char/tpm: tpm_infineon no longer loaded for HP 2510p laptop
While comparing the loaded modules for 2.6.26 and 2.6.27-rc3 for my HP 2510p, I noticed that the tpm_infineon module and related modules no longer get loaded automatically.
The difference seems to be that 2.6.26 listed:
/lib/modules/2.6.26.2/modules.alias:alias pnp:dIFX0102* tpm_infineon /lib/modules/2.6.26.2/modules.alias:alias pnp:dIFX0101* tpm_infineon
while 2.6.27 has:
/lib/modules/2.6.27-rc3/modules.alias:alias acpi*:IFX0101:* tpm_infineon /lib/modules/2.6.27-rc3/modules.alias:alias pnp:dIFX0101* tpm_infineon
My system has:
$ grep IFX /sys/bus/pnp/devices/*/id /sys/bus/pnp/devices/00:02/id:IFX0102
And if I modprobe the module manually the device really does seem to be supported:
pnp: the driver 'tpm_inf_pnp' has been registered tpm_inf_pnp 00:02: Found C284 with ID IFX0102tpm_inf_pnp 00:02: TPM found: config base 0x560, data base 0x570, chip version 0x000b, vendor id 0x15d1 (Infineon), product id 0x000b (SLB 9635 TT 1.2)tpm_inf_pnp 00:02: driver attached
Relevant bit of the kernel config:CONFIG_TCG_TPM=mCONFIG_TCG_TIS=m
# CONFIG_TCG_NSC is not set
# CONFIG_TCG_ATMEL is not setCONFIG_TCG_INFINEON=m
I've taken a quick look at changes from 2.6.26 in drivers/char/tpm/ but did not see really anything there that could explain this difference in behavior.
Cheers,
FJP
Quoted 8 lines · Frans Pop
While comparing the loaded modules for 2.6.26 and 2.6.27-rc3 for my HP 2510p, I noticed that the
tpm_infineonmodule and related modules no longer get loaded automatically.The difference seems to be that 2.6.26 listed:
/lib/modules/2.6.26.2/modules.alias:alias pnp:dIFX0102*tpm_infineon/lib/modules/2.6.26.2/modules.alias:alias pnp:dIFX0101*tpm_infineon
Quoted 7 lines
while 2.6.27 has:
/lib/modules/2.6.27-rc3/modules.alias:alias acpi*:IFX0101:*tpm_infineon/lib/modules/2.6.27-rc3/modules.alias:alias pnp:dIFX0101*tpm_infineonMy system has:
$ grep IFX /sys/bus/pnp/devices/*/id /sys/bus/pnp/devices/00:02/id:IFX0102
drivers/char/tpm/tpm_infineon.c hasn't changed since v2.6.26. I think the problem is more likely related to commit 22454cb99fc39f2629a:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=22454cb99fc39f2629ad06a7eccb3df312f8830e
I don't know enough about modaliases to understand the point of this patch, but I am suspicious of this part of it:
- do_table(symval, sym->st_size,
- sizeof(struct pnp_device_id), "pnp",
- do_pnp_entry, mod);
+ do_pnp_device_entry(symval, sym->st_size, mod);
That suggests to me that where we used to generate an alias for every PNP ID in the table:
static const struct pnp_device_id tpm_pnp_tbl[] = {
{"IFX0101", 0},
{"IFX0102", 0},
{"", 0}
};possibly the new code only does it for the first entry (IFX0101).
Apart from problem Frans points out, I'd like to understand the reason for generating the "acpi*" aliases for PNP drivers. I'd like to move away from ACPI being visible to userland, and this file2alias.c change makes it *more* visible.
Bjorn
Quoted 13 lines · Bjorn Helgaas
Quoted 12 lines · Frans Pop
While comparing the loaded modules for 2.6.26 and 2.6.27-rc3 for my HP 2510p, I noticed that the
tpm_infineonmodule and related modules no longer get loaded automatically.The difference seems to be that 2.6.26 listed:
/lib/modules/2.6.26.2/modules.alias:alias pnp:dIFX0102*tpm_infineon/lib/modules/2.6.26.2/modules.alias:alias pnp:dIFX0101*tpm_infineonwhile 2.6.27 has:
/lib/modules/2.6.27-rc3/modules.alias:alias acpi*:IFX0101:*tpm_infineon/lib/modules/2.6.27-rc3/modules.alias:alias pnp:dIFX0101*tpm_infineon
Quoted 7 lines
Quoted 3 lines
My system has:
$ grep IFX /sys/bus/pnp/devices/*/id /sys/bus/pnp/devices/00:02/id:IFX0102
drivers/char/tpm/tpm_infineon.chasn't changed since v2.6.26. I think the problem is more likely related to commit22454cb99fc39f2629a:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=22454cb99fc39f2629ad06a7eccb3df312f8830e
Frans, could you verify if reverting commit 22454cb99fc39f2629a fixes the problem for you?
