Skip to contentSkip to conversation

remaining tpmdd fixes for Linux 4.5

GIT PULL Jarkko Sakkinen 4 participants 12 messages Feb 20, 2016 – Feb 26, 2016

lore
Share & tools
lore
#0 lore raw

Hi James,

I'm sorry for the late pull request for 4.5. The reason for this was the latency in my previous one. I picked with care the absolutely critical fixes so that we can make a sound tpmdd release.

I really hope you can still pick these as one of them is absolutely critical to get authorization policy sealing API right (kernel keeps it finger out of user space created objects).

BR,
/Jarkko

The following changes since commit 8e0ee3c9faed7ca68807ea45141775856c438ac0:

tpm: fix the cleanup of struct tpm_chip (2016-02-10 04:12:08 +0200)

are available in the git repository at:

https://github.com/jsakkine/linux-tpmdd.git tags/tpmdd-next-20160220

for you to fetch changes up to 99cda8cb4639de81cde785b5bab9bc52e916e594:

tpm_crb: tpm2_shutdown() must be called before tpm_chip_unregister() (2016-02-20 09:59:33 +0200)

----------------------------------------------------------------
remaining tpmdd fixes for Linux 4.5

----------------------------------------------------------------
Harald Hoyer (1):
tpm_eventlog.c: fix binary_bios_measurements

Jarkko Sakkinen (4):
tpm: fix: set continueSession attribute for the unseal operation
tpm: fix: return rc when devm_add_action() fails tpm_crb/tis: fix: use dev_name() for /proc/iomem
tpm_crb: tpm2_shutdown() must be called before tpm_chip_unregister()

 drivers/char/tpm/tpm-chip.c     |  7 ++++++-
 drivers/char/tpm/tpm2-cmd.c     | 10 +++++++---
 drivers/char/tpm/tpm_crb.c      |  8 +++++---
 drivers/char/tpm/tpm_eventlog.c | 14 ++++++++++----
 drivers/char/tpm/tpm_tis.c      |  4 +++-
 5 files changed, 31 insertions(+), 12 deletions(-)
#1 lore raw
Quoted 9 lines · Jarkko Sakkinen

Hi James,

I'm sorry for the late pull request for 4.5. The reason for this was the latency in my previous one. I picked with care the absolutely critical fixes so that we can make a sound tpmdd release.

I really hope you can still pick these as one of them is absolutely critical to get authorization policy sealing API right (kernel keeps it finger out of user space created objects).

Pushed to next for more testing and review.

This really is getting too late in the development cycle for so many fixes. It means the code was not ready to be merged in the first place.

Also, any idea why I'm seeing this:

drivers/char/tpm/tpm_tis.c:838: warning: ‘tpm_tis_resume’ defined but not used

Signature · 2 lines
James Morris
<jmorris-gx6/[email protected]>

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140

_______________________________________________
tpmdd-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/tpmdd-devel

#2 lore raw
Quoted 8 lines · James Morris
Quoted 6 lines · Jarkko Sakkinen

Hi James,

I'm sorry for the late pull request for 4.5. The reason for this was the latency in my previous one. I picked with care the absolutely critical fixes so that we can make a sound tpmdd release.

Quoted 8 lines
Quoted 3 lines

I really hope you can still pick these as one of them is absolutely critical to get authorization policy sealing API right (kernel keeps it finger out of user space created objects).

Pushed to next for more testing and review.

This really is getting too late in the development cycle for so many fixes. It means the code was not ready to be merged in the first place.

I fully agree what you're saying. I'll learn the lesson here and take factors more conservative attitude from now on. No excuses. I'm sorry about this.

Partly the reason for recent increase in regressions has been increased real-world use of TPM2 and thus issues have started to pop up that's a lame excuse anyway.

Quoted 4 lines

Also, any idea why I'm seeing this:

drivers/char/tpm/tpm_tis.c:838: warning: ‘tpm_tis_resume’ defined but not used

Bisected the patch: 00194826e6be

Do you want me to send a pull request containing a fix for the build warning or reverting the whole commit? My call would be to apply the fix because this commit has been tested both TPM 1.2 by Martin and with TPM 2.0 by me and things have worked well.

I can live with either option.

I already pushed a fix to my master for this issue:

https://github.com/jsakkine/linux-tpmdd/commit/6386544ad7bceb3d0248b85da29d4d99eebe9161

Quoted 3 lines
Signature · 2 lines
James Morris
<[email protected]>

I've been recently working on a custom BR environment that bundles my latest master with initramfs user space [1]. At minimum I'll start using this environment to create builds of this env with and without PM for release testing and run the images both 1.2 and 2.0 HW.

This should prevent the warning you experienced never happening again.

[1] http://git.infradead.org/users/jjs/buildroot-tpmdd.git

/Jarkko

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
_______________________________________________
tpmdd-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/tpmdd-devel

#3 lore raw
Quoted 3 lines · Jarkko Sakkinen

I already pushed a fix to my master for this issue:

https://github.com/jsakkine/linux-tpmdd/commit/6386544ad7bceb3d0248b85da29d4d99eebe9161

The goal is to reduce the number of #ifdef'd code segments so we have fewer problems in future with a large .config test matrix.

I'd rather see a __maybe_unused annotation instead.

Jason

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140

#4 lore raw
Quoted 10 lines · Jason Gunthorpe
Quoted 3 lines · Jarkko Sakkinen

I already pushed a fix to my master for this issue:

https://github.com/jsakkine/linux-tpmdd/commit/6386544ad7bceb3d0248b85da29d4d99eebe9161

The goal is to reduce the number of #ifdef'd code segments so we have fewer problems in future with a large .config test matrix.

I'd rather see a __maybe_unused annotation instead.

Agreed that it's a better form but at this point it's probably revert the breaking change and move to that later on. Otherwise, I don't see reason not to include the patch that you authored to the release. I've used it in my test kernels for quite some time now and it has worked without issues.

I sent my fix for review now.

Quoted 1 line

Jason

/Jarkko

#5 lore raw
Quoted 12 lines · Jarkko Sakkinen
Quoted 10 lines · Jason Gunthorpe
Quoted 3 lines · Jarkko Sakkinen

I already pushed a fix to my master for this issue:

https://github.com/jsakkine/linux-tpmdd/commit/6386544ad7bceb3d0248b85da29d4d99eebe9161

The goal is to reduce the number of #ifdef'd code segments so we have fewer problems in future with a large .config test matrix.

I'd rather see a __maybe_unused annotation instead.

Quoted 7 lines

Agreed that it's a better form but at this point it's probably revert the breaking change and move to that later on. Otherwise, I don't see reason not to include the patch that you authored to the release. I've used it in my test kernels for quite some time now and it has worked without issues.

I sent my fix for review now.

A warning with some kconfigs is very minor, we can take the time to fix it properly for 4.6.

I am surprised the 0day -next builds/etc didn't notice this - Jarkko is your tree included in that process somehow? (sorry, I don't remember the details)

Jason

#6 lore raw
Quoted 19 lines · Jason Gunthorpe
Quoted 18 lines · Jarkko Sakkinen
Quoted 10 lines · Jason Gunthorpe
Quoted 3 lines · Jarkko Sakkinen

I already pushed a fix to my master for this issue:

https://github.com/jsakkine/linux-tpmdd/commit/6386544ad7bceb3d0248b85da29d4d99eebe9161

The goal is to reduce the number of #ifdef'd code segments so we have fewer problems in future with a large .config test matrix.

I'd rather see a __maybe_unused annotation instead.

Agreed that it's a better form but at this point it's probably revert the breaking change and move to that later on. Otherwise, I don't see reason not to include the patch that you authored to the release. I've used it in my test kernels for quite some time now and it has worked without issues.

Quoted 4 lines
Quoted 1 line

I sent my fix for review now.

A warning with some kconfigs is very minor, we can take the time to fix it properly for 4.6.

I don't see any problem adding those too ifdefs back for 4.5 release.

Quoted 3 lines

I am surprised the 0day -next builds/etc didn't notice this - Jarkko is your tree included in that process somehow? (sorry, I don't remember the details)

It has been and I've received notifications from there from time to time about my master branch and fixed issues accordingly.

Quoted 1 line

Jason

/Jarkko

#7 lore raw
In reply to Jarkko Sakkinen
Quoted 18 lines · Jarkko Sakkinen
Quoted 8 lines · James Morris
Quoted 6 lines · Jarkko Sakkinen

Hi James,

I'm sorry for the late pull request for 4.5. The reason for this was the latency in my previous one. I picked with care the absolutely critical fixes so that we can make a sound tpmdd release.

Quoted 8 lines
Quoted 3 lines

I really hope you can still pick these as one of them is absolutely critical to get authorization policy sealing API right (kernel keeps it finger out of user space created objects).

