Re: [PATCH] perf/genelf: Switch deprecated openssl MD5_* functions to new EVP API

[Arnaldo Carvalho de Melo <acme@kernel.org>]

Em Tue, Sep 06, 2022 at 09:33:35AM -0300, Arnaldo Carvalho de Melo escreveu:
> Em Tue, Sep 06, 2022 at 05:40:24PM +0800, Zixuan Tan escreveu:
> > Hi Arnaldo, are there any updates?
> 
> Trying to apply it manually:

Done. Thanks.

- Arnaldo
 
> ⬢[acme@toolbox perf-urgent]$ am
> Applying: perf/genelf: Switch deprecated openssl MD5_* functions to new EVP API
> error: corrupt patch at line 31
> Patch failed at 0001 perf/genelf: Switch deprecated openssl MD5_* functions to new EVP API
> hint: Use 'git am --show-current-patch=diff' to see the failed patch
> When you have resolved this problem, run "git am --continue".
> If you prefer to skip this patch, run "git am --skip" instead.
> To restore the original branch and stop patching, run "git am --abort".
> ⬢[acme@toolbox perf-urgent]$ git am --abort
> ⬢[acme@toolbox perf-urgent]$
> ⬢[acme@toolbox perf-urgent]$ patch -p1 < ~/wb/1.patch
> patching file tools/perf/util/genelf.c
> patch: **** malformed patch at line 141: const void *code, size_t csize)
> 
> ⬢[acme@toolbox perf-urgent]$
> 
> ⬢[acme@toolbox perf-urgent]$ tail -30  ~/wb/1.patch
>  #endif
> @@ -142,15 +139,20 @@ gen_build_id(struct buildid_note *note,
>  static void
>  gen_build_id(struct buildid_note *note, unsigned long load_addr,
> const void *code, size_t csize)
>  {
> -       MD5_CTX context;
> +       EVP_MD_CTX *mdctx;
> 
>         if (sizeof(note->build_id) < 16)
>                 errx(1, "build_id too small for MD5");
> 
> -       MD5_Init(&context);
> -       MD5_Update(&context, &load_addr, sizeof(load_addr));
> -       MD5_Update(&context, code, csize);
> -       MD5_Final((unsigned char *)note->build_id, &context);
> +       mdctx = EVP_MD_CTX_new();
> +       if (!mdctx)
> +               errx(2, "failed to create EVP_MD_CTX");
> +
> +       EVP_DigestInit_ex(mdctx, EVP_md5(), NULL);
> +       EVP_DigestUpdate(mdctx, &load_addr, sizeof(load_addr));
> +       EVP_DigestUpdate(mdctx, code, csize);
> +       EVP_DigestFinal_ex(mdctx, (unsigned char *)note->build_id, NULL);
> +       EVP_MD_CTX_free(mdctx);
>  }
>  #endif
> 
> --
> 2.34.1
> ⬢[acme@toolbox perf-urgent]$
>  
> > Thanks,
> > - Zixuan
> > 
> > On Sat, Aug 27, 2022 at 2:32 AM Namhyung Kim <namhyung@kernel.org> wrote:
> > >
> > > Hello,
> > >
> > > On Fri, Aug 26, 2022 at 10:22 AM Zixuan Tan <tanzixuan.me@gmail.com> wrote:
> > > >
> > > > On Fri, Aug 26, 2022 at 4:17 AM Arnaldo Carvalho de Melo
> > > > <arnaldo.melo@gmail.com> wrote:
> > > > >
> > > > > Em Fri, Aug 26, 2022 at 01:00:58AM +0800, Zixuan Tan escreveu:
> > > > > > Switch to the flavored EVP API like in test-libcrypto.c, and remove the
> > > > > > bad gcc #pragma.
> > > > > >
> > > > > > Inspired-By: 5b245985a6de ("tools build: Switch to new openssl API for
> > > > > > test-libcrypto")
> > > > >
> > > > > How did you test the end result? Can you please describe step by step?
> > > > >
> > > > > Also please consider adding a 'perf test' entry to make sure this
> > > > > doesn't regress.
> > > >
> > > > Sorry but I don't get what you mean, what results do I need to test?
> > > >
> > > > These EVP_* APIs are just replacements for the deprecated MD5_* APIs in
> > > > openssl v3 [1][2]. With the same input, they produce the same MD5 digest.
> > > >
> > > > And this patch just does the migration work for the upgrade and does not
> > > > change the logic of the code. so...what should I test?
> > >
> > > Yeah, I understand that this merely changes the MD5 APIs.
> > > While it's good to have a test case for the genelf code, I don't think
> > > it belongs to this patch.  So,
> > >
> > > Acked-by: Namhyung Kim <namhyung@kernel.org>
> > >
> > >
> > > >
> > > > Links:
> > > > [1] https://www.openssl.org/docs/man3.0/man3/MD5.html
> > > > [2] https://stackoverflow.com/questions/69806220/advice-needed-for-migration-of-low-level-openssl-api-to-high-level-openssl-apis
> 
> -- 
> 
> - Arnaldo

-- 

- Arnaldo