From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Attak, Hamza" Subject: [PATCH] msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers Date: Tue, 30 May 2017 17:50:21 +0000 Message-ID: Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="_004_DF4PR84MB00286E6F6F1E51633C8D65C4F2F00DF4PR84MB0028NAMP_" Return-path: Content-Language: en-US List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: tpmdd-devel-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org To: "tpmdd-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org" Cc: "Jacquin, Ludovic" , "Edwards, Nigel" List-Id: tpmdd-devel@lists.sourceforge.net --_004_DF4PR84MB00286E6F6F1E51633C8D65C4F2F00DF4PR84MB0028NAMP_ Content-Type: multipart/alternative; boundary="_000_DF4PR84MB00286E6F6F1E51633C8D65C4F2F00DF4PR84MB0028NAMP_" --_000_DF4PR84MB00286E6F6F1E51633C8D65C4F2F00DF4PR84MB0028NAMP_ Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Following the recent Nuvoton patch submitted by Mimi Zohar: https://sourceforge.net/p/tpmdd/mailman/message/35685260/ Based on this work, I decided to apply a similar modification to the generi= c TPM drivers in the kernel. The patch simply replaces all msleep function calls with usleep_range calls= in the generic drivers. As a former patch already exists for the Nuvoton T= PM, it is therefore not included in this new patch. Tested with an Infineon TPM 1.2, using the generic tpm-tis module, for a th= ousand PCR extends, we see results going from 1m57s unpatched to 40s with t= he new patch. We obtain similar results when using the original and patched= tpm_infineon driver, which is also part of the patch. Similarly with a STM TPM 2.0, using the CRB driver, it takes about 20ms per= extend unpatched and around 7ms with the new patch. Note that the PCR consistency is untouched with this patch, each TPM has be= en tested with 10 million extends and the aggregated PCR value is continuou= sly verified to be correct. As an extension of this work, this could potentially and easily be applied = to other vendor's drivers. Still, these changes are not included in the pro= posed patch as they are untested. Thanks, Hamza ATTAK. --_000_DF4PR84MB00286E6F6F1E51633C8D65C4F2F00DF4PR84MB0028NAMP_ Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable

Following the recent Nuvoton patch submitted by Mimi= Zohar:

https://sourceforge.net/p/tpmdd/mailman/message/3568= 5260/

Based on this work, I decided to apply a similar mod= ification to the generic TPM drivers in the kernel.

The patch simply replaces all msleep function calls = with usleep_range calls in the generic drivers. As a former patch already e= xists for the Nuvoton TPM, it is therefore not included in this new patch.<= o:p>

Tested with an Infineon TPM 1.2, using the generic t= pm-tis module, for a thousand PCR extends, we see results going from 1m57s = unpatched to 40s with the new patch. We obtain similar results when using t= he original and patched tpm_infineon driver, which is also part of the patch.

Similarly with a STM TPM 2.0, using the CRB driver, = it takes about 20ms per extend unpatched and around 7ms with the new patch.=

Note that the PCR consistency is untouched with this= patch, each TPM has been tested with 10 million extends and the aggregated= PCR value is continuously verified to be correct.

As an extension of this work, this could potentially= and easily be applied to other vendor's drivers. Still, these changes are = not included in the proposed patch as they are untested.

Thanks,

Hamza ATTAK.

--_000_DF4PR84MB00286E6F6F1E51633C8D65C4F2F00DF4PR84MB0028NAMP_-- --_004_DF4PR84MB00286E6F6F1E51633C8D65C4F2F00DF4PR84MB0028NAMP_ Content-Type: application/octet-stream; name="tpm_usleep_range_tested.patch" Content-Description: tpm_usleep_range_tested.patch Content-Disposition: attachment; filename="tpm_usleep_range_tested.patch"; size=4432; creation-date="Tue, 30 May 2017 16:53:22 GMT"; modification-date="Tue, 30 May 2017 15:32:34 GMT" Content-Transfer-Encoding: base64 ZGlmZiAtLWdpdCBhL2RyaXZlcnMvY2hhci90cG0vdHBtLWludGVyZmFjZS5jIGIvZHJpdmVycy9j aGFyL3RwbS90cG0taW50ZXJmYWNlLmMKaW5kZXggYmQyMTI4ZS4uZjAzZGFmNSAxMDA2NDQKLS0t IGEvZHJpdmVycy9jaGFyL3RwbS90cG0taW50ZXJmYWNlLmMKKysrIGIvZHJpdmVycy9jaGFyL3Rw bS90cG0taW50ZXJmYWNlLmMKQEAgLTM5NSw3ICszOTUsNyBAQCBzc2l6ZV90IHRwbV90cmFuc21p dChzdHJ1Y3QgdHBtX2NoaXAgKmNoaXAsIGNvbnN0IHU4ICpidWYsIHNpemVfdCBidWZzaXosCiAJ CQlnb3RvIG91dDsKIAkJfQogCi0JCW1zbGVlcChUUE1fVElNRU9VVCk7CS8qIENIRUNLICovCisJ CXVzbGVlcF9yYW5nZShUUE1fVElNRU9VVCAqIDEwMDAsIChUUE1fVElNRU9VVCAqIDEwMDApICsg MzAwKTsKIAkJcm1iKCk7CiAJfSB3aGlsZSAodGltZV9iZWZvcmUoamlmZmllcywgc3RvcCkpOwog CkBAIC04NjIsNyArODYyLDcgQEAgaW50IHRwbV9kb19zZWxmdGVzdChzdHJ1Y3QgdHBtX2NoaXAg KmNoaXApCiAJCQlkZXZfaW5mbygKIAkJCSAgICAmY2hpcC0+ZGV2LCBIV19FUlIKIAkJCSAgICAi VFBNIGNvbW1hbmQgdGltZWQgb3V0IGR1cmluZyBjb250aW51ZSBzZWxmIHRlc3QiKTsKLQkJCW1z bGVlcChkZWxheV9tc2VjKTsKKwkJCXVzbGVlcF9yYW5nZShkZWxheV9tc2VjICogMTAwMCwgKGRl bGF5X21zZWMgKiAxMDAwKSArIDMwMCk7CiAJCQljb250aW51ZTsKIAkJfQogCkBAIC04NzcsNyAr ODc3LDcgQEAgaW50IHRwbV9kb19zZWxmdGVzdChzdHJ1Y3QgdHBtX2NoaXAgKmNoaXApCiAJCX0K IAkJaWYgKHJjICE9IFRQTV9XQVJOX0RPSU5HX1NFTEZURVNUKQogCQkJcmV0dXJuIHJjOwotCQlt c2xlZXAoZGVsYXlfbXNlYyk7CisJCXVzbGVlcF9yYW5nZShkZWxheV9tc2VjICogMTAwMCwgKGRl bGF5X21zZWMgKiAxMDAwKSArIDMwMCk7CiAJfSB3aGlsZSAoLS1sb29wcyA+IDApOwogCiAJcmV0 dXJuIHJjOwpAQCAtOTc3LDcgKzk3Nyw3IEBAIGludCB3YWl0X2Zvcl90cG1fc3RhdChzdHJ1Y3Qg dHBtX2NoaXAgKmNoaXAsIHU4IG1hc2ssIHVuc2lnbmVkIGxvbmcgdGltZW91dCwKIAkJfQogCX0g ZWxzZSB7CiAJCWRvIHsKLQkJCW1zbGVlcChUUE1fVElNRU9VVCk7CisJCQl1c2xlZXBfcmFuZ2Uo VFBNX1RJTUVPVVQgKiAxMDAwLCAoVFBNX1RJTUVPVVQgKiAxMDAwKSArIDMwMCk7CiAJCQlzdGF0 dXMgPSBjaGlwLT5vcHMtPnN0YXR1cyhjaGlwKTsKIAkJCWlmICgoc3RhdHVzICYgbWFzaykgPT0g bWFzaykKIAkJCQlyZXR1cm4gMDsKQEAgLTEwNDUsNyArMTA0NSw3IEBAIGludCB0cG1fcG1fc3Vz cGVuZChzdHJ1Y3QgZGV2aWNlICpkZXYpCiAJCSAqLwogCQlpZiAocmMgIT0gVFBNX1dBUk5fUkVU UlkpCiAJCQlicmVhazsKLQkJbXNsZWVwKFRQTV9USU1FT1VUX1JFVFJZKTsKKwkJdXNsZWVwX3Jh bmdlKFRQTV9USU1FT1VUX1JFVFJZICogMTAwMCwgKFRQTV9USU1FT1VUX1JFVFJZICogMTAwMCkg KyAzMDApOwogCX0KIAogCWlmIChyYykKZGlmZiAtLWdpdCBhL2RyaXZlcnMvY2hhci90cG0vdHBt Mi1jbWQuYyBiL2RyaXZlcnMvY2hhci90cG0vdHBtMi1jbWQuYwppbmRleCA4ODFhZWE5Li5iZGFl MjA1IDEwMDY0NAotLS0gYS9kcml2ZXJzL2NoYXIvdHBtL3RwbTItY21kLmMKKysrIGIvZHJpdmVy cy9jaGFyL3RwbS90cG0yLWNtZC5jCkBAIC05NjEsNyArOTYxLDcgQEAgc3RhdGljIGludCB0cG0y X2RvX3NlbGZ0ZXN0KHN0cnVjdCB0cG1fY2hpcCAqY2hpcCkKIAkJaWYgKHJjICE9IFRQTTJfUkNf VEVTVElORykKIAkJCWJyZWFrOwogCi0JCW1zbGVlcChkZWxheV9tc2VjKTsKKwkJdXNsZWVwX3Jh bmdlKGRlbGF5X21zZWMgKiAxMDAwLCAoZGVsYXlfbXNlYyAqIDEwMDApICsgMzAwKTsKIAl9CiAK IAlyZXR1cm4gcmM7CmRpZmYgLS1naXQgYS9kcml2ZXJzL2NoYXIvdHBtL3RwbV9pbmZpbmVvbi5j IGIvZHJpdmVycy9jaGFyL3RwbS90cG1faW5maW5lb24uYwppbmRleCBlM2NmOWYzLi44NmYwZjNl IDEwMDY0NAotLS0gYS9kcml2ZXJzL2NoYXIvdHBtL3RwbV9pbmZpbmVvbi5jCisrKyBiL2RyaXZl cnMvY2hhci90cG0vdHBtX2luZmluZW9uLmMKQEAgLTE5MSw3ICsxOTEsNyBAQCBzdGF0aWMgaW50 IHdhaXQoc3RydWN0IHRwbV9jaGlwICpjaGlwLCBpbnQgd2FpdF9mb3JfYml0KQogCQkvKiBjaGVj ayB0aGUgc3RhdHVzLXJlZ2lzdGVyIGlmIHdhaXRfZm9yX2JpdCBpcyBzZXQgKi8KIAkJaWYgKHN0 YXR1cyAmIDEgPDwgd2FpdF9mb3JfYml0KQogCQkJYnJlYWs7Ci0JCW1zbGVlcChUUE1fTVNMRUVQ X1RJTUUpOworCQl1c2xlZXBfcmFuZ2UoVFBNX01TTEVFUF9USU1FICogMTAwMCwgKFRQTV9NU0xF RVBfVElNRSAqIDEwMDApICsgMzAwKTsKIAl9CiAJaWYgKGkgPT0gVFBNX01BWF9UUklFUykgewkv KiB0aW1lb3V0IG9jY3VycyAqLwogCQlpZiAod2FpdF9mb3JfYml0ID09IFNUQVRfWEZFKQpAQCAt MjI2LDcgKzIyNiw3IEBAIHN0YXRpYyB2b2lkIHRwbV93dHgoc3RydWN0IHRwbV9jaGlwICpjaGlw KQogCXdhaXRfYW5kX3NlbmQoY2hpcCwgVFBNX0NUUkxfV1RYKTsKIAl3YWl0X2FuZF9zZW5kKGNo aXAsIDB4MDApOwogCXdhaXRfYW5kX3NlbmQoY2hpcCwgMHgwMCk7Ci0JbXNsZWVwKFRQTV9XVFhf TVNMRUVQX1RJTUUpOworCXVzbGVlcF9yYW5nZShUUE1fV1RYX01TTEVFUF9USU1FICogMTAwMCwg KFRQTV9XVFhfTVNMRUVQX1RJTUUgKiAxMDAwKSArIDMwMCk7CiB9CiAKIHN0YXRpYyB2b2lkIHRw bV93dHhfYWJvcnQoc3RydWN0IHRwbV9jaGlwICpjaGlwKQpAQCAtMjM3LDcgKzIzNyw3IEBAIHN0 YXRpYyB2b2lkIHRwbV93dHhfYWJvcnQoc3RydWN0IHRwbV9jaGlwICpjaGlwKQogCXdhaXRfYW5k X3NlbmQoY2hpcCwgMHgwMCk7CiAJd2FpdF9hbmRfc2VuZChjaGlwLCAweDAwKTsKIAludW1iZXJf b2Zfd3R4ID0gMDsKLQltc2xlZXAoVFBNX1dUWF9NU0xFRVBfVElNRSk7CisJdXNsZWVwX3Jhbmdl KFRQTV9XVFhfTVNMRUVQX1RJTUUgKiAxMDAwLCAoVFBNX1dUWF9NU0xFRVBfVElNRSAqIDEwMDAp ICsgMzAwKTsKIH0KIAogc3RhdGljIGludCB0cG1faW5mX3JlY3Yoc3RydWN0IHRwbV9jaGlwICpj aGlwLCB1OCAqIGJ1Ziwgc2l6ZV90IGNvdW50KQpkaWZmIC0tZ2l0IGEvZHJpdmVycy9jaGFyL3Rw bS90cG1fdGlzX2NvcmUuYyBiL2RyaXZlcnMvY2hhci90cG0vdHBtX3Rpc19jb3JlLmMKaW5kZXgg YzBmMjk2Yi4uZTdkMzA0NiAxMDA2NDQKLS0tIGEvZHJpdmVycy9jaGFyL3RwbS90cG1fdGlzX2Nv cmUuYworKysgYi9kcml2ZXJzL2NoYXIvdHBtL3RwbV90aXNfY29yZS5jCkBAIC01MSw3ICs1MSw3 IEBAIHN0YXRpYyBpbnQgd2FpdF9zdGFydHVwKHN0cnVjdCB0cG1fY2hpcCAqY2hpcCwgaW50IGwp CiAKIAkJaWYgKGFjY2VzcyAmIFRQTV9BQ0NFU1NfVkFMSUQpCiAJCQlyZXR1cm4gMDsKLQkJbXNs ZWVwKFRQTV9USU1FT1VUKTsKKwkJdXNsZWVwX3JhbmdlKFRQTV9USU1FT1VUICogMTAwMCwgKFRQ TV9USU1FT1VUICogMTAwMCkgKyAzMDApOwogCX0gd2hpbGUgKHRpbWVfYmVmb3JlKGppZmZpZXMs IHN0b3ApKTsKIAlyZXR1cm4gLTE7CiB9CkBAIC0xMjUsNyArMTI1LDcgQEAgc3RhdGljIGludCBy ZXF1ZXN0X2xvY2FsaXR5KHN0cnVjdCB0cG1fY2hpcCAqY2hpcCwgaW50IGwpCiAJCWRvIHsKIAkJ CWlmIChjaGVja19sb2NhbGl0eShjaGlwLCBsKSA+PSAwKQogCQkJCXJldHVybiBsOwotCQkJbXNs ZWVwKFRQTV9USU1FT1VUKTsKKwkJCXVzbGVlcF9yYW5nZShUUE1fVElNRU9VVCAqIDEwMDAsIChU UE1fVElNRU9VVCAqIDEwMDApICsgMzAwKTsKIAkJfSB3aGlsZSAodGltZV9iZWZvcmUoamlmZmll cywgc3RvcCkpOwogCX0KIAlyZXR1cm4gLTE7CkBAIC0xNzAsNyArMTcwLDcgQEAgc3RhdGljIGlu dCBnZXRfYnVyc3Rjb3VudChzdHJ1Y3QgdHBtX2NoaXAgKmNoaXApCiAJCWJ1cnN0Y250ID0gKHZh bHVlID4+IDgpICYgMHhGRkZGOwogCQlpZiAoYnVyc3RjbnQpCiAJCQlyZXR1cm4gYnVyc3RjbnQ7 Ci0JCW1zbGVlcChUUE1fVElNRU9VVCk7CisJCXVzbGVlcF9yYW5nZShUUE1fVElNRU9VVCAqIDEw MDAsIChUUE1fVElNRU9VVCAqIDEwMDApICsgMzAwKTsKIAl9IHdoaWxlICh0aW1lX2JlZm9yZShq aWZmaWVzLCBzdG9wKSk7CiAJcmV0dXJuIC1FQlVTWTsKIH0KQEAgLTQwOCw3ICs0MDgsNyBAQCBz