lkml.org 
[lkml]   [2008]   [Jun]   [6]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] Added in user-injected messages into blk traces
On Thu, Jun 05 2008, Alan D. Brunelle wrote:

> This allows a user to annotate the blk trace stream: writing a suitable
> message to {/sys/kernel/debug}/block/<dsf>/msg will have it propagated
> into the trace stream.

Looks good to me, I think this can be useful for ease of trace
annotation. Comments below.

> Signed-off-by: Alan D. Brunelle <alan.brunelle@hp.com>
> ---
> block/blktrace.c | 42 ++++++++++++++++++++++++++++++++++++++++++
> include/linux/blktrace_api.h | 1 +
> 2 files changed, 43 insertions(+), 0 deletions(-)
>
> diff --git a/block/blktrace.c b/block/blktrace.c
> index 38e6b83..b9c4df2 100644
> --- a/block/blktrace.c
> +++ b/block/blktrace.c
> @@ -253,6 +253,7 @@ err:
> static void blk_trace_cleanup(struct blk_trace *bt)
> {
> relay_close(bt->rchan);
> + debugfs_remove(bt->msg_file);
> debugfs_remove(bt->dropped_file);
> blk_remove_tree(bt->dir);
> free_percpu(bt->sequence);
> @@ -299,6 +300,41 @@ static const struct file_operations blk_dropped_fops = {
> .read = blk_dropped_read,
> };
>
> +static int blk_msg_open(struct inode *inode, struct file *filp)
> +{
> + filp->private_data = inode->i_private;
> +
> + return 0;
> +}
> +
> +static ssize_t blk_msg_write(struct file *filp, const char __user *buffer,
> + size_t count, loff_t *ppos)
> +{
> + ssize_t ret = count;
> +
> + if (count > BLK_TN_MAX_MSG)
> + ret = -EINVAL;
> + else {
> + char *msg = kmalloc(BLK_TN_MAX_MSG, GFP_KERNEL);
> +
> + if (copy_from_user(msg, buffer, count))
> + ret = -EFAULT;
> + else {
> + struct blk_trace *bt = filp->private_data;
> + __trace_note_message(bt, "%s", msg);
> + }
> + kfree(msg);

Would be cleaner with an explicit !msg check, though I think
copy_from_user() will notice and it'll work as-is.

if (count > BLK_TN_MAX_MSG)
return -EINVAL;

msg = kmalloc(BLK_TN_MAX_MSG, GFP_KERNEL);
if (!msg)
return -ENOMEM;

if (copy_from_user(msg, buffer, count))
return -EFAULT;

...

is just a lot easier to directly follow imho. Otherwise I have no
problems with it, if you fix that up we can queue it for 2.6.27
inclusion.

--
Jens Axboe



\
 
 \ /
  Last update: 2008-06-06 11:27    [W:0.057 / U:1.492 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site