Re: [RFC PATCH 01/39] blktrace_api: add new trace definitions

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2021/02/25 16:03, Chaitanya Kulkarni wrote:
> This patch adds a new trace categories, trace actions adds a new
> version number for the trace extentions, adds new trace extension
> structure to hold actual trace along with structure nedded to execute
> various IOCTLs to configure trace from user space.
> 
> Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@xxxxxxx>
> ---
>  include/uapi/linux/blktrace_api.h | 110 ++++++++++++++++++++++++------
>  1 file changed, 89 insertions(+), 21 deletions(-)
> 
> diff --git a/include/uapi/linux/blktrace_api.h b/include/uapi/linux/blktrace_api.h
> index 690621b610e5..fdb3a5cdfa22 100644
> --- a/include/uapi/linux/blktrace_api.h
> +++ b/include/uapi/linux/blktrace_api.h
> @@ -8,30 +8,41 @@
>   * Trace categories
>   */
>  enum blktrace_cat {
> -	BLK_TC_READ	= 1 << 0,	/* reads */
> -	BLK_TC_WRITE	= 1 << 1,	/* writes */
> -	BLK_TC_FLUSH	= 1 << 2,	/* flush */
> -	BLK_TC_SYNC	= 1 << 3,	/* sync IO */
> -	BLK_TC_SYNCIO	= BLK_TC_SYNC,
> -	BLK_TC_QUEUE	= 1 << 4,	/* queueing/merging */
> -	BLK_TC_REQUEUE	= 1 << 5,	/* requeueing */
> -	BLK_TC_ISSUE	= 1 << 6,	/* issue */
> -	BLK_TC_COMPLETE	= 1 << 7,	/* completions */
> -	BLK_TC_FS	= 1 << 8,	/* fs requests */
> -	BLK_TC_PC	= 1 << 9,	/* pc requests */
> -	BLK_TC_NOTIFY	= 1 << 10,	/* special message */
> -	BLK_TC_AHEAD	= 1 << 11,	/* readahead */
> -	BLK_TC_META	= 1 << 12,	/* metadata */
> -	BLK_TC_DISCARD	= 1 << 13,	/* discard requests */
> -	BLK_TC_DRV_DATA	= 1 << 14,	/* binary per-driver data */
> -	BLK_TC_FUA	= 1 << 15,	/* fua requests */
> -
> -	BLK_TC_END	= 1 << 15,	/* we've run out of bits! */
> +	BLK_TC_READ		= 1 << 0,	/* reads */
> +	BLK_TC_WRITE		= 1 << 1,	/* writes */
> +	BLK_TC_FLUSH		= 1 << 2,	/* flush */
> +	BLK_TC_SYNC		= 1 << 3,	/* sync IO */
> +	BLK_TC_SYNCIO		= BLK_TC_SYNC,
> +	BLK_TC_QUEUE		= 1 << 4,	/* queueing/merging */
> +	BLK_TC_REQUEUE		= 1 << 5,	/* requeueing */
> +	BLK_TC_ISSUE		= 1 << 6,	/* issue */
> +	BLK_TC_COMPLETE		= 1 << 7,	/* completions */
> +	BLK_TC_FS		= 1 << 8,	/* fs requests */
> +	BLK_TC_PC		= 1 << 9,	/* pc requests */
> +	BLK_TC_NOTIFY		= 1 << 10,	/* special message */
> +	BLK_TC_AHEAD		= 1 << 11,	/* readahead */
> +	BLK_TC_META		= 1 << 12,	/* metadata */
> +	BLK_TC_DISCARD		= 1 << 13,	/* discard requests */
> +	BLK_TC_DRV_DATA		= 1 << 14,	/* binary per-driver data */
> +	BLK_TC_FUA		= 1 << 15,	/* fua requests */
> +	BLK_TC_WRITE_ZEROES	= 1 << 16,	/* write-zeores */
> +	BLK_TC_ZONE_RESET	= 1 << 17,	/* zone-reset */
> +	BLK_TC_ZONE_RESET_ALL	= 1 << 18,	/* zone-reset-all */
> +	BLK_TC_ZONE_APPEND	= 1 << 19,	/* zone-append */
> +	BLK_TC_ZONE_OPEN	= 1 << 20,	/* zone-open */
> +	BLK_TC_ZONE_CLOSE	= 1 << 21,	/* zone-close */
> +	BLK_TC_ZONE_FINISH	= 1 << 22,	/* zone-finish */
> +
> +	BLK_TC_END		= 1 << 15,	/* we've run out of bits! */

