Fix incorrect error return in ppc_aes_gcm_cipher_update decrypt path by Scottcjn · Pull Request #30452 · openssl/openssl

Scottcjn · 2026-03-16T17:31:10Z

Summary

Fix incorrect -1 error return in the PPC AES-GCM decrypt pre-alignment path. The function contract is to return 1 on success and 0 on failure, but the decrypt path returned -1.

Problem

In ppc_aes_gcm_cipher_update(), the decrypt pre-alignment check at line 122 returns -1 when CRYPTO_gcm128_decrypt() fails:

// Decrypt path (WRONG — was returning -1):
if (CRYPTO_gcm128_decrypt(&ctx->gcm, in, out, res))
    return -1;

// Encrypt path (CORRECT — returns 0):
if (CRYPTO_gcm128_encrypt(&ctx->gcm, in, out, res))
    return 0;

The caller at ciphercommon_gcm.c:460 checks if (!hw->cipherupdate(...)). Since !(-1) evaluates to 0 (false in C), the goto err branch is not taken and GCM processing continues with potentially corrupt state.

This is the only return -1 in all GCM cipher implementation files. All five other error returns in the same function correctly return 0.

Fix

Change return -1 to return 0 to match the encrypt path and the function's documented contract.

Fixes #30380

CLA: Signed (on file from PR #9932 on wolfSSL, also applicable here — OpenSSL CLA was completed for #30437).

nhorman · 2026-03-16T18:46:03Z

You don't appear to have a CLA on file, #30437 was marked as trivial (though it perhaps shouldn't have been), while #9932 was not authored by you.

t8m · 2026-03-16T19:02:30Z

Please amend the commit message so it contains CLA: trivial. As this is AI generated you're lucky that this is within the CLA: trivial limit. Otherwise it would not be an acceptable contribution.

ppc_aes_gcm_cipher_update() returns 1 on success and 0 on failure. The decrypt pre-alignment path (line 122) incorrectly returned -1 instead of 0 when CRYPTO_gcm128_decrypt() failed. Since the caller checks `if (!hw->cipherupdate(...))`, and !(-1) evaluates to 0 (false) in C, the error was silently swallowed and GCM processing continued with potentially corrupt state. The encrypt path at line 98 correctly returns 0. This was likely a copy-paste error when the decrypt path was added. Fixes openssl#30380 CLA: trivial

Scottcjn · 2026-03-16T19:15:16Z

Done — amended commit message to include CLA: trivial. Thank you @t8m and @nhorman for the guidance.

paulidale

Okay with trivial

esyr · 2026-03-17T00:12:07Z

Fixes: 9ab6b64ac856 "Fix AES-GCM on Power 8 CPUs"

esyr

OK with CLA: trivial.

openssl-machine · 2026-03-18T01:00:48Z

This pull request is ready to merge

ppc_aes_gcm_cipher_update() returns 1 on success and 0 on failure. The decrypt pre-alignment path (line 122) incorrectly returned -1 instead of 0 when CRYPTO_gcm128_decrypt() failed. Since the caller checks `if (!hw->cipherupdate(...))`, and !(-1) evaluates to 0 (false) in C, the error was silently swallowed and GCM processing continued with potentially corrupt state. The encrypt path at line 98 correctly returns 0. This was likely a copy-paste error when the decrypt path was added. Fixes #30380 CLA: trivial Reviewed-by: Paul Dale <[email protected]> Reviewed-by: Eugene Syromiatnikov <[email protected]> MergeDate: Wed Mar 18 07:03:41 2026 (Merged from #30452)

ppc_aes_gcm_cipher_update() returns 1 on success and 0 on failure. The decrypt pre-alignment path (line 122) incorrectly returned -1 instead of 0 when CRYPTO_gcm128_decrypt() failed. Since the caller checks `if (!hw->cipherupdate(...))`, and !(-1) evaluates to 0 (false) in C, the error was silently swallowed and GCM processing continued with potentially corrupt state. The encrypt path at line 98 correctly returns 0. This was likely a copy-paste error when the decrypt path was added. Fixes #30380 CLA: trivial Reviewed-by: Paul Dale <[email protected]> Reviewed-by: Eugene Syromiatnikov <[email protected]> MergeDate: Wed Mar 18 07:04:47 2026 (Merged from #30452)

ppc_aes_gcm_cipher_update() returns 1 on success and 0 on failure. The decrypt pre-alignment path (line 122) incorrectly returned -1 instead of 0 when CRYPTO_gcm128_decrypt() failed. Since the caller checks `if (!hw->cipherupdate(...))`, and !(-1) evaluates to 0 (false) in C, the error was silently swallowed and GCM processing continued with potentially corrupt state. The encrypt path at line 98 correctly returns 0. This was likely a copy-paste error when the decrypt path was added. Fixes #30380 CLA: trivial Reviewed-by: Paul Dale <[email protected]> Reviewed-by: Eugene Syromiatnikov <[email protected]> MergeDate: Wed Mar 18 07:05:52 2026 (Merged from #30452)

ppc_aes_gcm_cipher_update() returns 1 on success and 0 on failure. The decrypt pre-alignment path (line 122) incorrectly returned -1 instead of 0 when CRYPTO_gcm128_decrypt() failed. Since the caller checks `if (!hw->cipherupdate(...))`, and !(-1) evaluates to 0 (false) in C, the error was silently swallowed and GCM processing continued with potentially corrupt state. The encrypt path at line 98 correctly returns 0. This was likely a copy-paste error when the decrypt path was added. Fixes #30380 CLA: trivial Reviewed-by: Paul Dale <[email protected]> Reviewed-by: Eugene Syromiatnikov <[email protected]> MergeDate: Wed Mar 18 07:05:12 2026 (Merged from #30452)

ppc_aes_gcm_cipher_update() returns 1 on success and 0 on failure. The decrypt pre-alignment path (line 122) incorrectly returned -1 instead of 0 when CRYPTO_gcm128_decrypt() failed. Since the caller checks `if (!hw->cipherupdate(...))`, and !(-1) evaluates to 0 (false) in C, the error was silently swallowed and GCM processing continued with potentially corrupt state. The encrypt path at line 98 correctly returns 0. This was likely a copy-paste error when the decrypt path was added. Fixes #30380 CLA: trivial Reviewed-by: Paul Dale <[email protected]> Reviewed-by: Eugene Syromiatnikov <[email protected]> MergeDate: Wed Mar 18 07:05:30 2026 (Merged from #30452)

jogme · 2026-03-18T07:06:46Z

Merged to branches 3.3<=. Thank you!

ppc_aes_gcm_cipher_update() returns 1 on success and 0 on failure. The decrypt pre-alignment path (line 122) incorrectly returned -1 instead of 0 when CRYPTO_gcm128_decrypt() failed. Since the caller checks `if (!hw->cipherupdate(...))`, and !(-1) evaluates to 0 (false) in C, the error was silently swallowed and GCM processing continued with potentially corrupt state. The encrypt path at line 98 correctly returns 0. This was likely a copy-paste error when the decrypt path was added. Fixes #30380 CLA: trivial Reviewed-by: Paul Dale <[email protected]> Reviewed-by: Eugene Syromiatnikov <[email protected]> MergeDate: Wed Mar 18 07:06:12 2026 (Merged from #30452)

Scottcjn · 2026-03-18T15:23:55Z

Thank you to @paulidale, @esyr, @jogme, and @t8m for the reviews and merge. This complements #30437 by fixing the decrypt return path in the same AES-GCM code. Appreciate the guidance on the CLA process as well — lesson learned for next time.

— Scott Boudreaux, Elyan Labs

openssl-machine added the hold: cla required The contributor needs to submit a license agreement label Mar 16, 2026

Scottcjn force-pushed the fix/ppc-gcm-decrypt-error-return branch from 12be1bc to dae1e23 Compare March 16, 2026 19:15

openssl-machine removed the hold: cla required The contributor needs to submit a license agreement label Mar 16, 2026

paulidale approved these changes Mar 16, 2026

View reviewed changes

paulidale added the cla: trivial One of the commits is marked as 'CLA: trivial' label Mar 16, 2026

esyr approved these changes Mar 17, 2026

View reviewed changes

jogme closed this Mar 18, 2026

hopkdj mentioned this pull request Mar 31, 2026

[BOUNTY: 3 RTC] Translate RustChain README Into Your Language Scottcjn/rustchain-bounties#725

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Fix incorrect error return in ppc_aes_gcm_cipher_update decrypt path#30452

Fix incorrect error return in ppc_aes_gcm_cipher_update decrypt path#30452
Scottcjn wants to merge 1 commit intoopenssl:masterfrom
Scottcjn:fix/ppc-gcm-decrypt-error-return

Scottcjn commented Mar 16, 2026

Uh oh!

nhorman commented Mar 16, 2026

Uh oh!

t8m commented Mar 16, 2026

Uh oh!

Scottcjn commented Mar 16, 2026

Uh oh!

paulidale left a comment

Uh oh!

esyr commented Mar 17, 2026

Uh oh!

esyr left a comment

Uh oh!

openssl-machine commented Mar 18, 2026

Uh oh!

jogme commented Mar 18, 2026

Uh oh!

Scottcjn commented Mar 18, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

Uh oh!

Conversation

Scottcjn commented Mar 16, 2026

Summary

Problem

Fix

Uh oh!

nhorman commented Mar 16, 2026

Uh oh!

t8m commented Mar 16, 2026

Uh oh!

Scottcjn commented Mar 16, 2026

Uh oh!

paulidale left a comment

Choose a reason for hiding this comment

Uh oh!

esyr commented Mar 17, 2026

Uh oh!

esyr left a comment

Choose a reason for hiding this comment

Uh oh!

openssl-machine commented Mar 18, 2026

Uh oh!

jogme commented Mar 18, 2026

Uh oh!

Scottcjn commented Mar 18, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants