| Message ID | 20161203175213.10210-4-jarkko.sakkinen@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
CC linux-security-module On Sat, Dec 03, 2016 at 07:52:13PM +0200, Jarkko Sakkinen wrote: > Request and relinquish locality for the driver use in order to be > a better citizen in a multi locality environment like TXT. The > locality is requested and relinquished as part of going into and > waking up from idle. > > Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> > --- > drivers/char/tpm/tpm_crb.c | 36 ++++++++++++++++++++++++------------ > 1 file changed, 24 insertions(+), 12 deletions(-) > > diff --git a/drivers/char/tpm/tpm_crb.c b/drivers/char/tpm/tpm_crb.c > index f986d02..f6e0beb 100644 > --- a/drivers/char/tpm/tpm_crb.c > +++ b/drivers/char/tpm/tpm_crb.c > @@ -34,6 +34,15 @@ enum crb_defaults { > CRB_ACPI_START_INDEX = 1, > }; > > +enum crb_loc_ctrl { > + CRB_LOC_CTRL_REQUEST_ACCESS = BIT(0), > + CRB_LOC_CTRL_RELINQUISH = BIT(1), > +}; > + > +enum crb_loc_state { > + CRB_LOC_STATE_LOC_ASSIGNED = BIT(1), > +}; > + > enum crb_ctrl_req { > CRB_CTRL_REQ_CMD_READY = BIT(0), > CRB_CTRL_REQ_GO_IDLE = BIT(1), > @@ -101,12 +110,8 @@ struct crb_priv { > * @dev: crb device > * @priv: crb private data > * > - * Write CRB_CTRL_REQ_GO_IDLE to TPM_CRB_CTRL_REQ > - * The device should respond within TIMEOUT_C by clearing the bit. > - * Anyhow, we do not wait here as a consequent CMD_READY request > - * will be handled correctly even if idle was not completed. > - * > - * The function does nothing for devices with ACPI-start method. > + * Put device to the idle state and relinquish locality. The function does > + * nothing for devices with the ACPI-start method. > * > * Return: 0 always > */ > @@ -115,6 +120,7 @@ static int __maybe_unused crb_go_idle(struct device *dev, struct crb_priv *priv) > if (priv->flags & CRB_FL_ACPI_START) > return 0; > > + iowrite32(CRB_LOC_CTRL_RELINQUISH, &priv->regs_h->loc_ctrl); > iowrite32(CRB_CTRL_REQ_GO_IDLE, &priv->regs_t->ctrl_req); > /* we don't really care when this settles */ > > @@ -146,11 +152,8 @@ static bool crb_wait_for_reg_32(u32 __iomem *reg, u32 mask, u32 value, > * @dev: crb device > * @priv: crb private data > * > - * Write CRB_CTRL_REQ_CMD_READY to TPM_CRB_CTRL_REQ > - * and poll till the device acknowledge it by clearing the bit. > - * The device should respond within TIMEOUT_C. > - * > - * The function does nothing for devices with ACPI-start method > + * Try to wake up the device and request locality. The function does nothing > + * for devices with the ACPI-start method. > * > * Return: 0 on success -ETIME on timeout; > */ > @@ -165,7 +168,16 @@ static int __maybe_unused crb_cmd_ready(struct device *dev, > CRB_CTRL_REQ_CMD_READY /* mask */, > 0, /* value */ > TPM2_TIMEOUT_C)) { > - dev_warn(dev, "cmdReady timed out\n"); > + dev_warn(dev, "TPM_CRB_CTRL_REQ_x.cmdReady timed out\n"); > + return -ETIME; > + } > + > + iowrite32(CRB_LOC_CTRL_REQUEST_ACCESS, &priv->regs_h->loc_ctrl); > + if (!crb_wait_for_reg_32(&priv->regs_h->loc_state, > + CRB_LOC_STATE_LOC_ASSIGNED, /* mask */ > + CRB_LOC_STATE_LOC_ASSIGNED, /* value */ > + TPM2_TIMEOUT_C)) { > + dev_warn(dev, "TPM_LOC_STATE_x.requestAccess timed out\n"); > return -ETIME; > } > > -- > 2.9.3 > ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot
On Sat, Dec 03, 2016 at 07:52:13PM +0200, Jarkko Sakkinen wrote:
> + iowrite32(CRB_LOC_CTRL_RELINQUISH, &priv->regs_h->loc_ctrl);
Since regs_h can be null shouldn't there be some guards ?
Jason
------------------------------------------------------------------------------
> > CC linux-security-module > > On Sat, Dec 03, 2016 at 07:52:13PM +0200, Jarkko Sakkinen wrote: > > Request and relinquish locality for the driver use in order to be a > > better citizen in a multi locality environment like TXT. The locality > > is requested and relinquished as part of going into and waking up from > > idle. > This has nothing to do with power management, please note, that this is run via runtime_pm handlers, which can be disabled via sysfs and actually on server platforms runtime pm is not relevant, while you cannot disable locality acquire/relinquish flow. NACK > > Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> > > --- > > drivers/char/tpm/tpm_crb.c | 36 ++++++++++++++++++++++++------------ > > 1 file changed, 24 insertions(+), 12 deletions(-) > > > > diff --git a/drivers/char/tpm/tpm_crb.c b/drivers/char/tpm/tpm_crb.c > > index f986d02..f6e0beb 100644 > > --- a/drivers/char/tpm/tpm_crb.c > > +++ b/drivers/char/tpm/tpm_crb.c > > @@ -34,6 +34,15 @@ enum crb_defaults { > > CRB_ACPI_START_INDEX = 1, > > }; > > > > +enum crb_loc_ctrl { > > + CRB_LOC_CTRL_REQUEST_ACCESS = BIT(0), > > + CRB_LOC_CTRL_RELINQUISH = BIT(1), > > +}; > > + > > +enum crb_loc_state { > > + CRB_LOC_STATE_LOC_ASSIGNED = BIT(1), > > +}; > > + > > enum crb_ctrl_req { > > CRB_CTRL_REQ_CMD_READY = BIT(0), > > CRB_CTRL_REQ_GO_IDLE = BIT(1), > > @@ -101,12 +110,8 @@ struct crb_priv { > > * @dev: crb device > > * @priv: crb private data > > * > > - * Write CRB_CTRL_REQ_GO_IDLE to TPM_CRB_CTRL_REQ > > - * The device should respond within TIMEOUT_C by clearing the bit. > > - * Anyhow, we do not wait here as a consequent CMD_READY request > > - * will be handled correctly even if idle was not completed. > > - * > > - * The function does nothing for devices with ACPI-start method. > > + * Put device to the idle state and relinquish locality. The function > > + does > > + * nothing for devices with the ACPI-start method. > > * > > * Return: 0 always > > */ > > @@ -115,6 +120,7 @@ static int __maybe_unused crb_go_idle(struct device > *dev, struct crb_priv *priv) > > if (priv->flags & CRB_FL_ACPI_START) > > return 0; > > > > + iowrite32(CRB_LOC_CTRL_RELINQUISH, &priv->regs_h->loc_ctrl); This is misleading this has nothing to do with go idle, what is the > > iowrite32(CRB_CTRL_REQ_GO_IDLE, &priv->regs_t->ctrl_req); > > /* we don't really care when this settles */ > > > > @@ -146,11 +152,8 @@ static bool crb_wait_for_reg_32(u32 __iomem *reg, > u32 mask, u32 value, > > * @dev: crb device > > * @priv: crb private data > > * > > - * Write CRB_CTRL_REQ_CMD_READY to TPM_CRB_CTRL_REQ > > - * and poll till the device acknowledge it by clearing the bit. > > - * The device should respond within TIMEOUT_C. > > - * > > - * The function does nothing for devices with ACPI-start method > > + * Try to wake up the device and request locality. The function does > > + nothing > > + * for devices with the ACPI-start method. > > * > > * Return: 0 on success -ETIME on timeout; > > */ > > @@ -165,7 +168,16 @@ static int __maybe_unused crb_cmd_ready(struct > device *dev, > > CRB_CTRL_REQ_CMD_READY /* mask */, > > 0, /* value */ > > TPM2_TIMEOUT_C)) { > > - dev_warn(dev, "cmdReady timed out\n"); > > + dev_warn(dev, "TPM_CRB_CTRL_REQ_x.cmdReady timed > out\n"); We are always in locality 0 here, right? > > + return -ETIME; > > + } > > + > > + iowrite32(CRB_LOC_CTRL_REQUEST_ACCESS, &priv->regs_h- > >loc_ctrl); > > + if (!crb_wait_for_reg_32(&priv->regs_h->loc_state, > > + CRB_LOC_STATE_LOC_ASSIGNED, /* mask */ > > + CRB_LOC_STATE_LOC_ASSIGNED, /* value */ > > + TPM2_TIMEOUT_C)) { > > + dev_warn(dev, "TPM_LOC_STATE_x.requestAccess timed > out\n"); > > return -ETIME; > > } ------------------------------------------------------------------------------ Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today.http://sdm.link/xeonphi