dGF0aWMgaW50IHRwbV90aXNfc2VuZChzdHJ1Y3QgdHBtX2NoaXAgKmNoaXAsIHU4ICpidWYsIHNp emVfdCBsZW4pCiAJcHJpdi0+aXJxID0gaXJxOwogCWNoaXAtPmZsYWdzIHw9IFRQTV9DSElQX0ZM QUdfSVJROwogCWlmICghcHJpdi0+aXJxX3Rlc3RlZCkKLQkJbXNsZWVwKDEpOworCQl1c2xlZXBf cmFuZ2UoMTAwMCwgMTMwMCk7CiAJaWYgKCFwcml2LT5pcnFfdGVzdGVkKQogCQlkaXNhYmxlX2lu dGVycnVwdHMoY2hpcCk7CiAJcHJpdi0+aXJxX3Rlc3RlZCA9IHRydWU7Cg== --_004_DF4PR84MB00286E6F6F1E51633C8D65C4F2F00DF4PR84MB0028NAMP_ Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot --_004_DF4PR84MB00286E6F6F1E51633C8D65C4F2F00DF4PR84MB0028NAMP_ Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ tpmdd-devel mailing list tpmdd-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org https://lists.sourceforge.net/lists/listinfo/tpmdd-devel --_004_DF4PR84MB00286E6F6F1E51633C8D65C4F2F00DF4PR84MB0028NAMP_-- From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Attak, Hamza" Subject: [PATCH] msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers Date: Mon, 10 Jul 2017 13:26:51 +0000 Message-ID: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Language: en-US List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: tpmdd-devel-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org To: "tpmdd-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org" Cc: "Jacquin, Ludovic" , "Edwards, Nigel" List-Id: tpmdd-devel@lists.sourceforge.net msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers Following the recent Nuvoton patch submitted by Mimi Zohar: https://sourceforge.net/p/tpmdd/mailman/message/35685260/ Based on this work, I decided to apply a similar modification to the generic TPM drivers in the kernel. The patch simply replaces all msleep function calls with usleep_range calls in the generic drivers. As a former patch already exists for the Nuvoton TPM, it is therefore not included in this new patch. Tested with an Infineon TPM 1.2, using the generic tpm-tis module, for a thousand PCR extends, we see results going from 1m57s unpatched to 40s with the new patch. We obtain similar results when using the original and patched tpm_infineon driver, which is also part of the patch. Similarly with a STM TPM 2.0, using the CRB driver, it takes about 20ms per extend unpatched and around 7ms with the new patch. Note that the PCR consistency is untouched with this patch, each TPM has been tested with 10 million extends and the aggregated PCR value is continuously verified to be correct. As an extension of this work, this could potentially and easily be applied to other vendor's drivers. Still, these changes are not included in the proposed patch as they are untested. Signed-off-by: Hamza Attak --- drivers/char/tpm/tpm-interface.c | 10 +++++----- drivers/char/tpm/tpm2-cmd.c | 2 +- drivers/char/tpm/tpm_infineon.c | 6 +++--- drivers/char/tpm/tpm_tis_core.c | 8 ++++---- 4 files changed, 13 insertions(+), 13 deletions(-) diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c index bd2128e..f03daf5 100644 --- a/drivers/char/tpm/tpm-interface.c +++ b/drivers/char/tpm/tpm-interface.c @@ -395,7 +395,7 @@ ssize_t tpm_transmit(struct tpm_chip *chip, const u8 *buf, size_t bufsiz, goto out; } - msleep(TPM_TIMEOUT); /* CHECK */ + usleep_range(TPM_TIMEOUT * 1000, (TPM_TIMEOUT * 1000) + 300); rmb(); } while (time_before(jiffies, stop)); @@ -862,7 +862,7 @@ int tpm_do_selftest(struct tpm_chip *chip) dev_info( &chip->dev, HW_ERR "TPM command timed out during continue self test"); - msleep(delay_msec); + usleep_range(delay_msec * 1000, (delay_msec * 1000) + 300); continue; } @@ -877,7 +877,7 @@ int tpm_do_selftest(struct tpm_chip *chip) } if (rc != TPM_WARN_DOING_SELFTEST) return rc; - msleep(delay_msec); + usleep_range(delay_msec * 1000, (delay_msec * 1000) + 300); } while (--loops > 0); return rc; @@ -977,7 +977,7 @@ again: } } else { do { - msleep(TPM_TIMEOUT); + usleep_range(TPM_TIMEOUT * 1000, (TPM_TIMEOUT * 1000) + 300); status = chip->ops->status(chip); if ((status & mask) == mask) return 0; @@ -1045,7 +1045,7 @@ int tpm_pm_suspend(struct device *dev) */ if (rc != TPM_WARN_RETRY) break; - msleep(TPM_TIMEOUT_RETRY); + usleep_range(TPM_TIMEOUT_RETRY * 1000, (TPM_TIMEOUT_RETRY * 1000) + 300); } if (rc) diff --git a/drivers/char/tpm/tpm2-cmd.c b/drivers/char/tpm/tpm2-cmd.c index 881aea9..bdae205 100644 --- a/drivers/char/tpm/tpm2-cmd.c +++ b/drivers/char/tpm/tpm2-cmd.c @@ -961,7 +961,7 @@ static int tpm2_do_selftest(struct tpm_chip *chip) if (rc != TPM2_RC_TESTING) break; - msleep(delay_msec); + usleep_range(delay_msec * 1000, (delay_msec * 1000) + 300); } return rc; diff --git a/drivers/char/tpm/tpm_infineon.c b/drivers/char/tpm/tpm_infineon.c index e3cf9f3..86f0f3e 100644 --- a/drivers/char/tpm/tpm_infineon.c +++ b/drivers/char/tpm/tpm_infineon.c @@ -191,7 +191,7 @@ static int wait(struct tpm_chip *chip, int wait_for_bit) /* check the status-register if wait_for_bit