Skip to content

crypto: handle exceptions in hmac/hash.digest #12164

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
tniessen wants to merge 4 commits into from
Closed

crypto: handle exceptions in hmac/hash.digest #12164

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
768c986
crypto: handle exceptions in hmac/hash.digest
tniessen Apr 2, 2017
23b71f5
Disallow symbols as encoding
tniessen Apr 3, 2017
23f95ab
Add checks and regression test
tniessen Apr 3, 2017
27d359c
Punctuate commit in crypto.js
tniessen Apr 5, 2017
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
Add checks and regression test
  • Loading branch information
@tniessen
tniessen committed Apr 3, 2017
commit 23f95ab393e519336cea4d514ee0bb49016fc7e1
2 changes: 2 additions & 0 deletions src/node_crypto.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3797,6 +3797,7 @@ void Hmac::HmacDigest(const FunctionCallbackInfo<Value>& args) {

enum encoding encoding = BUFFER;
if (args.Length() >= 1) {
CHECK(args[0]->IsString());
encoding = ParseEncoding(env->isolate(), args[0], BUFFER);
Copy link
Member

@addaleax addaleax Apr 3, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe leave a CHECK(args[i]->IsString()); line before the ParseEncoding calls?

Copy link
Member Author

@tniessen tniessen Apr 3, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If I remember correctly, most functions just ignore invalid encoding values and use the default value instead. Currently, we are forcing the conversion to a string, but just because it is a string doesn't make it a valid encoding value. Adding a CHECK shouldn't do any harm, until we decide to finally drop support for toString() for consistency with other APIs. However, if we do not add a chack and the parameter is not a string (which won't happen now, but might as soon as we decide to drop support for toString()), there won't be any problems, the value will be ignored (ParseEncoding will return instantly).

The above only applies to hmac and hash. For sign and verify, the encoding parameter seems to always be null, so the CHECK would fail, right?

Copy link
Member

@addaleax addaleax Apr 3, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@tniessen Still, if we drop toString() support it’s easy to just drop the checks again. (Likewise, it would be good if you included the test you said you had prepared in this PR – it’s easy to change things later, but for now it’s best to test the current behaviour.)

The above only applies to hmac and hash. For sign and verify, the encoding parameter seems to always be null, so the CHECK would fail, right?

It looks that way, yes… maybe you could fix that up and drop the unused code bits? That also simplifies SignFinal(), right now the StringBytes::Encode() encode call there seems like a completely unnecessary extra copy operation.

Copy link
Member Author

@tniessen tniessen Apr 3, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@addaleax You are right, I will add the CHECKs and add the regression test.

maybe you could fix that up and drop the unused code bits?

Is it okay to do that in a separate PR or will that create too much organizational overhead?

Copy link
Member

@addaleax addaleax Apr 3, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is it okay to do that in a separate PR or will that create too much organizational overhead?

That’s your call… you can also maintain multiple commits in a single PR, which may or may not be easier for you. :) I don’t think there are other open pull requests for this code that would generate merge conflicts, and everything we talked about (except maybe dropping toString() support) semver-patch.

Copy link
Member Author

@tniessen tniessen Apr 3, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I will create a separate PR after this gets merged to avoid conflicts and keep it simple, thank you for your support :)

}

Expand Down Expand Up @@ -3919,6 +3920,7 @@ void Hash::HashDigest(const FunctionCallbackInfo<Value>& args) {

enum encoding encoding = BUFFER;
if (args.Length() >= 1) {
CHECK(args[0]->IsString());
encoding = ParseEncoding(env->isolate(), args[0], BUFFER);
}

Expand Down
24 changes: 24 additions & 0 deletions test/parallel/test-regress-GH-9819.js
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,24 @@
'use strict';
const common = require('../common');
const assert = require('assert');
const execFile = require('child_process').execFile;

if (!common.hasCrypto) {
common.skip('missing crypto');
return;
}

const setup = 'const enc = { toString: () => { throw new Error("xyz"); } };';

const scripts = [
'crypto.createHash("sha256").digest(enc)',
'crypto.createHmac("sha256", "msg").digest(enc)'
];
Copy link
Member

@jasnell jasnell Apr 3, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Adding the code in a test/fixtures file would be a bit more readable.

Copy link
Member Author

@tniessen tniessen Apr 3, 2017 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you mean adding the code as a JSON file or do you mean a JS file which can then be executed within the spawned node process? We are talking about two lines of code here...

Copy link
Member

@jasnell jasnell Apr 4, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, the latter. I'm fine with this the way it is, I'd just generally prefer a fixture. Feel free to ignore tho :-)

Copy link
Member Author

@tniessen tniessen Apr 4, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Mhhh I think I will keep it this way unless someone expresses a strong opposing opinion, but thank you for your feedback :)


scripts.forEach((script) => {
const node = process.execPath;
const code = setup + ';' + script;
execFile(node, [ '-e', code ], common.mustCall((err, stdout, stderr) => {
assert(stderr.includes('Error: xyz'), 'digest crashes');
}));
});