TPM: Work around buggy TPMs that block during continue self test
PATCH 1/1 TPM: Work around buggy TPMs that block during continue self test
We've been testing an alternative TPM for our embedded products and found random kernel boot failures due to time outs after the continue self test command.
This was happening randomly, and has been *very* hard to track down, but it looks like with this chip there is some kind of race with the tpm_tis_status() check of TPM_STS_COMMAND_READY. If things get there 'too fast' then it sees the chip is ready, or tpm_tis_ready() works. Otherwise it takes somewhere over 400ms before the chip will return TPM_STS_COMMAND_READY.
Adding some delay after tpm_continue_selftest() makes things reliably hit the failure path, otherwise it is a crapshot.
The spec says it should be returning TPM_WARN_DOING_SELFTEST, not holding off on ready..
Boot log during this event looks like this:
tpm_tis 70030000.tpm_tis: 1.2 TPM (device-id 0x3204, rev-id 64)tpm_tis 70030000.tpm_tis: Issuing TPM_STARTUPtpm_tis 70030000.tpm_tis: tpm_transmit: tpm_send: error -62tpm_tis 70030000.tpm_tis: [Hardware Error]: TPM command timed out during continue self test tpm_tis 70030000.tpm_tis: tpm_transmit: tpm_send: error -62tpm_tis 70030000.tpm_tis: [Hardware Error]: TPM command timed out during continue self test tpm_tis 70030000.tpm_tis: tpm_transmit: tpm_send: error -62tpm_tis 70030000.tpm_tis: [Hardware Error]: TPM command timed out during continue self test tpm_tis 70030000.tpm_tis: tpm_transmit: tpm_send: error -62tpm_tis 70030000.tpm_tis: [Hardware Error]: TPM command timed out during continue self test
The other TPM vendor we use doesn't show this wonky behaviour:tpm_tis 70030000.tpm_tis: 1.2 TPM (device-id 0xFE, rev-id 70)
--- drivers/char/tpm/tpm.c | 10 +++++++++- 1 files changed, 9 insertions(+), 1 deletions(-)
drivers/char/tpm/tpm.c +9 -1
@@ -857,7 +857,7 @@ int tpm_do_selftest(struct tpm_chip *chip) {int rc;unsigned int loops; -unsigned int delay_msec =1000 ; +unsigned int delay_msec =100 ;unsigned long duration;struct tpm_cmd_t cmd;
@@ -878,6 +878,14 @@ int tpm_do_selftest(struct tpm_chip *chip) cmd.header.in = pcrread_header; cmd.params.pcrread_in.pcr_idx =cpu_to_be32 (0 ); rc =tpm_transmit (chip, (u8 *) &cmd,READ_PCR_RESULT_SIZE ); +/* Some buggy TPMs will not respond to tpm_tis_ready() for +* around 300ms while the self test is ongoing, keep trying +* until the self test duration expires. */ +if (rc == -ETIME ) { +dev_info (chip->dev,HW_ERR "TPM command timed out during continue self test" ); +msleep (delay_msec); +continue ; + }if (rc <TPM_HEADER_SIZE )return -EFAULT ;
Signature · 1 line
1.7.5.4
Hi Jason,
Quoted 4 lines · Jason Gunthorpe
We've been testing an alternative TPM for our embedded products and found random kernel boot failures due to time outs after the continue self test command.
Quoted 8 lines
This was happening randomly, and has been *very* hard to track down, but it looks like with this chip there is some kind of race with the
tpm_tis_status()check ofTPM_STS_COMMAND_READY. If things get there 'too fast' then it sees the chip is ready, ortpm_tis_ready()works. Otherwise it takes somewhere over 400ms before the chip will returnTPM_STS_COMMAND_READY.Adding some delay after
tpm_continue_selftest()makes things reliably hit the failure path, otherwise it is a crapshot.
I've staged this patch here, please test:
https://github.com/shpedoikal/linux.git tpmdd-01-22-13
Thanks,
Kent
Quoted 14 lines · Kent Yoder
Hi Jason,
Quoted 10 lines · Jason Gunthorpe
We've been testing an alternative TPM for our embedded products and found random kernel boot failures due to time outs after the continue self test command.
This was happening randomly, and has been *very* hard to track down, but it looks like with this chip there is some kind of race with the
tpm_tis_status()check ofTPM_STS_COMMAND_READY. If things get there 'too fast' then it sees the chip is ready, ortpm_tis_ready()works. Otherwise it takes somewhere over 400ms before the chip will returnTPM_STS_COMMAND_READY.