is set */ if (status & 1 << wait_for_bit) break; - msleep(TPM_MSLEEP_TIME); + usleep_range(TPM_MSLEEP_TIME * 1000, (TPM_MSLEEP_TIME * 1000) + 300); } if (i == TPM_MAX_TRIES) { /* timeout occurs */ if (wait_for_bit == STAT_XFE) @@ -226,7 +226,7 @@ static void tpm_wtx(struct tpm_chip *chip) wait_and_send(chip, TPM_CTRL_WTX); wait_and_send(chip, 0x00); wait_and_send(chip, 0x00); - msleep(TPM_WTX_MSLEEP_TIME); + usleep_range(TPM_WTX_MSLEEP_TIME * 1000, (TPM_WTX_MSLEEP_TIME * 1000) + 300); } static void tpm_wtx_abort(struct tpm_chip *chip) @@ -237,7 +237,7 @@ static void tpm_wtx_abort(struct tpm_chip *chip) wait_and_send(chip, 0x00); wait_and_send(chip, 0x00); number_of_wtx = 0; - msleep(TPM_WTX_MSLEEP_TIME); + usleep_range(TPM_WTX_MSLEEP_TIME * 1000, (TPM_WTX_MSLEEP_TIME * 1000) + 300); } static int tpm_inf_recv(struct tpm_chip *chip, u8 * buf, size_t count) diff --git a/drivers/char/tpm/tpm_tis_core.c b/drivers/char/tpm/tpm_tis_core.c index c0f296b..e7d3046 100644 --- a/drivers/char/tpm/tpm_tis_core.c +++ b/drivers/char/tpm/tpm_tis_core.c @@ -51,7 +51,7 @@ static int wait_startup(struct tpm_chip *chip, int l) if (access & TPM_ACCESS_VALID) return 0; - msleep(TPM_TIMEOUT); + usleep_range(TPM_TIMEOUT * 1000, (TPM_TIMEOUT * 1000) + 300); } while (time_before(jiffies, stop)); return -1; } @@ -125,7 +125,7 @@ again: do { if (check_locality(chip, l) >= 0) return l; - msleep(TPM_TIMEOUT); + usleep_range(TPM_TIMEOUT * 1000, (TPM_TIMEOUT * 1000) + 300); } while (time_before(jiffies, stop)); } return -1; @@ -170,7 +170,7 @@ static int get_burstcount(struct tpm_chip *chip) burstcnt = (value >> 8) & 0xFFFF; if (burstcnt) return burstcnt; - msleep(TPM_TIMEOUT); + usleep_range(TPM_TIMEOUT * 1000, (TPM_TIMEOUT * 1000) + 300); } while (time_before(jiffies, stop)); return -EBUSY; } @@ -408,7 +408,7 @@ static int tpm_tis_send(struct tpm_chip *chip, u8 *buf, size_t len) priv->irq = irq; chip->flags |= TPM_CHIP_FLAG_IRQ; if (!priv->irq_tested) - msleep(1); + usleep_range(1000, 1300); if (!priv->irq_tested) disable_interrupts(chip); priv->irq_tested = true; -- 2.7.4 ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jarkko Sakkinen Subject: Re: [PATCH] msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers Date: Mon, 10 Jul 2017 22:15:27 +0300 Message-ID: <20170710191527.qlurrqit6naoaotp@linux.intel.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Disposition: inline In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: tpmdd-devel-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org To: "Attak, Hamza" Cc: "Jacquin, Ludovic" , "tpmdd-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org" , "Edwards, Nigel" List-Id: tpmdd-devel@lists.sourceforge.net On Mon, Jul 10, 2017 at 01:26:51PM +0000, Attak, Hamza wrote: > msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers > > Following the recent Nuvoton patch submitted by Mimi Zohar: > https://sourceforge.net/p/tpmdd/mailman/message/35685260/ 1. Please do not your work on a patch that was never applied. This one was https://patchwork.kernel.org/patch/9617835/ 2. Please break your lines properly in the commit message 3. Please use commit ID in the mainline instead of URL. Have you run scripts/checkpatch.pl? /Jarkko ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Attak, Hamza" Subject: Re: [PATCH] msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers Date: Wed, 12 Jul 2017 12:33:17 +0000 Message-ID: References: <20170710191527.qlurrqit6naoaotp@linux.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20170710191527.qlurrqit6naoaotp-VuQAYsv1563Yd54FQh9/CA@public.gmane.org> Content-Language: en-US List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: tpmdd-devel-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org To: Jarkko Sakkinen Cc: "Jacquin, Ludovic" , "tpmdd-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org" , "Edwards, Nigel" List-Id: tpmdd-devel@lists.sourceforge.net 1. I guess you meant not to "base" my work on an unapplied patch. The patch we are submitting is not dependant on the referenced, earlier patch. It is also not including any of its changes. The former patch was only mentioned to give context, as it is a similar approach, however it is addressing another problem (Nuvoton driver). This reference can as well be removed if misleading. 