BLK_TC_FUA has the same value. Is that intentional ?

> +	BLK_TC_END_EXT		= 1 << 31,	/* we've run out of bits! */
>  };
>  
>  #define BLK_TC_SHIFT		(16)
>  #define BLK_TC_ACT(act)		((act) << BLK_TC_SHIFT)
>  
> +#define BLK_TC_SHIFT_EXT   	(32)
> +#define BLK_TC_ACT_EXT(act)		(((u64)act) << BLK_TC_SHIFT_EXT)
> +
>  /*
>   * Basic trace actions
>   */
> @@ -88,12 +99,38 @@ enum blktrace_notify {
>  #define BLK_TA_ABORT		(__BLK_TA_ABORT | BLK_TC_ACT(BLK_TC_QUEUE))
>  #define BLK_TA_DRV_DATA	(__BLK_TA_DRV_DATA | BLK_TC_ACT(BLK_TC_DRV_DATA))
>  
> +#define BLK_TA_QUEUE_EXT	(__BLK_TA_QUEUE | BLK_TC_ACT_EXT(BLK_TC_QUEUE))
> +#define BLK_TA_BACKMERGE_EXT	(__BLK_TA_BACKMERGE | BLK_TC_ACT_EXT(BLK_TC_QUEUE))
> +#define BLK_TA_FRONTMERGE_EXT	(__BLK_TA_FRONTMERGE | BLK_TC_ACT_EXT(BLK_TC_QUEUE))
> +#define BLK_TA_GETRQ_EXT	(__BLK_TA_GETRQ | BLK_TC_ACT_EXT(BLK_TC_QUEUE))
> +#define BLK_TA_SLEEPRQ_EXT	(__BLK_TA_SLEEPRQ | BLK_TC_ACT_EXT(BLK_TC_QUEUE))
> +#define BLK_TA_REQUEUE_EXT	(__BLK_TA_REQUEUE | BLK_TC_ACT_EXT(BLK_TC_REQUEUE))
> +#define BLK_TA_ISSUE_EXT	(__BLK_TA_ISSUE | BLK_TC_ACT_EXT(BLK_TC_ISSUE))
> +#define BLK_TA_COMPLETE_EXT	(__BLK_TA_COMPLETE | BLK_TC_ACT_EXT(BLK_TC_COMPLETE))
> +#define BLK_TA_PLUG_EXT		(__BLK_TA_PLUG | BLK_TC_ACT_EXT(BLK_TC_QUEUE))
> +#define BLK_TA_UNPLUG_IO_EXT	(__BLK_TA_UNPLUG_IO | BLK_TC_ACT_EXT(BLK_TC_QUEUE))
> +#define BLK_TA_UNPLUG_TIMER_EXT		\
> +	(__BLK_TA_UNPLUG_TIMER | BLK_TC_ACT_EXT(BLK_TC_QUEUE))
> +#define BLK_TA_INSERT_EXT	(__BLK_TA_INSERT | BLK_TC_ACT_EXT(BLK_TC_QUEUE))
> +#define BLK_TA_SPLIT_EXT	(__BLK_TA_SPLIT)
> +#define BLK_TA_BOUNCE_EXT	(__BLK_TA_BOUNCE)
> +#define BLK_TA_REMAP_EXT	(__BLK_TA_REMAP | BLK_TC_ACT_EXT(BLK_TC_QUEUE))
> +#define BLK_TA_ABORT_EXT	(__BLK_TA_ABORT | BLK_TC_ACT_EXT(BLK_TC_QUEUE))
> +#define BLK_TA_DRV_DATA_EXT	\
> +	(__BLK_TA_DRV_DATA | BLK_TC_ACT_EXT(BLK_TC_DRV_DATA))
> +
>  #define BLK_TN_PROCESS		(__BLK_TN_PROCESS | BLK_TC_ACT(BLK_TC_NOTIFY))
>  #define BLK_TN_TIMESTAMP	(__BLK_TN_TIMESTAMP | BLK_TC_ACT(BLK_TC_NOTIFY))
>  #define BLK_TN_MESSAGE		(__BLK_TN_MESSAGE | BLK_TC_ACT(BLK_TC_NOTIFY))
>  
> -#define BLK_IO_TRACE_MAGIC	0x65617400
> -#define BLK_IO_TRACE_VERSION	0x07
> +#define BLK_TN_PROCESS_EXT	(__BLK_TN_PROCESS | BLK_TC_ACT_EXT(BLK_TC_NOTIFY))
> +#define BLK_TN_TIMESTAMP_EXT	(__BLK_TN_TIMESTAMP | BLK_TC_ACT_EXT(BLK_TC_NOTIFY))
> +#define BLK_TN_MESSAGE_EXT	(__BLK_TN_MESSAGE | BLK_TC_ACT_EXT(BLK_TC_NOTIFY))
> +
> +#define BLK_IO_TRACE_MAGIC             0x65617400
> +#define BLK_IO_TRACE_VERSION           0x07
> +#define BLK_IO_TRACE_VERSION_EXT       0x08

