> > Worth me looking into ?
>
> I have to give this a lot more thought.
>
> The universality of n_tty is important, and costs real cycles on servers and
> such. It's not just about typing speed.
For most systems its about ppp performance and nothing much else in real
use. I'm not arguing for thi
On 12/04/2013 07:13 PM, One Thousand Gnomes wrote:
Not so much confused as simply merged. Input processing is inherently
single-threaded; it makes sense to rely on that at the highest level
possible.
I would disagree entirely. You want to minimise the areas affected by a
given lock. You also wa
On 12/08/2013 08:12 PM, Greg Kroah-Hartman wrote:
On Mon, Dec 02, 2013 at 04:12:01PM -0500, Peter Hurley wrote:
Greg,
Sometimes when interrupting terminal output, the '^C' won't be echoed
until more output is echoed. This is fairly repeatable by interrupting
'cat large-file'.
The common reason
On Mon, Dec 02, 2013 at 04:12:01PM -0500, Peter Hurley wrote:
> Greg,
>
> Sometimes when interrupting terminal output, the '^C' won't be echoed
> until more output is echoed. This is fairly repeatable by interrupting
> 'cat large-file'.
>
> The common reason for this is because the tty write buff
> Not so much confused as simply merged. Input processing is inherently
> single-threaded; it makes sense to rely on that at the highest level
> possible.
I would disagree entirely. You want to minimise the areas affected by a
given lock. You also want to lock data not code. Correctness comes befo
On 12/03/2013 09:20 AM, One Thousand Gnomes wrote:
These types of nested lock problems are common when different layers use
the same interface (the fb subsystem's use of the vt driver is another
example).
They are, and they end up nasty and eventually become impossible to fix.
Better to fix the
On 12/03/2013 12:23 PM, Peter Hurley wrote:
On 12/03/2013 09:20 AM, One Thousand Gnomes wrote:
As a PS: I think the termios is probably an RCU problem. The semaphore
seemed to make sense when I did it, but in hindsight I think I made the
wrong call.
Unfortunately, not without overhauling the c
On 12/03/2013 09:20 AM, One Thousand Gnomes wrote:
As a PS: I think the termios is probably an RCU problem. The semaphore
seemed to make sense when I did it, but in hindsight I think I made the
wrong call.
Alan,
Converting safe termios access to RCU is a really good idea.
Massive changeset th
> These types of nested lock problems are common when different layers use
> the same interface (the fb subsystem's use of the vt driver is another
> example).
They are, and they end up nasty and eventually become impossible to fix.
Better to fix the underlying fundamental error as and when we can
On 12/02/2013 07:01 PM, One Thousand Gnomes wrote:
I cc'd you because of your recent involvement in other
tty patches/bug fixes and because it's your FIXME comment.
Feel free to ignore and/or let me know you would prefer not to
be bothered.
It does seem horribly convoluted and likely to dig big
> I cc'd you because of your recent involvement in other
> tty patches/bug fixes and because it's your FIXME comment.
> Feel free to ignore and/or let me know you would prefer not to
> be bothered.
It does seem horribly convoluted and likely to dig bigger long term holes
than the one its filling.
Greg,
Sometimes when interrupting terminal output, the '^C' won't be echoed
until more output is echoed. This is fairly repeatable by interrupting
'cat large-file'.
The common reason for this is because the tty write buffer is full,
even though the write buffer _should_ have been flushed already.
12 matches
Mail list logo