2. Linebreaks in commit message should be fixed now. 3. The URL has been removed, so as the reference to the former patch. No, I didn't run checkpatch.pl when first submitting the patch. Testing it with checkpatch.pl, the output gave me warning about lines over 80 chars. They are now solved. I am sending an updated version of the patch. Thanks, Hamza ATTAK. -----Original Message----- From: Jarkko Sakkinen [mailto:jarkko.sakkinen-VuQAYsv1563Yd54FQh9/CA@public.gmane.org] Sent: 10 July 2017 20:15 To: Attak, Hamza Cc: tpmdd-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org; Jacquin, Ludovic ; Edwards, Nigel Subject: Re: [tpmdd-devel] [PATCH] msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers On Mon, Jul 10, 2017 at 01:26:51PM +0000, Attak, Hamza wrote: > msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic > drivers > > Following the recent Nuvoton patch submitted by Mimi Zohar: > https://sourceforge.net/p/tpmdd/mailman/message/35685260/ 1. Please do not your work on a patch that was never applied. This one was https://patchwork.kernel.org/patch/9617835/ 2. Please break your lines properly in the commit message 3. Please use commit ID in the mainline instead of URL. Have you run scripts/checkpatch.pl? /Jarkko ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Attak, Hamza" Subject: Re: [PATCH] msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers Date: Wed, 12 Jul 2017 12:34:19 +0000 Message-ID: References: <20170710191527.qlurrqit6naoaotp@linux.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20170710191527.qlurrqit6naoaotp-VuQAYsv1563Yd54FQh9/CA@public.gmane.org> Content-Language: en-US List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: tpmdd-devel-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org To: Jarkko Sakkinen Cc: "Jacquin, Ludovic" , "tpmdd-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org" , "Edwards, Nigel" List-Id: tpmdd-devel@lists.sourceforge.net msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers The patch simply replaces all msleep function calls with usleep_range calls in the generic drivers. Tested with an Infineon TPM 1.2, using the generic tpm-tis module, for a thousand PCR extends, we see results going from 1m57s unpatched to 40s with the new patch. We obtain similar results when using the original and patched tpm_infineon driver, which is also part of the patch. Similarly with a STM TPM 2.0, using the CRB driver, it takes about 20ms per extend unpatched and around 7ms with the new patch. Note that the PCR consistency is untouched with this patch, each TPM has been tested with 10 million extends and the aggregated PCR value is continuously verified to be correct. As an extension of this work, this could potentially and easily be applied to other vendor's drivers. Still, these changes are not included in the proposed patch as they are untested. Signed-off-by: Hamza Attak --- drivers/char/tpm/tpm-interface.c | 13 ++++++++----- drivers/char/tpm/tpm2-cmd.c | 2 +- drivers/char/tpm/tpm_infineon.c | 9 ++++++--- drivers/char/tpm/tpm_tis_core.c | 9 +++++---- 4 files changed, 20 insertions(+), 13 deletions(-) diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c index bd2128e..513b801 100644 --- a/drivers/char/tpm/tpm-interface.c +++ b/drivers/char/tpm/tpm-interface.c @@ -395,7 +395,7 @@ ssize_t tpm_transmit(struct tpm_chip *chip, const u8 *buf, size_t bufsiz, goto out; } - msleep(TPM_TIMEOUT); /* CHECK */ + usleep_range(TPM_TIMEOUT * 1000, (TPM_TIMEOUT * 1000) + 300); rmb(); } while (time_before(jiffies, stop)); @@ -862,7 +862,8 @@ int tpm_do_selftest(struct tpm_chip *chip) dev_info( &chip->dev, HW_ERR "TPM command timed out during continue self test"); - msleep(delay_msec); + usleep_range(delay_msec * 1000, + (delay_msec * 1000) + 300); continue; } @@ -877,7 +878,7 @@ int tpm_do_selftest(struct tpm_chip *chip) } if (rc != TPM_WARN_DOING_SELFTEST) return rc; - msleep(delay_msec); + usleep_range(delay_msec * 1000, (delay_msec * 1000) + 300); } while (--loops > 0); return rc; @@ -977,7 +978,8 @@ again: } } else { do { - msleep(TPM_TIMEOUT); + usleep_range(TPM_TIMEOUT * 1000, + (TPM_TIMEOUT * 1000) + 300); status = chip->ops->status(chip); if ((status & mask) == mask) return 0; @@ -1045,7 +1047,8 @@ int tpm_pm_suspend(struct device *dev) */ if (rc != TPM_WARN_RETRY) break; - msleep(TPM_TIMEOUT_RETRY); + usleep_range(TPM_TIMEOUT_RETRY * 1000, + (TPM_TIMEOUT_RETRY * 1000) + 300); } if (rc) diff --git a/drivers/char/tpm/tpm2-cmd.c b/drivers/char/tpm/tpm2-cmd.c index 881aea9..bdae205 100644 --- a/drivers/char/tpm/tpm2-cmd.c +++ b/drivers/char/tpm/tpm2-cmd.c @@ -961,7 +961,7 @@ static int tpm2_do_selftest(struct tpm_chip *chip) if (rc != TPM2_RC_TESTING) break; - msleep(delay_msec); + usleep_range(delay_msec * 1000, (delay_msec * 1000) + 300); } return rc; diff --git a/drivers/char/tpm/tpm_infineon.c b/drivers/char/tpm/tpm_infineon.c index e3cf9f3..1846151 100644 --- a/drivers/char/tpm/tpm_infineon.c +++ b/drivers/char/tpm/tpm_infineon.c @@ -191,7 +191,8 @@ static int wait(struct tpm_chip *chip, int wait_for_bit) /* check the status-register if wait_for_bit is set */ if (status & 1 << wait_for_bit) break; - msleep(TPM_MSLEEP_TIME); + usleep_range(TPM_MSLEEP_TIME * 1000, + (TPM_MSLEEP_TIME * 1000) + 300); } if (i == TPM_MAX_TRIES) { /* timeout occurs */ if (wait_for_bit == STAT_XFE) @@ -226,7 +227,8 @@ static void tpm_wtx(struct tpm_chip *chip) wait_and_send(chip, TPM_CTRL_WTX); wait_and_send(chip, 0x00); wait_and_send(chip, 0x00); - msleep(TPM_WTX_MSLEEP_TIME); + usleep_range(TPM_WTX_MSLEEP_TIME * 1000, + (TPM_WTX_MSLEEP_TIME * 1000) + 300); } static void tpm_wtx_abort(struct tpm_chip *chip) @@ -237,7 +239,8 @@ static void tpm_wtx_abort(struct tpm_chip *chip) wait_and_send(chip, 0x00); wait_and_send(chip, 0x00); number_of_wtx = 0; - msleep(TPM_WTX_MSLEEP_TIME); + usleep_range(TPM_WTX_MSLEEP_TIME * 1000, + (TPM_WTX_MSLEEP_TIME * 1000) + 300); } static int tpm_inf_recv(struct tpm_chip *chip, u8 * buf, size_t count) diff --git a/drivers/char/tpm/tpm_tis_core.c b/drivers/char/tpm/tpm_tis_core.c index c0f296b..62c0915 100644 --- a/drivers/char/tpm/tpm_tis_core.c +++ b/drivers/char/tpm/tpm_tis_core.c @@ -51,7 +51,7 @@ static int wait_startup(struct tpm_chip *chip, int l) if (access & TPM_ACCESS_VALID) return 0; - msleep(TPM_TIMEOUT); + usleep_range(TPM_TIMEOUT * 1000, (TPM_TIMEOUT * 1000) + 300); } while (time_before(jiffies, stop)); return -1; } @@ -125,7 +125,8 @@ again: do { if (check_locality(chip, l) >= 0) return l; - msleep(TPM_TIMEOUT); + usleep_range(TPM_TIMEOUT * 1000, + (TPM_TIMEOUT * 1000) + 300); } while (time_before(jiffies, stop)); } return -1; @@ -170,7 +171,7 @@ static int get_burstcount(struct tpm_chip *chip) burstcnt = (value >> 8) & 0xFFFF; if (burstcnt) return burstcnt; - msleep(TPM_TIMEOUT); + usleep_range(TPM_TIMEOUT * 1000, (TPM_TIMEOUT * 1000) + 300); } while (time_before(jiffies, stop)); return -EBUSY; } @@ -408,7 +409,7 @@ static int tpm_tis_send(struct tpm_chip *chip, u8 *buf, size_t len) priv->irq = irq; chip->flags |= TPM_CHIP_FLAG_IRQ; if (!priv->irq_tested) - msleep(1); + usleep_range(1000, 1300); if (!priv->irq_tested) disable_interrupts(chip); priv->irq_tested = true; -- 2.7.4 ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jarkko Sakkinen Subject: Re: [PATCH] msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers Date: Sun, 16 Jul 2017 13:49:07 +0300 Message-ID: <20170716104907.foav6tnuda6gubca@linux.intel.com> References: <20170710191527.qlurrqit6naoaotp@linux.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Disposition: inline In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: tpmdd-devel-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org To: "Attak, Hamza" Cc: "Jacquin, Ludovic" , "tpmdd-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org" , "Edwards, Nigel" List-Id: tpmdd-devel@lists.sourceforge.net Your subject line looks wrong. This cannot be cleanly applied with git am. Please go through https://kernelnewbies.org/FirstKernelPatch On Wed, Jul 12, 2017 at 12:34:19PM +0000, Attak, Hamza wrote: > msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers > > The patch simply replaces all msleep function calls with usleep_range calls in > the generic drivers. > > Tested with an Infineon TPM 1.2, using the generic tpm-tis module, for a > thousand PCR extends, we see results going from 1m57s unpatched to 40s with > the new patch. We obtain similar results when using the original and patched > tpm_infineon driver, which is also part of the patch. Similarly with a STM TPM > 2.0, using the CRB driver, it takes about 20ms per extend unpatched and around > 7ms with the new patch. > > Note that the PCR consistency is untouched with this patch, each TPM has been > tested with 10 million extends and the aggregated PCR value is continuously > verified to be correct. > > As an extension of this work, this could potentially and easily be applied to > other vendor's drivers. Still, these changes are not included in the proposed > patch as they are untested. > > Signed-off-by: Hamza Attak > --- > drivers/char/tpm/tpm-interface.c | 13 ++++++++----- > drivers/char/tpm/tpm2-cmd.c | 2 +- > drivers/char/tpm/tpm_infineon.c | 9 ++++++--- > drivers/char/tpm/tpm_tis_core.c | 9 +++++---- > 4 files changed, 20 insertions(+), 13 deletions(-) > > diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c > index bd2128e..513b801 100644 > --- a/drivers/char/tpm/tpm-interface.c > +++ b/drivers/char/tpm/tpm-interface.c > @@ -395,7 +395,7 @@ ssize_t tpm_transmit(struct tpm_chip *chip, const u8 *buf, size_t bufsiz, > goto out; > } > > - msleep(TPM_TIMEOUT); /* CHECK */ > + usleep_range(TPM_TIMEOUT * 1000, (TPM_TIMEOUT * 1000) + 300); Plese look into. https://patchwork.kernel.org/patch/9617823/ Now you are sprinkling magic numbers everywhere. > rmb(); > } while (time_before(jiffies, stop)); > > @@ -862,7 +862,8 @@ int tpm_do_selftest(struct tpm_chip *chip) > dev_info( > &chip->dev, HW_ERR > "TPM command timed out during continue self test"); > - msleep(delay_msec); > + usleep_range(delay_msec * 1000, > + (delay_msec * 1000) + 300); The second line dooes not look properly aligned. There's probably