[PATCH v2] bootm: update os_entry from uimage

[Sascha Hauer]

On Mon, Sep 23, 2013 at 10:12:43AM +0200, Jean-Christophe PLAGNIOL-VILLARD wrote:
> On 10:07 Mon 23 Sep     , Jean-Christophe PLAGNIOL-VILLARD wrote:
> > On 08:33 Mon 23 Sep     , Sascha Hauer wrote:
> > > On Mon, Sep 23, 2013 at 06:40:25AM +0200, Jean-Christophe PLAGNIOL-VILLARD wrote:
> > > > On 20:17 Sun 22 Sep     , Sascha Hauer wrote:
> > > > >  common/bootm.c | 2 ++
> > > > >  1 file changed, 2 insertions(+)
> > > > > 
> > > > > diff --git a/common/bootm.c b/common/bootm.c
> > > > > index 1987a09..4e62006 100644
> > > > > --- a/common/bootm.c
> > > > > +++ b/common/bootm.c
> > > > > @@ -73,6 +73,8 @@ static int bootm_open_os_uimage(struct image_data *data)
> > > > >  	if (data->os_address == UIMAGE_SOME_ADDRESS)
> > > > >  		data->os_address = data->os->header.ih_load;
> > > > >  
> > > > > +	data->os_entry = data->os->header.ih_ep - data->os->header.ih_load;
> > > > > +
> > > > >  	if (data->os_address != UIMAGE_INVALID_ADDRESS) {
> > > > >  		data->os_res = uimage_load_to_sdram(data->os, 0,
> > > > >  				data->os_address);
> > > > 
> > > > this does not work on INVALID_ADDRESS as this time teh ih_load is as
> > > > 0xffffffff and the ih_ep is relative not absolute
> > > 
> > > Right. How about this one then?
> > > 
> > > 
> > > From da3192d5c763361ec08a51ebaa3cc2c3884c9827 Mon Sep 17 00:00:00 2001
> > > From: Sascha Hauer <s.hauer at pengutronix.de>
> > > Date: Sun, 22 Sep 2013 19:51:53 +0200
> > > Subject: [PATCH] bootm: handle image entry point correctly
> > > 
> > > This fixes starting of uImages which have an entry point other than
> > > the load address.
> > > 
> > > The uImage format stores the entry point to the image in the ih_ep
> > > field. Normally it has the absolute address, so calculate the relative
> > > offset to the image start address and put it into data->os_entry. In
> > > case the uImage has a load address of 0xffffffff (UIMAGE_INVALID_ADDRESS)
> > > the entry point is stored relative to the image start, so just transfer
> > > it to data->os_address.
> > > 
> > > Signed-off-by: Sascha Hauer <s.hauer at pengutronix.de>
> > > ---
> > >  common/bootm.c | 6 +++++-
> > >  1 file changed, 5 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/common/bootm.c b/common/bootm.c
> > > index 1987a09..16d3416 100644
> > > --- a/common/bootm.c
> > > +++ b/common/bootm.c
> > > @@ -73,7 +73,11 @@ static int bootm_open_os_uimage(struct image_data *data)
> > >  	if (data->os_address == UIMAGE_SOME_ADDRESS)
> > >  		data->os_address = data->os->header.ih_load;
> > >  
> > > -	if (data->os_address != UIMAGE_INVALID_ADDRESS) {
> > > +	if (data->os_address == UIMAGE_INVALID_ADDRESS) {
> > > +		data->os_entry = data->os->header.ih_ep;
> > > +	} else {
> > > +		data->os_entry = data->os->header.ih_ep -
> > > +			data->os->header.ih_load;
> > >  		data->os_res = uimage_load_to_sdram(data->os, 0,
> > >  				data->os_address);
> > >  		if (!data->os_res) {
> > 
> > same as my patch do
> 
> except one point you forget to reset data->os_entry in the case of a non
> uimage

To what should data->os_entry be resetted in the case of non uImage? The
code above is never executed for non uImages, so data->os_entry is never
modified in this case.

data->os_entry is expected to be an offset to the image start. Your
patch sets this to an absolute address first and adjusts this to an
offset later which is quite hard to read and understand.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |