mirror of
https://github.com/DarkFlippers/unleashed-firmware.git
synced 2026-08-07 01:59:50 +00:00
b1b1f94c42
* nfc: fix double-free of card_number_str in Plantain parser printf_plantain_data() freed purse->card_number_str, then the caller freed the same (non-NULL) pointer again at the end of the parse -> double-free / heap corruption on every Plantain card. Remove the free inside the print helper; the caller that allocated it owns the free. Part of #1029 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * nfc: fix double-free + secondary-ticket leak in SZPPK parser In the second-ticket branch the cleanup freed primary_ticket.* (already freed just above) instead of the freshly-allocated secondary_ticket.*, causing a double-free of the primary strings and a leak of all four secondary strings on two-trip tickets. Free the secondary_ticket.* members. Part of #1029 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * nfc: fix double-free + secondary-ticket leak in SEVPPK parser Same copy-paste bug as SZPPK: the second-ticket branch freed primary_ticket.* (already freed) instead of secondary_ticket.*, double-freeing the primary strings and leaking the secondary ones on two-trip tickets. Free the secondary_ticket.* members. Part of #1029 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * nfc: fix double-free + secondary-ticket leak in SK parser Same copy-paste bug as SZPPK/SEVPPK: the second-ticket branch freed primary_ticket.* (already freed) instead of secondary_ticket.*, double-freeing the primary strings and leaking the secondary ones on two-trip tickets. Free the secondary_ticket.* members. Part of #1029 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * nfc: fix restricted_weekday_string leak in Saflok parser restricted_weekday_string was allocated and only read via furi_string_get_cstr; the file had no furi_string_free at all, leaking one FuriString on every Saflok card parse (both MFC and UL-C variants). Free it before the do/while(false) exits (after #endif so it applies to both SL_PROTO builds). Part of #1029 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * nfc: free api_resolver in nfc_app_free (fix per-launch leak) instance->api_resolver (composite_api_resolver_alloc + two composite_api_resolver_add nodes) was never freed; nfc_supported_cards only borrows it. Leaked the resolver struct plus its two list nodes on every NFC-app open->close. Free it after nfc_supported_cards_free (its only user). Part of #1029 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * nfc: fix Stream + storage-record leak on CUID dict open failure When a CUID dictionary passed keys_dict_check_presence but buffered_file_stream_open then failed, cleanup did only buffered_file_stream_close + free(dict), leaking the Stream object and the RECORD_STORAGE reference opened just above. Use keys_dict_free(dict), which closes+frees the stream, frees the dict, and closes RECORD_STORAGE, matching the total_keys==0 path. Part of #1029 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * nfc: fix ~15x RAM over-allocation of DESFire file_data array mf_desfire_file_data_array_config.type_size used sizeof(MfDesfireData) (~60+ bytes) but the element type is MfDesfireFileData ({ SimpleArray* data; } = 4 bytes), so every file slot over-allocated ~15x, scaled by (files x applications) on the card. Use sizeof(MfDesfireFileData). Consistent stride, so no behavior change beyond reduced heap use. Part of #1029 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * changelog: NFC memory leaks & double-frees (PR #1030, Fixes #1029) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>