#28885 reports other problems which aren't being fixed by this PR.

All allocations are happening taking place in file bn_lib.c in a function called bin2bn. The current call sequence to allocate (for example s->srp_ctx.a) are as follows

1. Should we update BN_new() to BN_secure_new inside the bin2bn.
2. Or should we pass in an additional flag to the function bin2bin that determines if we should use BN_new() or BN_secure_new
3. Or should we create a new function called bin2bn_secure that duplicates everything from bin2bn but uses the BN_secure_new instead of BN_new and use this function to allocate a,b,s,v

All allocations are happening taking place in file bn_lib.c in a function called bin2bn. The current call sequence to allocate (for example s->srp_ctx.a) are as follows

File function (line#) Call

1 openssl/ssl/tls_srp.c ssl_srp_calc_a_param_intern (458) s->srp_ctx.a = BN_bin2bn(rnd, sizeof(rnd), s->srp_ctx.a);
2 openssl/crypto/bn/bn_lib.c BN_bin2bn (539) bin2bn
3 openssl/crypto/bn/bn_lib.c bin2bn (451) ret = bn = BN_new();

1. Should we update `BN_new()` to `BN_secure_new` inside the `bin2bn`.

2. Or should we pass in an additional flag to the function `bin2bin` that determines if we should use `BN_new()` or `BN_secure_new`

3. Or should we create a new function called `bin2bn_secure` that duplicates everything from `bin2bn` but uses the `BN_secure_new` instead of `BN_new` and use this function to allocate a,b,s,v

Please do not change this. These are ephemeral secrets and thus should not be stored in secure memory.

Please squash the commits, drop the merge commits and rebase.

This pull contains releasing the allocated memory on failures only. Should I create a brand new pull request?

Updated.

This PR is in a state where it requires action by @openssl/committers but the last update was 30 days ago

This pull request is ready to merge

Pushed to master. Thank you for the contribution.