| Message ID | 20200728000313.31130-1-festevam@gmail.com |
|---|---|
| State | Accepted |
| Commit | aaa91a4e4b8a5d74f1317e18aa47d2a7a72e0c43 |
| Delegated to: | Tom Rini |
| Headers | show |
| Series | [v2] fit_image: Use calloc() to fix reproducibility issue | expand |
On 2020-07-27, Fabio Estevam wrote: > Vagrant Cascadian reported that mx6cuboxi target no longer builds > reproducibility on Debian. > > One example of builds mismatches: > > 00096680: 696e 6700 736f 756e 642d 6461 6900 6465 ing.sound-dai.de > -00096690: 7465 6374 2d67 7069 6f73 0000 tect-gpios.. > +00096690: 7465 6374 2d67 7069 6f73 0061 tect-gpios.a > > This problem happens because all the buffers in fit_image.c are > allocated via malloc(), which does not zero out the allocated buffer. > > Using calloc() fixes this unpredictable behaviour as it guarantees > that the allocated buffer are zero initialized. > > Reported-by: Vagrant Cascadian <vagrant@reproducible-builds.org> > Suggested-by: Tom Rini <trini@konsulko.com> > Signed-off-by: Fabio Estevam <festevam@gmail.com> Tested that it boots on the mx6cuboxi supported board hummingboard i2ex, and that it fixes the reproducibility issues. Thanks! Tested-by: Vagrant Cascadian <vagrant@reproducible-builds.org> live well, vagrant > --- > Changes since v1: > - Improve the commit log description by stating why calloc() helps. > > tools/fit_image.c | 8 ++++---- > 1 file changed, 4 insertions(+), 4 deletions(-) > > diff --git a/tools/fit_image.c b/tools/fit_image.c > index a082d9386d..0c6185d892 100644 > --- a/tools/fit_image.c > +++ b/tools/fit_image.c > @@ -388,7 +388,7 @@ static int fit_build(struct image_tool_params *params, const char *fname) > size = fit_calc_size(params); > if (size < 0) > return -1; > - buf = malloc(size); > + buf = calloc(1, size); > if (!buf) { > fprintf(stderr, "%s: Out of memory (%d bytes)\n", > params->cmdname, size); > @@ -467,7 +467,7 @@ static int fit_extract_data(struct image_tool_params *params, const char *fname) > * Allocate space to hold the image data we will extract, > * extral space allocate for image alignment to prevent overflow. > */ > - buf = malloc(fit_size + (align_size * image_number)); > + buf = calloc(1, fit_size + (align_size * image_number)); > if (!buf) { > ret = -ENOMEM; > goto err_munmap; > @@ -572,7 +572,7 @@ static int fit_import_data(struct image_tool_params *params, const char *fname) > > /* Allocate space to hold the new FIT */ > size = sbuf.st_size + 16384; > - fdt = malloc(size); > + fdt = calloc(1, size); > if (!fdt) { > fprintf(stderr, "%s: Failed to allocate memory (%d bytes)\n", > __func__, size); > @@ -673,7 +673,7 @@ static int copyfile(const char *src, const char *dst) > goto out; > } > > - buf = malloc(512); > + buf = calloc(1, 512); > if (!buf) { > printf("Can't allocate buffer to copy file\n"); > goto out; > -- > 2.17.1
On Mon, Jul 27, 2020 at 09:03:13PM -0300, Fabio Estevam wrote: > Vagrant Cascadian reported that mx6cuboxi target no longer builds > reproducibility on Debian. > > One example of builds mismatches: > > 00096680: 696e 6700 736f 756e 642d 6461 6900 6465 ing.sound-dai.de > -00096690: 7465 6374 2d67 7069 6f73 0000 tect-gpios.. > +00096690: 7465 6374 2d67 7069 6f73 0061 tect-gpios.a > > This problem happens because all the buffers in fit_image.c are > allocated via malloc(), which does not zero out the allocated buffer. > > Using calloc() fixes this unpredictable behaviour as it guarantees > that the allocated buffer are zero initialized. > > Reported-by: Vagrant Cascadian <vagrant@reproducible-builds.org> > Suggested-by: Tom Rini <trini@konsulko.com> > Signed-off-by: Fabio Estevam <festevam@gmail.com> > Tested-by: Vagrant Cascadian <vagrant@reproducible-builds.org> Applied to u-boot/master, thanks!
diff --git a/tools/fit_image.c b/tools/fit_image.c index a082d9386d..0c6185d892 100644 --- a/tools/fit_image.c +++ b/tools/fit_image.c @@ -388,7 +388,7 @@ static int fit_build(struct image_tool_params *params, const char *fname) size = fit_calc_size(params); if (size < 0) return -1; - buf = malloc(size); + buf = calloc(1, size); if (!buf) { fprintf(stderr, "%s: Out of memory (%d bytes)\n", params->cmdname, size); @@ -467,7 +467,7 @@ static int fit_extract_data(struct image_tool_params *params, const char *fname) * Allocate space to hold the image data we will extract, * extral space allocate for image alignment to prevent overflow. */ - buf = malloc(fit_size + (align_size * image_number)); + buf = calloc(1, fit_size + (align_size * image_number)); if (!buf) { ret = -ENOMEM; goto err_munmap; @@ -572,7 +572,7 @@ static int fit_import_data(struct image_tool_params *params, const char *fname) /* Allocate space to hold the new FIT */ size = sbuf.st_size + 16384; - fdt = malloc(size); + fdt = calloc(1, size); if (!fdt) { fprintf(stderr, "%s: Failed to allocate memory (%d bytes)\n", __func__, size); @@ -673,7 +673,7 @@ static int copyfile(const char *src, const char *dst) goto out; } - buf = malloc(512); + buf = calloc(1, 512); if (!buf) { printf("Can't allocate buffer to copy file\n"); goto out;
Vagrant Cascadian reported that mx6cuboxi target no longer builds reproducibility on Debian. One example of builds mismatches: 00096680: 696e 6700 736f 756e 642d 6461 6900 6465 ing.sound-dai.de -00096690: 7465 6374 2d67 7069 6f73 0000 tect-gpios.. +00096690: 7465 6374 2d67 7069 6f73 0061 tect-gpios.a This problem happens because all the buffers in fit_image.c are allocated via malloc(), which does not zero out the allocated buffer. Using calloc() fixes this unpredictable behaviour as it guarantees that the allocated buffer are zero initialized. Reported-by: Vagrant Cascadian <vagrant@reproducible-builds.org> Suggested-by: Tom Rini <trini@konsulko.com> Signed-off-by: Fabio Estevam <festevam@gmail.com> --- Changes since v1: - Improve the commit log description by stating why calloc() helps. tools/fit_image.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)