[RFC][PATCH] mkimage: fit: Do not tail-pad fitImage with external data
There is no reason to tail-pad fitImage with external data to 4-bytes, while fitImage without external data does not have any such padding and is often unaligned. DT spec also does not mandate any such padding.
Moreover, the tail-pad fills the last few bytes with uninitialized data, which could lead to a potential information leak.
Test: echo -n x > /tmp/data ; ./tools/mkimage -f auto -d /tmp/data /tmp/fitImage ; hexdump -vC /tmp/fitImage echo -n x > /tmp/data ; ./tools/mkimage -E -f auto -d /tmp/data /tmp/fitImage ; hexdump -vC /tmp/fitImage
Signed-off-by: Marek Vasut marex@denx.de Cc: Heinrich Schuchardt xypron.glpk@gmx.de Cc: Simon Glass sjg@chromium.org Cc: Tom Rini trini@konsulko.com --- tools/fit_image.c | 1 - 1 file changed, 1 deletion(-)
diff --git a/tools/fit_image.c b/tools/fit_image.c index dd61a816c9..2fc4c3630d 100644 --- a/tools/fit_image.c +++ b/tools/fit_image.c @@ -488,7 +488,6 @@ static int fit_extract_data(struct image_tool_params *params, const char *fname) debug("Size reduced from %x to %x\n", fit_size, fdt_totalsize(fdt)); debug("External data size %x\n", buf_ptr); new_size = fdt_totalsize(fdt); - new_size = (new_size + 3) & ~3; munmap(fdt, sbuf.st_size);
if (ftruncate(fd, new_size)) {
On Fri, 24 Apr 2020 at 04:19, Marek Vasut marex@denx.de wrote:
There is no reason to tail-pad fitImage with external data to 4-bytes, while fitImage without external data does not have any such padding and is often unaligned. DT spec also does not mandate any such padding.
Moreover, the tail-pad fills the last few bytes with uninitialized data, which could lead to a potential information leak.
Test: echo -n x > /tmp/data ; ./tools/mkimage -f auto -d /tmp/data /tmp/fitImage ; hexdump -vC /tmp/fitImage echo -n x > /tmp/data ; ./tools/mkimage -E -f auto -d /tmp/data /tmp/fitImage ; hexdump -vC /tmp/fitImage
Signed-off-by: Marek Vasut marex@denx.de Cc: Heinrich Schuchardt xypron.glpk@gmx.de Cc: Simon Glass sjg@chromium.org Cc: Tom Rini trini@konsulko.com
tools/fit_image.c | 1 - 1 file changed, 1 deletion(-)
Reviewed-by: Simon Glass sjg@chromium.org
Would be good to get this in early for testing.
On 4/24/20 6:02 PM, Simon Glass wrote:
On Fri, 24 Apr 2020 at 04:19, Marek Vasut marex@denx.de wrote:
There is no reason to tail-pad fitImage with external data to 4-bytes, while fitImage without external data does not have any such padding and is often unaligned. DT spec also does not mandate any such padding.
Moreover, the tail-pad fills the last few bytes with uninitialized data, which could lead to a potential information leak.
Test: echo -n x > /tmp/data ; ./tools/mkimage -f auto -d /tmp/data /tmp/fitImage ; hexdump -vC /tmp/fitImage echo -n x > /tmp/data ; ./tools/mkimage -E -f auto -d /tmp/data /tmp/fitImage ; hexdump -vC /tmp/fitImage
Signed-off-by: Marek Vasut marex@denx.de Cc: Heinrich Schuchardt xypron.glpk@gmx.de Cc: Simon Glass sjg@chromium.org Cc: Tom Rini trini@konsulko.com
tools/fit_image.c | 1 - 1 file changed, 1 deletion(-)
Reviewed-by: Simon Glass sjg@chromium.org
Would be good to get this in early for testing.
Tom ? This missed rc1 ...
On Fri, Apr 24, 2020 at 12:19:00PM +0200, Marek Vasut wrote:
There is no reason to tail-pad fitImage with external data to 4-bytes, while fitImage without external data does not have any such padding and is often unaligned. DT spec also does not mandate any such padding.
Moreover, the tail-pad fills the last few bytes with uninitialized data, which could lead to a potential information leak.
Test: echo -n x > /tmp/data ; ./tools/mkimage -f auto -d /tmp/data /tmp/fitImage ; hexdump -vC /tmp/fitImage echo -n x > /tmp/data ; ./tools/mkimage -E -f auto -d /tmp/data /tmp/fitImage ; hexdump -vC /tmp/fitImage
Signed-off-by: Marek Vasut marex@denx.de Cc: Heinrich Schuchardt xypron.glpk@gmx.de Cc: Simon Glass sjg@chromium.org Cc: Tom Rini trini@konsulko.com Reviewed-by: Simon Glass sjg@chromium.org
tools/fit_image.c | 1 - 1 file changed, 1 deletion(-)
diff --git a/tools/fit_image.c b/tools/fit_image.c index dd61a816c9..2fc4c3630d 100644 --- a/tools/fit_image.c +++ b/tools/fit_image.c @@ -488,7 +488,6 @@ static int fit_extract_data(struct image_tool_params *params, const char *fname) debug("Size reduced from %x to %x\n", fit_size, fdt_totalsize(fdt)); debug("External data size %x\n", buf_ptr); new_size = fdt_totalsize(fdt);
new_size = (new_size + 3) & ~3; munmap(fdt, sbuf.st_size);
if (ftruncate(fd, new_size)) {
As of: commit ebfe611be91e0075c040588a30a9996519d30aa6 Author: Kever Yang kever.yang@rock-chips.com Date: Mon Mar 30 11:56:24 2020 +0800
mkimage: fit_image: Add option to make fit header align
this doesn't apply and the area in question has changed. Can you please see if your use-case is supported correctly now? Or v2 this, thanks!
participants (3)
-
Marek Vasut -
Simon Glass -
Tom Rini