On Mon, Dec 05, 2016 at 09:32:35AM -0700, Jason Gunthorpe wrote: > On Sat, Dec 03, 2016 at 07:52:13PM +0200, Jarkko Sakkinen wrote: > > + iowrite32(CRB_LOC_CTRL_RELINQUISH, &priv->regs_h->loc_ctrl); > > Since regs_h can be null shouldn't there be some guards ? Yes, there should. Thanks for pointing this out! /Jarkko ------------------------------------------------------------------------------ Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today.http://sdm.link/xeonphi
On Mon, Dec 05, 2016 at 09:51:42PM +0000, Winkler, Tomas wrote: > > > > CC linux-security-module > > > > On Sat, Dec 03, 2016 at 07:52:13PM +0200, Jarkko Sakkinen wrote: > > > Request and relinquish locality for the driver use in order to be a > > > better citizen in a multi locality environment like TXT. The locality > > > is requested and relinquished as part of going into and waking up from > > > idle. > > > This has nothing to do with power management, please note, that this is run via runtime_pm handlers, > which can be disabled via sysfs and actually on server platforms runtime pm is not relevant, > while you cannot disable locality acquire/relinquish flow. > > NACK Good point. I'll move it as part of tpm_transmit() /Jarkko > > > Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> > > > --- > > > drivers/char/tpm/tpm_crb.c | 36 ++++++++++++++++++++++++------------ > > > 1 file changed, 24 insertions(+), 12 deletions(-) > > > > > > diff --git a/drivers/char/tpm/tpm_crb.c b/drivers/char/tpm/tpm_crb.c > > > index f986d02..f6e0beb 100644 > > > --- a/drivers/char/tpm/tpm_crb.c > > > +++ b/drivers/char/tpm/tpm_crb.c > > > @@ -34,6 +34,15 @@ enum crb_defaults { > > > CRB_ACPI_START_INDEX = 1, > > > }; > > > > > > +enum crb_loc_ctrl { > > > + CRB_LOC_CTRL_REQUEST_ACCESS = BIT(0), > > > + CRB_LOC_CTRL_RELINQUISH = BIT(1), > > > +}; > > > + > > > +enum crb_loc_state { > > > + CRB_LOC_STATE_LOC_ASSIGNED = BIT(1), > > > +}; > > > + > > > enum crb_ctrl_req { > > > CRB_CTRL_REQ_CMD_READY = BIT(0), > > > CRB_CTRL_REQ_GO_IDLE = BIT(1), > > > @@ -101,12 +110,8 @@ struct crb_priv { > > > * @dev: crb device > > > * @priv: crb private data > > > * > > > - * Write CRB_CTRL_REQ_GO_IDLE to TPM_CRB_CTRL_REQ > > > - * The device should respond within TIMEOUT_C by clearing the bit. > > > - * Anyhow, we do not wait here as a consequent CMD_READY request > > > - * will be handled correctly even if idle was not completed. > > > - * > > > - * The function does nothing for devices with ACPI-start method. > > > + * Put device to the idle state and relinquish locality. The function > > > + does > > > + * nothing for devices with the ACPI-start method. > > > * > > > * Return: 0 always > > > */ > > > @@ -115,6 +120,7 @@ static int __maybe_unused crb_go_idle(struct device > > *dev, struct crb_priv *priv) > > > if (priv->flags & CRB_FL_ACPI_START) > > > return 0; > > > > > > + iowrite32(CRB_LOC_CTRL_RELINQUISH, &priv->regs_h->loc_ctrl); > This is misleading this has nothing to do with go idle, what is the > > > iowrite32(CRB_CTRL_REQ_GO_IDLE, &priv->regs_t->ctrl_req); > > > /* we don't really care when this settles */ > > > > > > @@ -146,11 +152,8 @@ static bool crb_wait_for_reg_32(u32 __iomem *reg, > > u32 mask, u32 value, > > > * @dev: crb device > > > * @priv: crb private data > > > * > > > - * Write CRB_CTRL_REQ_CMD_READY to TPM_CRB_CTRL_REQ > > > - * and poll till the device acknowledge it by clearing the bit. > > > - * The device should respond within TIMEOUT_C. > > > - * > > > - * The function does nothing for devices with ACPI-start method > > > + * Try to wake up the device and request locality. The function does > > > + nothing > > > + * for devices with the ACPI-start method. > > > * > > > * Return: 