It is a little weird to have 2 versions. Why not simply increase the version
number ? BLK_IO_TRACE_VERSION == 7 means "support only old trace format" and
BLK_IO_TRACE_VERSION == 8 means "support old and new extended trace format"
would be better. From just the code in this patch, not sure how this is being
used though.

> +

blank line not needed.

>  
>  /*
>   * The trace itself
> @@ -113,6 +150,23 @@ struct blk_io_trace {
>  	/* cgroup id will be stored here if exists */
>  };
>  
> +struct blk_io_trace_ext {
> +	__u32 magic;		/* MAGIC << 8 | version */
> +	__u32 sequence;		/* event number */
> +	__u64 time;		/* in nanoseconds */
> +	__u64 sector;		/* disk offset */
> +	__u32 bytes;		/* transfer length */
> +	__u64 action;		/* what happened */
> +	__u32 ioprio;		/* I/O priority */
> +	__u32 pid;		/* who did it */
> +	__u32 device;		/* device number */
> +	__u32 cpu;		/* on what cpu did it happen */
> +	__u16 error;		/* completion error */
> +	__u16 pdu_len;		/* length of data after this trace */
> +	/* cgroup id will be stored here if exists */
> +};
> +
> +

extra blank line not needed.

>  /*
>   * The remap event
>   */
> @@ -143,4 +197,18 @@ struct blk_user_trace_setup {
>  	__u32 pid;
>  };
>  
> +/*
> + * User setup structure passed with BLKTRACESETUP_EXT
> + */
> +struct blk_user_trace_setup_ext {
> +	char name[BLKTRACE_BDEV_SIZE];	/* output */
> +	__u64 act_mask;			/* input */
> +	__u32 prio_mask;		/* input */
> +	__u32 buf_size;			/* input */
> +	__u32 buf_nr;			/* input */
> +	__u64 start_lba;
> +	__u64 end_lba;
> +	__u32 pid;
> +};
> +
>  #endif /* _UAPIBLKTRACE_H */
> 


-- 
Damien Le Moal
Western Digital Research




[Index of Archives]     [Linux RAID]     [Linux SCSI]     [Linux ATA RAID]     [IDE]     [Linux Wireless]     [Linux Kernel]     [ATH6KL]     [Linux Bluetooth]     [Linux Netdev]     [Kernel Newbies]     [Security]     [Git]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Device Mapper]

  Powered by Linux