Quoted 6 lines
Quoted 2 lines
Adding some delay after
tpm_continue_selftest()makes things reliably hit the failure path, otherwise it is a crapshot.I've staged this patch here, please test:
https://github.com/shpedoikal/linux.git tpmdd-01-22-13
Thanks Kent, I will try to test your branch next week, if I am able.
Can you also grab
https://github.com/jgunthorpe/linux/commit/98b2a198b43b41b0535200bf475160786398f114
And did you have any comments on:
https://github.com/jgunthorpe/linux/commit/9981e3e622bf702394982117134bed731ffd6f7e
Both were posted to the list a bit ago.
Regards,
Signature · 2 lines
Jason Gunthorpe <[email protected]> (780)4406067x832 Chief Technology Officer, Obsidian Research Corp Edmonton, Canada
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 23 lines · Kent Yoder
Quoted 22 lines · Jason Gunthorpe
Quoted 20 lines · Kent Yoder
Hi Jason,
Quoted 4 lines · Jason Gunthorpe
We've been testing an alternative TPM for our embedded products and found random kernel boot failures due to time outs after the continue self test command.
Quoted 8 lines
This was happening randomly, and has been *very* hard to track down, but it looks like with this chip there is some kind of race with the
tpm_tis_status()check ofTPM_STS_COMMAND_READY. If things get there 'too fast' then it sees the chip is ready, ortpm_tis_ready()works. Otherwise it takes somewhere over 400ms before the chip will returnTPM_STS_COMMAND_READY.Adding some delay after
tpm_continue_selftest()makes things reliably hit the failure path, otherwise it is a crapshot.I've staged this patch here, please test:
https://github.com/shpedoikal/linux.git tpmdd-01-22-13
Quoted 7 lines
Quoted 5 lines
Thanks Kent, I will try to test your branch next week, if I am able.
Can you also grab
https://github.com/jgunthorpe/linux/commit/98b2a198b43b41b0535200bf475160786398f114
Thanks, I missed this, I'll start testing it.
Ok, after yet another round of fixes to the stm i2c driver, I have a
staging tree. I'll submit this as-is next week unless something super ground-breaking comes up.
https://github.com/shpedoikal/linux.git tpmdd-01-31-13
Thanks,
Kent
Quoted 29 lines
Quoted 3 lines
And did you have any comments on:
https://github.com/jgunthorpe/linux/commit/9981e3e622bf702394982117134bed731ffd6f7e
This one is a bit out of date atm, for instance the __dev* stuff is
going away. The thing that most makes me hesitant though is the config if (X86 || OF). A hardware platform or a firmware type.. What platform should this actually target?Kent
Quoted 18 lines
Both were posted to the list a bit ago.
Regards,
--
Jason Gunthorpe <[email protected]> (780)4406067x832
Chief Technology Officer, Obsidian Research Corp Edmonton, Canada------------------------------------------------------------------------------
Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft MVPs and experts. ON SALE this month only -- learn more at:
http://p.sf.net/sfu/learnnow-d2d
_______________________________________________
tpmdd-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/tpmdd-devel
Quoted 6 lines · Kent Yoder
Quoted 6 lines
Quoted 6 lines
Quoted 1 line
https://github.com/shpedoikal/linux.git tpmdd-01-22-13
Thanks Kent, I will try to test your branch next week, if I am able.
Can you also grab
Quoted 7 lines
Quoted 3 lines
Thanks, I missed this, I'll start testing it.
Ok, after yet another round of fixes to the stm i2c driver, I have a
staging tree. I'll submit this as-is next week unless something super ground-breaking comes up.
Great, I'll put this on my list..
I didn't receive your prior message, thank you for quoting it..
Quoted 6 lines
Quoted 6 lines
This one is a bit out of date atm, for instance the __dev* stuff is
going away. The thing that most makes me hesitant though is the config if (X86 || OF). A hardware platform or a firmware type.. What platform should this actually target?
Right on the __dev stuff..
Well the 'depends on' buisness is only due to the unconditional probing of TIS_MEM_BASE for a tpm, and then the autodetection of an IRQ number. That procedure is very much X86 only.
My patch removes the unconditional probe when OF support is turned on, so it is safe to use on any platform only when OF is enabled.
Yes it is really wonky, but unconditionally probing HW in a driver is wonky these days too..
Jason
Am Freitag, 25. Januar 2013, 21:25:38 schrieb Jason Gunthorpe:
Quoted 9 lines
Quoted 1 line
https://github.com/shpedoikal/linux.git tpmdd-01-22-13
Thanks Kent, I will try to test your branch next week, if I am able.
Can you also grab
https://github.com/jgunthorpe/linux/commit/98b2a198b43b41b0535200bf47516078
6398f114
This one looks good to me.
For this one please add my
Thanks,
PeterH