similar style issues elsewhere in the patch (at least spotted one after this). /Jarkko ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jarkko Sakkinen Subject: Re: [PATCH] msleep() delays - replace with usleep_range() in TPM 1.2/2.0 generic drivers Date: Wed, 2 Aug 2017 15:28:52 +0300 Message-ID: <20170802122852.jmtsihjd6k4ezems@linux.intel.com> References: <20170726174642.GA16821@dev-HP-EliteBook-Folio-1040-G1> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Disposition: inline In-Reply-To: <20170726174642.GA16821@dev-HP-EliteBook-Folio-1040-G1> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: tpmdd-devel-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org To: Hamza Attak Cc: ludo-ZPxbGqLxI0U@public.gmane.org, tpmdd-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org, nigel.edwards-ZPxbGqLxI0U@public.gmane.org List-Id: tpmdd-devel@lists.sourceforge.net On Wed, Jul 26, 2017 at 06:46:42PM +0100, Hamza Attak wrote: > The patch simply replaces all msleep function calls with usleep_range calls > in the generic drivers. > > Tested with an Infineon TPM 1.2, using the generic tpm-tis module, for a > thousand PCR extends, we see results going from 1m57s unpatched to 40s > with the new patch. We obtain similar results when using the original and > patched tpm_infineon driver, which is also part of the patch. > Similarly with a STM TPM 2.0, using the CRB driver, it takes about 20ms per > extend unpatched and around 7ms with the new patch. > > Note that the PCR consistency is untouched with this patch, each TPM has > been tested with 10 million extends and the aggregated PCR value is > continuously verified to be correct. > > As an extension of this work, this could potentially and easily be applied > to other vendor's drivers. Still, these changes are not included in the > proposed patch as they are untested. > > Signed-off-by: Hamza Attak I think this would be a lot cleaner that we would have void tpm_msleep(unsigned long msecs) function that would call usleep_range(). The change makes in high-level sense but the implementation makes the driver code a mess. > --- > drivers/char/tpm/tpm-interface.c | 18 +++++++++++------- > drivers/char/tpm/tpm.h | 11 +++++++++-- > drivers/char/tpm/tpm2-cmd.c | 6 +++--- > drivers/char/tpm/tpm_infineon.c | 17 ++++++++++------- > drivers/char/tpm/tpm_tis_core.c | 11 +++++++---- > 5 files changed, 40 insertions(+), 23 deletions(-) > > diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c > index bd2128e..95276d1 100644 > --- a/drivers/char/tpm/tpm-interface.c > +++ b/drivers/char/tpm/tpm-interface.c > @@ -395,7 +395,8 @@ ssize_t tpm_transmit(struct tpm_chip *chip, const u8 *buf, size_t bufsiz, > goto out; > } > > - msleep(TPM_TIMEOUT); /* CHECK */ > + usleep_range(TPM_TIMEOUT_US, > + TPM_TIMEOUT_US + TPM_TIMEOUT_RANGE_US); > rmb(); > } while (time_before(jiffies, stop)); > > @@ -836,13 +837,13 @@ int tpm_do_selftest(struct tpm_chip *chip) > { > int rc; > unsigned int loops; > - unsigned int delay_msec = 100; > + unsigned int delay_usec = 100000; > unsigned long duration; > u8 dummy[TPM_DIGEST_SIZE]; > > duration = tpm_calc_ordinal_duration(chip, TPM_ORD_CONTINUE_SELFTEST); > > - loops = jiffies_to_msecs(duration) / delay_msec; > + loops = jiffies_to_usecs(duration) / delay_usec; > > rc = tpm_continue_selftest(chip); > /* This may fail if there was no TPM driver during a suspend/resume > @@ -862,7 +863,8 @@ int tpm_do_selftest(struct tpm_chip *chip) > dev_info( > &chip->dev, HW_ERR > "TPM command timed out during continue self test"); > - msleep(delay_msec); > + usleep_range(delay_usec, > + delay_usec + TPM_TIMEOUT_RANGE_US); > continue; > } > > @@ -877,7 +879,7 @@ int tpm_do_selftest(struct tpm_chip *chip) > } > if (rc != TPM_WARN_DOING_SELFTEST) > return rc; > - msleep(delay_msec); > + usleep_range(delay_usec, delay_usec + TPM_TIMEOUT_RANGE_US); > } while (--loops > 0); > > return rc; > @@ -977,7 +979,8 @@ again: > } > } else { > do { > - msleep(TPM_TIMEOUT); > + usleep_range(TPM_TIMEOUT_US, > + TPM_TIMEOUT_US + TPM_TIMEOUT_RANGE_US); > status = chip->ops->status(chip); > if ((status & mask) == mask) > return 0; > @@ -1045,7 +1048,8 @@ int tpm_pm_suspend(struct device *dev) > */ > if (rc != TPM_WARN_RETRY) > break; > - msleep(TPM_TIMEOUT_RETRY); > + usleep_range(TPM_TIMEOUT_RETRY_US, > + TPM_TIMEOUT_RETRY_US + TPM_TIMEOUT_RANGE_US); > } > > if (rc) > diff --git a/drivers/char/tpm/tpm.h b/drivers/char/tpm/tpm.h > index 4937b56..c1ace6c 100644 > --- a/drivers/char/tpm/tpm.h > +++ b/drivers/char/tpm/tpm.h > @@ -45,8 +45,15 @@ enum tpm_const { > }; > > enum tpm_timeout { > - TPM_TIMEOUT = 5, /* msecs */ > - TPM_TIMEOUT_RETRY = 100 /* msecs */ > + TPM_TIMEOUT = 5, /* msecs */ > + TPM_TIMEOUT_RETRY = 100, /* msecs */ > + /* Ideally all the drivers should be using the usecs values, > + * the