0 on success -ETIME on timeout; > > > */ > > > @@ -165,7 +168,16 @@ static int __maybe_unused crb_cmd_ready(struct > > device *dev, > > > CRB_CTRL_REQ_CMD_READY /* mask */, > > > 0, /* value */ > > > TPM2_TIMEOUT_C)) { > > > - dev_warn(dev, "cmdReady timed out\n"); > > > + dev_warn(dev, "TPM_CRB_CTRL_REQ_x.cmdReady timed > > out\n"); > We are always in locality 0 here, right? > > > > + return -ETIME; > > > + } > > > + > > > + iowrite32(CRB_LOC_CTRL_REQUEST_ACCESS, &priv->regs_h- > > >loc_ctrl); > > > + if (!crb_wait_for_reg_32(&priv->regs_h->loc_state, > > > + CRB_LOC_STATE_LOC_ASSIGNED, /* mask */ > > > + CRB_LOC_STATE_LOC_ASSIGNED, /* value */ > > > + TPM2_TIMEOUT_C)) { > > > + dev_warn(dev, "TPM_LOC_STATE_x.requestAccess timed > > out\n"); > > > > > return -ETIME; > > > } > > ------------------------------------------------------------------------------ Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today.http://sdm.link/xeonphi
diff --git a/drivers/char/tpm/tpm_crb.c b/drivers/char/tpm/tpm_crb.c index f986d02..f6e0beb 100644 --- a/drivers/char/tpm/tpm_crb.c +++ b/drivers/char/tpm/tpm_crb.c @@ -34,6 +34,15 @@ enum crb_defaults { CRB_ACPI_START_INDEX = 1, }; +enum crb_loc_ctrl { + CRB_LOC_CTRL_REQUEST_ACCESS = BIT(0), + CRB_LOC_CTRL_RELINQUISH = BIT(1), +}; + +enum crb_loc_state { + CRB_LOC_STATE_LOC_ASSIGNED = BIT(1), +}; + enum crb_ctrl_req { CRB_CTRL_REQ_CMD_READY = BIT(0), CRB_CTRL_REQ_GO_IDLE = BIT(1), @@ -101,12 +110,8 @@ struct crb_priv { * @dev: crb device * @priv: crb private data * - * Write CRB_CTRL_REQ_GO_IDLE to TPM_CRB_CTRL_REQ - * The device should respond within TIMEOUT_C by clearing the bit. - * Anyhow, we do not wait here as a consequent CMD_READY request - * will be handled correctly even if idle was not completed. - * - * The function does nothing for devices with ACPI-start method. + * Put device to the idle state and relinquish locality. The function does + * nothing for devices with the ACPI-start method. * * Return: 0 always */ @@ -115,6 +120,7 @@ static int __maybe_unused crb_go_idle(struct device *dev, struct crb_priv *priv) if (priv->flags & CRB_FL_ACPI_START) return 0; + iowrite32(CRB_LOC_CTRL_RELINQUISH, &priv->regs_h->loc_ctrl); iowrite32(CRB_CTRL_REQ_GO_IDLE, &priv->regs_t->ctrl_req); /* we don't really care when this settles */ @@ -146,11 +152,8 @@ static bool crb_wait_for_reg_32(u32 __iomem *reg, u32 mask, u32 value, * @dev: crb device * @priv: crb private data * - * Write CRB_CTRL_REQ_CMD_READY to TPM_CRB_CTRL_REQ - * and poll till the device acknowledge it by clearing the bit. - * The device should respond within TIMEOUT_C. - * - * The function does nothing for devices with ACPI-start method + * Try to wake up the device and request locality. The function does nothing + * for devices with the ACPI-start method. * * Return: 0 on success -ETIME on timeout; */ @@ -165,7 +168,16 @@ static int __maybe_unused crb_cmd_ready(struct device *dev, CRB_CTRL_REQ_CMD_READY /* mask */, 0, /* value */ TPM2_TIMEOUT_C)) { - dev_warn(dev, "cmdReady timed out\n"); + dev_warn(dev, "TPM_CRB_CTRL_REQ_x.cmdReady timed out\n"); + return -ETIME; + } + + iowrite32(CRB_LOC_CTRL_REQUEST_ACCESS, &priv->regs_h->loc_ctrl); + if (!crb_wait_for_reg_32(&priv->regs_h->loc_state, + CRB_LOC_STATE_LOC_ASSIGNED, /* mask */ + CRB_LOC_STATE_LOC_ASSIGNED, /* value */ + TPM2_TIMEOUT_C)) { + dev_warn(dev, "TPM_LOC_STATE_x.requestAccess timed out\n"); return -ETIME; }
Request and relinquish locality for the driver use in order to be a better citizen in a multi locality environment like TXT. The locality is requested and relinquished as part of going into and waking up from idle. Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> --- drivers/char/tpm/tpm_crb.c | 36 ++++++++++++++++++++++++------------ 1 file changed, 24 insertions(+), 12 deletions(-)