Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH v10, RESEND 4/6] tpm: move tpm_chip definition to
 include/linux/tpm.h
To:     Nathan Chancellor <natechancellor@gmail.com>
CC:     <jarkko.sakkinen@linux.intel.com>, <zohar@linux.ibm.com>,
        <david.safford@ge.com>, <monty.wiseman@ge.com>,
        <matthewgarrett@google.com>, <linux-integrity@vger.kernel.org>,
        <linux-security-module@vger.kernel.org>,
        <keyrings@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
        <silviu.vlasceanu@huawei.com>, <ndesaulniers@google.com>
References: <20190206162452.7749-1-roberto.sassu@huawei.com>
 <20190206162452.7749-5-roberto.sassu@huawei.com>
 <20190208042411.GA4822@archlinux-ryzen>
From:   Roberto Sassu <roberto.sassu@huawei.com>
Message-ID: <cded9b3a-e321-ed05-4d41-51b8dab45c8d@huawei.com>
Date:   Fri, 8 Feb 2019 09:41:14 +0100
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:60.0) Gecko/20100101
 Thunderbird/60.3.0
MIME-Version: 1.0
In-Reply-To: <20190208042411.GA4822@archlinux-ryzen>
Content-Type: text/plain; charset="utf-8"; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 2/8/2019 5:24 AM, Nathan Chancellor wrote:
>> diff --git a/include/linux/tpm.h b/include/linux/tpm.h
>> index afd022fc9d3d..816e686a73ac 100644
>> --- a/include/linux/tpm.h
>> +++ b/include/linux/tpm.h
>> @@ -22,6 +22,10 @@
>>   #ifndef __LINUX_TPM_H__
>>   #define __LINUX_TPM_H__
>>   
>> +#include <linux/hw_random.h>
>> +#include <linux/acpi.h>

Hi Nathan

I think the error comes from the line above.
security/integrity/ima/ima.h includes <linux/tpm.h>, which now includes
<linux/acpi.h>, which includes <asm/acpi.h>, which includes
<linux/efi.h> (for the arm64 architecture only). Both ima.h and efi.h
define 'NONE'.

The solution would be to rename one of them. I'm not familiar with the
EFI part. Renaming 'NONE' in IMA should be easy as it is not used
anywhere.

Thanks

Roberto


>> +#include <linux/cdev.h>
>> +#include <linux/fs.h>
>>   #include <crypto/hash_info.h>
>>   
>>   #define TPM_DIGEST_SIZE 20	/* Max TPM v1.2 PCR size */
>> @@ -75,6 +79,93 @@ struct tpm_class_ops {
>>   	void (*clk_enable)(struct tpm_chip *chip, bool value);
>>   };
>>   
>> +#define TPM_NUM_EVENT_LOG_FILES		3
>> +
>> +/* Indexes the duration array */
>> +enum tpm_duration {
>> +	TPM_SHORT = 0,
>> +	TPM_MEDIUM = 1,
>> +	TPM_LONG = 2,
>> +	TPM_LONG_LONG = 3,
>> +	TPM_UNDEFINED,
>> +	TPM_NUM_DURATIONS = TPM_UNDEFINED,
>> +};
>> +
>> +#define TPM_PPI_VERSION_LEN		3
>> +
>> +struct tpm_space {
>> +	u32 context_tbl[3];
>> +	u8 *context_buf;
>> +	u32 session_tbl[3];
>> +	u8 *session_buf;
>> +};
>> +
>> +struct tpm_bios_log {
>> +	void *bios_event_log;
>> +	void *bios_event_log_end;
>> +};
>> +
>> +struct tpm_chip_seqops {
>> +	struct tpm_chip *chip;
>> +	const struct seq_operations *seqops;
>> +};
>> +
>> +struct tpm_chip {
>> +	struct device dev;
>> +	struct device devs;
>> +	struct cdev cdev;
>> +	struct cdev cdevs;
>> +
>> +	/* A driver callback under ops cannot be run unless ops_sem is held
>> +	 * (sometimes implicitly, eg for the sysfs code). ops becomes null
>> +	 * when the driver is unregistered, see tpm_try_get_ops.
>> +	 */
>> +	struct rw_semaphore ops_sem;
>> +	const struct tpm_class_ops *ops;
>> +
>> +	struct tpm_bios_log log;
>> +	struct tpm_chip_seqops bin_log_seqops;
>> +	struct tpm_chip_seqops ascii_log_seqops;
>> +
>> +	unsigned int flags;
>> +
>> +	int dev_num;		/* /dev/tpm# */
>> +	unsigned long is_open;	/* only one allowed */
>> +
>> +	char hwrng_name[64];
>> +	struct hwrng hwrng;
>> +
>> +	struct mutex tpm_mutex;	/* tpm is processing */
>> +
>> +	unsigned long timeout_a; /* jiffies */
>> +	unsigned long timeout_b; /* jiffies */
>> +	unsigned long timeout_c; /* jiffies */
>> +	unsigned long timeout_d; /* jiffies */
>> +	bool timeout_adjusted;
>> +	unsigned long duration[TPM_NUM_DURATIONS]; /* jiffies */
>> +	bool duration_adjusted;
>> +
>> +	struct dentry *bios_dir[TPM_NUM_EVENT_LOG_FILES];
>> +
>> +	const struct attribute_group *groups[3];
>> +	unsigned int groups_cnt;
>> +
>> +	u32 nr_allocated_banks;
>> +	struct tpm_bank_info *allocated_banks;
>> +#ifdef CONFIG_ACPI
>> +	acpi_handle acpi_dev_handle;
>> +	char ppi_version[TPM_PPI_VERSION_LEN + 1];
>> +#endif /* CONFIG_ACPI */
>> +
>> +	struct tpm_space work_space;
>> +	u32 last_cc;
>> +	u32 nr_commands;
>> +	u32 *cc_attrs_tbl;
>> +
>> +	/* active locality */
>> +	int locality;
>> +};
>> +
>>   #if defined(CONFIG_TCG_TPM) || defined(CONFIG_TCG_TPM_MODULE)
>>   
>>   extern int tpm_is_tpm2(struct tpm_chip *chip);
>> -- 
>> 2.17.1
>>
> 
> Hi Robert,
> 
> This patch causes a build error with Clang (bisected on next-20190207):
> 
> security/integrity/ima/ima.h:191:2: error: redefinition of enumerator 'NONE'
>          __ima_hooks(__ima_hook_enumify)
>          ^
> security/integrity/ima/ima.h:176:7: note: expanded from macro '__ima_hooks'
>          hook(NONE)                      \
>               ^
> include/linux/efi.h:1709:2: note: previous definition is here
>          NONE,
>          ^
> 1 error generated.
> 
> I am not sure how to reconcile this otherwise I would have sent a patch.
> 
> Thanks,
> Nathan
> 

-- 
HUAWEI TECHNOLOGIES Duesseldorf GmbH, HRB 56063
Managing Director: Bo PENG, Jian LI, Yanli SHI