| Message ID | 24b427fa8ee7500d08b6ed412da70e597bf5061a.1436454776.git.baruch@tkos.co.il |
|---|---|
| State | Awaiting Upstream |
| Headers | show |
Hi Baruch, On 09/07/2015 17:19, Baruch Siach wrote: > When dcd_len is 0 the Write Data command that the set_dcd_rst_v2() routine > generates is empty. This causes HAB to complain that the command is invalid. > > --------- HAB Event 1 ----------------- > event data: > 0xdb 0x00 0x0c 0x41 0x33 0x06 0xc0 0x00 > 0xcc 0x00 0x04 0x04 > > To fix this set the DCD pointer in the IVT to NULL in this case. The DCD header > itself is still needed for detect_imximage_version() to determine the image > version. > > Signed-off-by: Baruch Siach <baruch@tkos.co.il> > --- > Resending with a proper To: header. > > tools/imximage.c | 6 +++++- > 1 file changed, 5 insertions(+), 1 deletion(-) > > diff --git a/tools/imximage.c b/tools/imximage.c > index 6f469ae6336c..7b91d09699ad 100644 > --- a/tools/imximage.c > +++ b/tools/imximage.c > @@ -250,7 +250,11 @@ static void set_imx_hdr_v2(struct imx_header *imxhdr, uint32_t dcd_len, > hdr_base = entry_point - imximage_init_loadsize + > flash_offset; > fhdr_v2->self = hdr_base; > - fhdr_v2->dcd_ptr = hdr_base + offsetof(imx_header_v2_t, dcd_table); > + if (dcd_len > 0) > + fhdr_v2->dcd_ptr = hdr_base > + + offsetof(imx_header_v2_t, dcd_table); > + else > + fhdr_v2->dcd_ptr = 0; > fhdr_v2->boot_data_ptr = hdr_base > + offsetof(imx_header_v2_t, boot_data); > hdr_v2->boot_data.start = entry_point - imximage_init_loadsize; > It looks ok for me. Acked-by: Stefano Babic <sbabic@denx.de> Best regards, Stefano Babic
Hi Stefano, On Thu, Jul 09, 2015 at 06:04:17PM +0200, Stefano Babic wrote: > On 09/07/2015 17:19, Baruch Siach wrote: > > When dcd_len is 0 the Write Data command that the set_dcd_rst_v2() routine > > generates is empty. This causes HAB to complain that the command is invalid. > > > > --------- HAB Event 1 ----------------- > > event data: > > 0xdb 0x00 0x0c 0x41 0x33 0x06 0xc0 0x00 > > 0xcc 0x00 0x04 0x04 > > > > To fix this set the DCD pointer in the IVT to NULL in this case. The DCD header > > itself is still needed for detect_imximage_version() to determine the image > > version. > > > > Signed-off-by: Baruch Siach <baruch@tkos.co.il> > > --- > > Resending with a proper To: header. > > > > tools/imximage.c | 6 +++++- > > 1 file changed, 5 insertions(+), 1 deletion(-) > > > > diff --git a/tools/imximage.c b/tools/imximage.c > > index 6f469ae6336c..7b91d09699ad 100644 > > --- a/tools/imximage.c > > +++ b/tools/imximage.c > > @@ -250,7 +250,11 @@ static void set_imx_hdr_v2(struct imx_header *imxhdr, uint32_t dcd_len, > > hdr_base = entry_point - imximage_init_loadsize + > > flash_offset; > > fhdr_v2->self = hdr_base; > > - fhdr_v2->dcd_ptr = hdr_base + offsetof(imx_header_v2_t, dcd_table); > > + if (dcd_len > 0) > > + fhdr_v2->dcd_ptr = hdr_base > > + + offsetof(imx_header_v2_t, dcd_table); > > + else > > + fhdr_v2->dcd_ptr = 0; > > fhdr_v2->boot_data_ptr = hdr_base > > + offsetof(imx_header_v2_t, boot_data); > > hdr_v2->boot_data.start = entry_point - imximage_init_loadsize; > > > > It looks ok for me. > > Acked-by: Stefano Babic <sbabic@denx.de> Thanks. Which tree is this patch going through? baruch
