46 ` [PATCH 2/3] usb: gadget: dfu: Progressive erase if file is a mtd Jules Maselbas
2020-03-26  6:03   ` Sascha Hauer [this message]
2020-03-26  9:08     ` Jules Maselbas
2020-03-24 15:46 ` [PATCH 3/3] usb: gadget: dfu: Reset global variables on unbind Jules Maselbas
2020-03-26  6:09   ` Sascha Hauer

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200326060311.GI27288@pengutronix.de \
    --to=s.hauer@pengutronix.de \
    --cc=barebox@lists.infradead.org \
    --cc=jmaselbas@kalray.eu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox