Messages in this thread |  | | | Date | Thu, 20 Sep 2012 17:18:57 -0400 | | From | Konrad Rzeszutek Wilk <> | | Subject | Re: 3.6rc6 slab corruption. |
| |
> >An alternative to this, though, might be to never test for *ppos == 0 in > >u32_array_read() and do the format_array_alloc() in u32_array_open() to > >initialize file->private_data. If that allocation fails, just return > >-ENOMEM. Then you never need to add a mutex in the read path. > > > > Tested-by: Raghavendra <raghavendra.kt@linux.vnet.ibm.com>
Excellent. Thank you!
>
|  |