On 14/07/2015 21:05, Baruch Siach wrote: > Hi Stefano, > > On Thu, Jul 09, 2015 at 06:04:17PM +0200, Stefano Babic wrote: >> On 09/07/2015 17:19, Baruch Siach wrote: >>> When dcd_len is 0 the Write Data command that the set_dcd_rst_v2() routine >>> generates is empty. This causes HAB to complain that the command is invalid. >>> >>> --------- HAB Event 1 ----------------- >>> event data: >>> 0xdb 0x00 0x0c 0x41 0x33 0x06 0xc0 0x00 >>> 0xcc 0x00 0x04 0x04 >>> >>> To fix this set the DCD pointer in the IVT to NULL in this case. The DCD header >>> itself is still needed for detect_imximage_version() to determine the image >>> version. >>> >>> Signed-off-by: Baruch Siach <baruch@tkos.co.il> >>> --- >>> Resending with a proper To: header. >>> >>> tools/imximage.c | 6 +++++- >>> 1 file changed, 5 insertions(+), 1 deletion(-) >>> >>> diff --git a/tools/imximage.c b/tools/imximage.c >>> index 6f469ae6336c..7b91d09699ad 100644 >>> --- a/tools/imximage.c >>> +++ b/tools/imximage.c >>> @@ -250,7 +250,11 @@ static void set_imx_hdr_v2(struct imx_header *imxhdr, uint32_t dcd_len, >>> hdr_base = entry_point - imximage_init_loadsize + >>> flash_offset; >>> fhdr_v2->self = hdr_base; >>> - fhdr_v2->dcd_ptr = hdr_base + offsetof(imx_header_v2_t, dcd_table); >>> + if (dcd_len > 0) >>> + fhdr_v2->dcd_ptr = hdr_base >>> + + offsetof(imx_header_v2_t, dcd_table); >>> + else >>> + fhdr_v2->dcd_ptr = 0; >>> fhdr_v2->boot_data_ptr = hdr_base >>> + offsetof(imx_header_v2_t, boot_data); >>> hdr_v2->boot_data.start = entry_point - imximage_init_loadsize; >>> >> >> It looks ok for me. >> >> Acked-by: Stefano Babic <sbabic@denx.de> > > Thanks. > > Which tree is this patch going through? This goes first into u-boot-imx, until I'll send a PR to Tom for inclusion in his tree, Best regards, Stefano Babic
Hi Stefano, On Wed, Jul 15, 2015 at 09:13:33AM +0200, Stefano Babic wrote: > On 14/07/2015 21:05, Baruch Siach wrote: > > On Thu, Jul 09, 2015 at 06:04:17PM +0200, Stefano Babic wrote: > >> On 09/07/2015 17:19, Baruch Siach wrote: > >>> When dcd_len is 0 the Write Data command that the set_dcd_rst_v2() routine > >>> generates is empty. This causes HAB to complain that the command is invalid. > >>> > >>> --------- HAB Event 1 ----------------- > >>> event data: > >>> 0xdb 0x00 0x0c 0x41 0x33 0x06 0xc0 0x00 > >>> 0xcc 0x00 0x04 0x04 > >>> > >>> To fix this set the DCD pointer in the IVT to NULL in this case. The DCD header > >>> itself is still needed for detect_imximage_version() to determine the image > >>> version. > >>> > >>> Signed-off-by: Baruch Siach <baruch@tkos.co.il> > >>> --- > >>> Resending with a proper To: header. > >>> > >>> tools/imximage.c | 6 +++++- > >>> 1 file changed, 5 insertions(+), 1 deletion(-) > >>> > >>> diff --git a/tools/imximage.c b/tools/imximage.c > >>> index 6f469ae6336c..7b91d09699ad 100644 > >>> --- a/tools/imximage.c > >>> +++ b/tools/imximage.c > >>> @@ -250,7 +250,11 @@ static void set_imx_hdr_v2(struct imx_header *imxhdr, uint32_t dcd_len, > >>> hdr_base = entry_point - imximage_init_loadsize + > >>> flash_offset; > >>> fhdr_v2->self = hdr_base; > >>> - fhdr_v2->dcd_ptr = hdr_base + offsetof(imx_header_v2_t, dcd_table); > >>> + if (dcd_len > 0) > >>> + fhdr_v2->dcd_ptr = hdr_base > >>> + + offsetof(imx_header_v2_t, dcd_table); > >>> + else > >>> + fhdr_v2->dcd_ptr = 0; > >>> fhdr_v2->boot_data_ptr = hdr_base > >>> + offsetof(imx_header_v2_t, boot_data); > >>> hdr_v2->boot_data.start = entry_point - imximage_init_loadsize; > >> > >> It looks ok for me. > >> > >> Acked-by: Stefano Babic <sbabic@denx.de> > > > > Which tree is this patch going through? > > This goes first into u-boot-imx, until I'll send a PR to Tom for > inclusion in his tree, Thanks again. Is there a reason this patch is not in the u-boot-imx git tree? baruch