msecs values are only kept there for compatibility purposes > + * with the remaining untested drivers. > + */ > + TPM_TIMEOUT_US = TPM_TIMEOUT * 1000, /* usecs */ > + TPM_TIMEOUT_RETRY_US = TPM_TIMEOUT_RETRY * 1000, /* usecs */ > + TPM_TIMEOUT_RANGE_US = 300 /* usecs */ These renames are not needed and the comment doesn't make any sense when you have _US postfix. Just keep them in msecs and implement the wrapper. The exception is of course the new constant TPM_TIMEOUT_RANGE_US, which you need to implement tpm_msleep(). > }; > > /* TPM addresses */ > diff --git a/drivers/char/tpm/tpm2-cmd.c b/drivers/char/tpm/tpm2-cmd.c > index 881aea9..425caa4 100644 > --- a/drivers/char/tpm/tpm2-cmd.c > +++ b/drivers/char/tpm/tpm2-cmd.c > @@ -930,14 +930,14 @@ static int tpm2_do_selftest(struct tpm_chip *chip) > { > int rc; > unsigned int loops; > - unsigned int delay_msec = 100; > + unsigned int delay_usec = 100000; > unsigned long duration; > struct tpm2_cmd cmd; > int i; > > duration = tpm2_calc_ordinal_duration(chip, TPM2_CC_SELF_TEST); > > - loops = jiffies_to_msecs(duration) / delay_msec; > + loops = jiffies_to_usecs(duration) / delay_usec; > > rc = tpm2_start_selftest(chip, true); > if (rc) > @@ -961,7 +961,7 @@ static int tpm2_do_selftest(struct tpm_chip *chip) > if (rc != TPM2_RC_TESTING) > break; > > - msleep(delay_msec); > + usleep_range(delay_usec, delay_usec + TPM_TIMEOUT_RANGE_US); > } > > return rc; > diff --git a/drivers/char/tpm/tpm_infineon.c b/drivers/char/tpm/tpm_infineon.c > index e3cf9f3..c9d2d13 100644 > --- a/drivers/char/tpm/tpm_infineon.c > +++ b/drivers/char/tpm/tpm_infineon.c > @@ -22,10 +22,10 @@ > /* Infineon specific definitions */ > /* maximum number of WTX-packages */ > #define TPM_MAX_WTX_PACKAGES 50 > -/* msleep-Time for WTX-packages */ > -#define TPM_WTX_MSLEEP_TIME 20 > -/* msleep-Time --> Interval to check status register */ > -#define TPM_MSLEEP_TIME 3 > +/* usleep-Time for WTX-packages */ > +#define TPM_WTX_USLEEP_TIME (20 * 1000) > +/* usleep-Time --> Interval to check status register */ > +#define TPM_USLEEP_TIME (3 * 1000) > /* gives number of max. msleep()-calls before throwing timeout */ > #define TPM_MAX_TRIES 5000 > #define TPM_INFINEON_DEV_VEN_VALUE 0x15D1 > @@ -191,7 +191,8 @@ static int wait(struct tpm_chip *chip, int wait_for_bit) > /* check the status-register if wait_for_bit is set */ > if (status & 1 << wait_for_bit) > break; > - msleep(TPM_MSLEEP_TIME); > + usleep_range(TPM_USLEEP_TIME, > + TPM_USLEEP_TIME + TPM_TIMEOUT_RANGE_US); > } > if (i == TPM_MAX_TRIES) { /* timeout occurs */ > if (wait_for_bit == STAT_XFE) > @@ -226,7 +227,8 @@ static void tpm_wtx(struct tpm_chip *chip) > wait_and_send(chip, TPM_CTRL_WTX); > wait_and_send(chip, 0x00); > wait_and_send(chip, 0x00); > - msleep(TPM_WTX_MSLEEP_TIME); > + usleep_range(TPM_WTX_USLEEP_TIME, > + TPM_WTX_USLEEP_TIME + TPM_TIMEOUT_RANGE_US); > } > > static void tpm_wtx_abort(struct tpm_chip *chip) > @@ -237,7 +239,8 @@ static void tpm_wtx_abort(struct tpm_chip *chip) > wait_and_send(chip, 0x00); > wait_and_send(chip, 0x00); > number_of_wtx = 0; > - msleep(TPM_WTX_MSLEEP_TIME); > + usleep_range(TPM_WTX_USLEEP_TIME, > + TPM_WTX_USLEEP_TIME + TPM_TIMEOUT_RANGE_US); > } > > static int tpm_inf_recv(struct tpm_chip *chip, u8 * buf, size_t count) > diff --git a/drivers/char/tpm/tpm_tis_core.c b/drivers/char/tpm/tpm_tis_core.c > index c0f296b..93e3233 100644 > --- a/drivers/char/tpm/tpm_tis_core.c > +++ b/drivers/char/tpm/tpm_tis_core.c > @@ -51,7 +51,8 @@ static int wait_startup(struct tpm_chip *chip, int l) > > if (access & TPM_ACCESS_VALID) > return 0; > - msleep(TPM_TIMEOUT); > + usleep_range(TPM_TIMEOUT_US, > + TPM_TIMEOUT_US + TPM_TIMEOUT_RANGE_US); > } while (time_before(jiffies, stop)); > return -1; > } > @@ -125,7 +126,8 @@ again: > do { > if (check_locality(chip, l) >= 0) > return l; > - msleep(TPM_TIMEOUT); > + usleep_range(TPM_TIMEOUT_US, > + TPM_TIMEOUT_US + TPM_TIMEOUT_RANGE_US); > } while (time_before(jiffies, stop)); > } > return -1; > @@ -170,7 +172,8 @@ static int get_burstcount(struct tpm_chip *chip) > burstcnt = (value >> 8) & 0xFFFF; > if (burstcnt) > return burstcnt; > - msleep(TPM_TIMEOUT); > + usleep_range(TPM_TIMEOUT_US, > + TPM_TIMEOUT_US + TPM_TIMEOUT_RANGE_US); > } while (time_before(jiffies, stop)); > return -EBUSY; > } > @@ -408,7 +411,7 @@ static int tpm_tis_send(struct tpm_chip *chip, u8 *buf, size_t len) > priv->irq = irq; > chip->flags |= TPM_CHIP_FLAG_IRQ; > if (!priv->irq_tested) > - msleep(1); > + usleep_range(1000, 1000 + TPM_TIMEOUT_RANGE_US); > if (!priv->irq_tested) > disable_interrupts(chip); > priv->irq_tested = true; > -- > 2.7.4 > PS. Please add linux-kernel and linux-security-module to CCs for the next version and thanks for working on this improvement. /Jarkko ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot