Hi Marek, On Fri, 13 Feb 2026 at 12:27, Marek Vasut <[email protected]> wrote: > > On 2/13/26 6:53 PM, Simon Glass wrote: > > Hello Simon, > > >> @@ -161,8 +165,14 @@ class Entry_nxp_imx8mcst(Entry_mkimage): > >> with open(cfg_fname, 'w') as cfgf: > >> config.write(cfgf) > >> > >> + # SSL is the default backend, PKCS11 backend is optional > >> + if self.backend == "pkcs11": > >> + cst_backend = "pkcs11" > >> + else: > >> + cst_backend = "ssl" > >> + > >> output_fname = > >> tools.get_output_filename(f'nxp.csf-output-blob.{uniq}') > >> - args = ['-i', cfg_fname, '-o', output_fname] > >> + args = ['-i', cfg_fname, '-o', output_fname, '-b', cst_backend] > >> if self.cst.run_cmd(*args) is not None: > >> outdata = tools.read_file(output_fname) > >> # fixme: 0x2000 should be CONFIG_CSF_SIZE > >> -- > >> 2.51.0 > >> > > > > This should be documented somewhere, likely in the comment for this > > entry type, so it ends up visible in the docs. > > I will send a V2 with documentation. > > > Also, this code path will need a test, please. > How do you propose to test this ?
Extend or copy one of the existing imx8 tests, perhaps with a new .dts - and I suppose set the env var and see that it works too? Note that imx8 has basic test coverage but there is quite a bit missing: tools/binman/btool/cst.py 18 4 78% tools/binman/etype/nxp_imx8mcst.py 89 69 22% tools/binman/etype/nxp_imx8mimage.py 38 6 84% I shed a small tear whenever I see that :-) Regards, Simon
