Messages in this thread | | | Date | Tue, 25 Sep 2018 14:23:00 +0200 | From | Petr Mladek <> | Subject | Re: [PATCH v2 1/2] printk: Fix panic caused by passing log_buf_len to command line |
| |
On Tue 2018-09-25 21:01:35, Sergey Senozhatsky wrote: > On (09/21/18 09:37), Petr Mladek wrote: > > > > I would personally keep the size as unsigned int. IMHO, a support > > for a log buffer bigger than 4GB is not worth the complexity. > > > > ftrace dumps are bothering me. > > Steven Rostedt wrote [0]: > > > > Especially when I have a machine with 240 CPUs. But it also has a ton of > > RAM, I could easily do log_buf_len=32G > > > > The systems are getting bigger, so log_buf_len=UINT_MAX+ might become > a norm at some point.
Thanks for pointing this out. Well, it seems that the change would require a new syscall to pass the buffer size as long. We need to be sure that people would use this in the real life.
This thread suggested this change to avoid a checkpatch warning. The 32GB was mentioned as an example one year ego. This is not enough for a new syscall from my point of view.
Best Regards, Petr
| |