fscrypt: avoid data race on fscrypt_mode::logged_impl_name
Eric Biggers
maintainerPATCH v2 1/1 fscrypt: avoid data race on fscrypt_mode::logged_impl_name
From: Eric Biggers <[email protected]>
The access to logged_impl_name is technically a data race, which tools like KCSAN could complain about in the future. See:
https://github.com/google/ktsan/wiki/READ_ONCE-and-WRITE_ONCE
Fix by using xchg(), which also ensures that only one thread does the logging.
This also required switching from bool to int, to avoid a build error on the RISC-V architecture which doesn't implement xchg on bytes.
--- fs/crypto/fscrypt_private.h | 2 +- fs/crypto/keysetup.c | 6 ++---- 2 files changed, 3 insertions(+), 5 deletions(-)
fs/crypto/fscrypt_private.h +1 -1
@@ -435,7 +435,7 @@ struct fscrypt_mode {const char *cipher_str;int keysize;int ivsize; -bool logged_impl_name; +int logged_impl_name; }; static inline bool
fs/crypto/keysetup.c +2 -4
@@ -81,15 +81,13 @@ struct crypto_skcipher *fscrypt_allocate_skcipher(struct fscrypt_mode *mode, mode->cipher_str, PTR_ERR(tfm));return tfm; } -if (unlikely (!mode->logged_impl_name)) { +if (!xchg (&mode->logged_impl_name,1 )) {/* * fscrypt performance can vary greatly depending on which * crypto algorithm implementation is used. Help people debug * performance problems by logging the ->cra_driver_name the -* first time a mode is used. Note that multiple threads can -* race here, but it doesn't really matter. +* first time a mode is used. */ - mode->logged_impl_name = true;pr_info ("fscrypt: %s using implementation \"%s\"\n" , mode->friendly_name,crypto_skcipher_alg (tfm)->base.cra_driver_name);
Signature · 1 line
2.23.0.866.gb869b98d4c-goog
Quoted 6 lines · Eric Biggers
From: Eric Biggers <[email protected]>
The access to
logged_impl_nameis technically a data race, which tools like KCSAN could complain about in the future. See:
https://github.com/google/ktsan/wiki/READ_ONCE-and-WRITE_ONCE
Quoted 7 lines
Fix by using
xchg(), which also ensures that only one thread does the logging.This also required switching from bool to int, to avoid a build error on the RISC-V architecture which doesn't implement xchg on bytes.
Signed-off-by Eric Biggers <[email protected]>
Applied to fscrypt.git#master for 5.5.
- Eric