Thanks,
Rafael
Quoted 18 lines · Rafael J. Wysocki
Quoted 17 lines · Bjorn Helgaas
Quoted 15 lines · Frans Pop
While comparing the loaded modules for 2.6.26 and 2.6.27-rc3 for my HP 2510p, I noticed that the
tpm_infineonmodule and related modules no longer get loaded automatically.The difference seems to be that 2.6.26 listed:
/lib/modules/2.6.26.2/modules.alias:alias pnp:dIFX0102*tpm_infineon/lib/modules/2.6.26.2/modules.alias:alias pnp:dIFX0101*tpm_infineonwhile 2.6.27 has:
/lib/modules/2.6.27-rc3/modules.alias:alias acpi*:IFX0101:*tpm_infineon/lib/modules/2.6.27-rc3/modules.alias:alias pnp:dIFX0101*tpm_infineonMy system has:
$ grep IFX /sys/bus/pnp/devices/*/id /sys/bus/pnp/devices/00:02/id:IFX0102
Quoted 6 lines
Quoted 3 lines
drivers/char/tpm/tpm_infineon.chasn't changed since v2.6.26. I think the problem is more likely related to commit22454cb99fc39f2629a:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=22454cb99fc39f2629ad06a7eccb3df312f8830eFrans, could you verify if reverting commit
22454cb99fc39f2629afixes the problem for you?
Yeah, we miss to loop over the list. This should fix it. Thanks!
From: Kay Sievers <[email protected]>
Subject: pnp: fix "add acpi:* modalias entries"
With 22454cb99fc39f2629ad06a7eccb3df312f8830e we added only the first entry of the device table. We need to loop over the whole device list.
---
file2alias.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-)
scripts/mod/file2alias.c +11 -5
@@ -344,14 +344,20 @@ static void do_pnp_device_entry(void *symval, unsigned long size, struct module *mod) {const unsigned long id_size =sizeof (struct pnp_device_id ); -const struct pnp_device_id *id = symval; +const unsigned int count = (size / id_size)-1 ; +const struct pnp_device_id *devs = symval; +unsigned int i;device_id_check (mod->name,"pnp" , size, id_size, symval); -buf_printf (&mod->dev_table_buf, -"MODULE_ALIAS(\"pnp:d%s*\");\n" , id->id); -buf_printf (&mod->dev_table_buf, -"MODULE_ALIAS(\"acpi*:%s:*\");\n" , id->id); +for (i =0 ; i < count; i++) { +const char *id = (char *)devs[i].id; + +buf_printf (&mod->dev_table_buf, +"MODULE_ALIAS(\"pnp:d%s*\");\n" , id); +buf_printf (&mod->dev_table_buf, +"MODULE_ALIAS(\"acpi*:%s:*\");\n" , id); + } }/* looks like: "pnp:dD" for every device of the card */
Quoted 1 line · Kay Sievers
Yeah, we miss to loop over the list. This should fix it. Thanks!
Thanks for the patch.
Can somebody elaborate on why we need to add "acpi*" aliases for all PNP devices? That broadens the kernel/user interface, so I'd like to understand why we need it.
Bjorn
Quoted 8 lines · Bjorn Helgaas
Quoted 1 line · Kay Sievers
Yeah, we miss to loop over the list. This should fix it. Thanks!
Thanks for the patch.
Can somebody elaborate on why we need to add "acpi*" aliases for all PNP devices? That broadens the kernel/user interface, so I'd like to understand why we need it.
We already do ACPI module autoloading by MODALIAS for other things than pnp. ACPI exports the pnp devices with modalias, but the modules do not have a matching alias, this add them.
PNP has no MODALIAS support at all, and the current pnp-aliases would not work for the standard modalias method, they would need to change their format.
The plan is to replace the current pnp modprobe shell script hack in udev, when ACPI devices load the right modules without any special userspace mangling.
Thanks,
Kay
Quoted 4 lines · Kay Sievers
Quoted 3 lines · Bjorn Helgaas
Can somebody elaborate on why we need to add "acpi*" aliases for all PNP devices? That broadens the kernel/user interface, so I'd like to understand why we need it.
Sorry, I should have prefaced my question with "I'm completely ignorant of this modalias stuff." Is there a "complete idiot's guide to modules and udev"? There's precious little in Documentation/ other than a bunch of sample rules for various subsystems.
Quoted 3 lines
We already do ACPI module autoloading by MODALIAS for other things than pnp. ACPI exports the pnp devices with modalias, but the modules do not have a matching alias, this add them.
I'm guessing this has something to do with acpi_device_uevent().
Quoted 3 lines
PNP has no MODALIAS support at all, and the current pnp-aliases would not work for the standard modalias method, they would need to change their format.
pnp_bus_type has no .uevent method. What if I added one? Would that help this situation?
It seems wrong for file2alias.c to take every PNP device (even if it's an ISAPNP or PNPBIOS device) and add "acpi*" aliases for it.
Quoted 3 lines
The plan is to replace the current pnp modprobe shell script hack in udev, when ACPI devices load the right modules without any special userspace mangling.
Is this the shell hack you mean (from etc/udev/rules.d/80-drivers.rules in udev-117)?
SUBSYSTEM=="pnp", DRIVER!="?*", ENV{MODALIAS}!="?*", \
RUN{ignore_error}+="/bin/sh -c '/sbin/modprobe -a $$(while read id; do echo pnp:d$$id; done < /sys$devpath/id)'"I agree that's gross. Could I fix this by implementing pnp_device_uevent()?
Bjorn
Quoted 8 lines · Kay Sievers
From: Kay Sievers <[email protected]>
Subject: pnp: fix "add acpi:* modalias entries"With
22454cb99fc39f2629ad06a7eccb3df312f8830ewe added only the first entry of the device table. We need to loop over the whole device list.Signed-off-by Kay Sievers <[email protected]>
This patch works for me (tested on top of -rc4).
$ grep IFX /lib/modules/2.6.27-rc4/modules.alias alias acpi*:IFX0102:* tpm_infineon alias pnp:dIFX0102* tpm_infineon alias acpi*:IFX0101:* tpm_infineon alias pnp:dIFX0101* tpm_infineon
And the tpm modules are automatically loaded again.
Cheers,
FJP