Hi Baruch, On 02/08/2015 20:45, Baruch Siach wrote: > Hi Stefano, > > On Wed, Jul 15, 2015 at 09:13:33AM +0200, Stefano Babic wrote: >> On 14/07/2015 21:05, Baruch Siach wrote: >>> On Thu, Jul 09, 2015 at 06:04:17PM +0200, Stefano Babic wrote: >>>> On 09/07/2015 17:19, Baruch Siach wrote: >>>>> When dcd_len is 0 the Write Data command that the set_dcd_rst_v2() routine >>>>> generates is empty. This causes HAB to complain that the command is invalid. >>>>> >>>>> --------- HAB Event 1 ----------------- >>>>> event data: >>>>> 0xdb 0x00 0x0c 0x41 0x33 0x06 0xc0 0x00 >>>>> 0xcc 0x00 0x04 0x04 >>>>> >>>>> To fix this set the DCD pointer in the IVT to NULL in this case. The DCD header >>>>> itself is still needed for detect_imximage_version() to determine the image >>>>> version. >>>>> >>>>> Signed-off-by: Baruch Siach <baruch@tkos.co.il> >>>>> --- >>>>> Resending with a proper To: header. >>>>> >>>>> tools/imximage.c | 6 +++++- >>>>> 1 file changed, 5 insertions(+), 1 deletion(-) >>>>> >>>>> diff --git a/tools/imximage.c b/tools/imximage.c >>>>> index 6f469ae6336c..7b91d09699ad 100644 >>>>> --- a/tools/imximage.c >>>>> +++ b/tools/imximage.c >>>>> @@ -250,7 +250,11 @@ static void set_imx_hdr_v2(struct imx_header *imxhdr, uint32_t dcd_len, >>>>> hdr_base = entry_point - imximage_init_loadsize + >>>>> flash_offset; >>>>> fhdr_v2->self = hdr_base; >>>>> - fhdr_v2->dcd_ptr = hdr_base + offsetof(imx_header_v2_t, dcd_table); >>>>> + if (dcd_len > 0) >>>>> + fhdr_v2->dcd_ptr = hdr_base >>>>> + + offsetof(imx_header_v2_t, dcd_table); >>>>> + else >>>>> + fhdr_v2->dcd_ptr = 0; >>>>> fhdr_v2->boot_data_ptr = hdr_base >>>>> + offsetof(imx_header_v2_t, boot_data); >>>>> hdr_v2->boot_data.start = entry_point - imximage_init_loadsize; >>>> >>>> It looks ok for me. >>>> >>>> Acked-by: Stefano Babic <sbabic@denx.de> >>> >>> Which tree is this patch going through? >> >> This goes first into u-boot-imx, until I'll send a PR to Tom for >> inclusion in his tree, > > Thanks again. Is there a reason this patch is not in the u-boot-imx git tree? The reason is that it was not assigned in patchwork as I supposed and this leads to forget it when I rework u-boot-imx - thanks for pointing out ! Applied to u-boot-imx, thanks ! Best regards, Stefano Babic
diff --git a/tools/imximage.c b/tools/imximage.c index 6f469ae6336c..7b91d09699ad 100644 --- a/tools/imximage.c +++ b/tools/imximage.c @@ -250,7 +250,11 @@ static void set_imx_hdr_v2(struct imx_header *imxhdr, uint32_t dcd_len, hdr_base = entry_point - imximage_init_loadsize + flash_offset; fhdr_v2->self = hdr_base; - fhdr_v2->dcd_ptr = hdr_base + offsetof(imx_header_v2_t, dcd_table); + if (dcd_len > 0) + fhdr_v2->dcd_ptr = hdr_base + + offsetof(imx_header_v2_t, dcd_table); + else + fhdr_v2->dcd_ptr = 0; fhdr_v2->boot_data_ptr = hdr_base + offsetof(imx_header_v2_t, boot_data); hdr_v2->boot_data.start = entry_point - imximage_init_loadsize;
When dcd_len is 0 the Write Data command that the set_dcd_rst_v2() routine generates is empty. This causes HAB to complain that the command is invalid. --------- HAB Event 1 ----------------- event data: 0xdb 0x00 0x0c 0x41 0x33 0x06 0xc0 0x00 0xcc 0x00 0x04 0x04 To fix this set the DCD pointer in the IVT to NULL in this case. The DCD header itself is still needed for detect_imximage_version() to determine the image version. Signed-off-by: Baruch Siach <baruch@tkos.co.il> --- Resending with a proper To: header. tools/imximage.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)