Pushed to next for more testing and review.

This really is getting too late in the development cycle for so many fixes. It means the code was not ready to be merged in the first place.

Quoted 8 lines

I fully agree what you're saying. I'll learn the lesson here and take factors more conservative attitude from now on. No excuses. I'm sorry about this.

Partly the reason for recent increase in regressions has been increased real-world use of TPM2 and thus issues have started to pop up that's a lame excuse anyway.

Would it be worthwhile to have a tpm branch that gets pulled by -next directly so changes will have already been going through the paces in -next prior to the pull reuqest to James?

snits

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140

#8 lore raw
Quoted 23 lines · Jerry Snitselaar
Quoted 22 lines · Jarkko Sakkinen
Quoted 8 lines · James Morris
Quoted 6 lines · Jarkko Sakkinen

Hi James,

I'm sorry for the late pull request for 4.5. The reason for this was the latency in my previous one. I picked with care the absolutely critical fixes so that we can make a sound tpmdd release.

Quoted 8 lines
Quoted 3 lines

I really hope you can still pick these as one of them is absolutely critical to get authorization policy sealing API right (kernel keeps it finger out of user space created objects).

Pushed to next for more testing and review.

This really is getting too late in the development cycle for so many fixes. It means the code was not ready to be merged in the first place.

I fully agree what you're saying. I'll learn the lesson here and take factors more conservative attitude from now on. No excuses. I'm sorry about this.

Quoted 8 lines
Quoted 4 lines

Partly the reason for recent increase in regressions has been increased real-world use of TPM2 and thus issues have started to pop up that's a lame excuse anyway.

Would it be worthwhile to have a tpm branch that gets pulled by -next directly so changes will have already been going through the paces in -next prior to the pull reuqest to James?

That would be useful, some other subsystems do that.

Signature · 10 lines
James Morris
<jmorris-gx6/[email protected]>


------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
#10 lore raw
In reply to Jarkko Sakkinen
Quoted 5 lines · Jarkko Sakkinen

Do you want me to send a pull request containing a fix for the build warning or reverting the whole commit? My call would be to apply the fix because this commit has been tested both TPM 1.2 by Martin and with TPM 2.0 by me and things have worked well.

Send me a pull request just for the fix.

I won't be pushing these changes to Linus for 4.5, they'll have to wait until the 4.6.

Signature · 10 lines
James Morris
<jmorris-gx6/[email protected]>


------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
#11 lore raw
Quoted 8 lines · James Morris
Quoted 5 lines · Jarkko Sakkinen

Do you want me to send a pull request containing a fix for the build warning or reverting the whole commit? My call would be to apply the fix because this commit has been tested both TPM 1.2 by Martin and with TPM 2.0 by me and things have worked well.

Quoted 4 lines

Send me a pull request just for the fix.

I won't be pushing these changes to Linus for 4.5, they'll have to wait until the 4.6.

OK

Quoted 3 lines
Signature · 2 lines
James Morris
<jmorris-gx6/[email protected]>

/Jarkko

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140

#9 lore raw
In reply to James Morris
Quoted 30 lines · James Morris
Quoted 28 lines · Jerry Snitselaar
Quoted 26 lines · Jarkko Sakkinen
Quoted 8 lines · James Morris
Quoted 6 lines · Jarkko Sakkinen

Hi James,

I'm sorry for the late pull request for 4.5. The reason for this was the latency in my previous one. I picked with care the absolutely critical fixes so that we can make a sound tpmdd release.

Quoted 8 lines
Quoted 3 lines

I really hope you can still pick these as one of them is absolutely critical to get authorization policy sealing API right (kernel keeps it finger out of user space created objects).

Pushed to next for more testing and review.

This really is getting too late in the development cycle for so many fixes. It means the code was not ready to be merged in the first place.

I fully agree what you're saying. I'll learn the lesson here and take factors more conservative attitude from now on. No excuses. I'm sorry about this.

Partly the reason for recent increase in regressions has been increased real-world use of TPM2 and thus issues have started to pop up that's a lame excuse anyway.

Quoted 5 lines
Quoted 3 lines

Would it be worthwhile to have a tpm branch that gets pulled by -next directly so changes will have already been going through the paces in -next prior to the pull reuqest to James?

That would be useful, some other subsystems do that.

I'll start using it.

Quoted 3 lines
Signature · 2 lines
James Morris
<jmorris-gx6/[email protected]>

/Jarkko

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140