Hello,
Just an update: I have fixed other issues with netdde crashing, the
latest fixes are available in 2:1.8+git20191029-7, to be built &
uploaded soon.
One thing that remains, however, is this: if netdde crashes while an irq
is pending, it will not make the hardware drop the irq. The interrupt
Samuel Thibault, le lun. 11 nov. 2019 19:06:21 +0100, a ecrit:
> Justus Winter, le lun. 11 nov. 2019 19:03:46 +0100, a ecrit:
> > I remember thinking about using a send-once right that when deallocated
> > enables the interrupt again. This should be a) well optimized, and b)
> > robust in the face
Justus Winter, le lun. 11 nov. 2019 19:03:46 +0100, a ecrit:
> I remember thinking about using a send-once right that when deallocated
> enables the interrupt again. This should be a) well optimized, and b)
> robust in the face of misbehaving userland.
I was thinking about something like that to
Samuel Thibault writes:
> A simpler way for now would be to make deliver_intr and
> device_intr_enable count how many "disables" we have for each irq.
> Both device_intr_register and deliver_intr would increase it (and thus
> disable the irq), and userland would have to call device_intr_enable
>
Samuel Thibault, le lun. 11 nov. 2019 17:52:36 +0100, a ecrit:
> That being said, now netdde with e1000 works only because linux_intr
> calls disable_irq before queue_intr. If that is not done, we keep
> getting interrupts. This really looks like an edge-vs-level trigger
> issue which we need to cl
Damien Zammit, le lun. 11 nov. 2019 21:23:15 +1100, a ecrit:
> On 11/11/19 3:08 am, Samuel Thibault wrote:
> > Samuel Thibault, le dim. 10 nov. 2019 16:25:22 +0100, a ecrit:
> >> Samuel Thibault, le dim. 10 nov. 2019 16:22:04 +0100, a ecrit:
> >>> Ok, I have restored some of the mask/unmask_irq, so
On 11/11/19 3:08 am, Samuel Thibault wrote:
> Samuel Thibault, le dim. 10 nov. 2019 16:25:22 +0100, a ecrit:
>> Samuel Thibault, le dim. 10 nov. 2019 16:22:04 +0100, a ecrit:
>>> Ok, I have restored some of the mask/unmask_irq, so that curr_pic_mask
>>> will still exist, but without the notion of s
BTW, this cleanup may help with kvm-in-kvm emulation such as used on
jenkins.debian.net: IIRC that was getting performance issues because of
interrupt simulation misperformance. Using cli/sti most of the time will
probably help here.
Samuel
Samuel Thibault, le dim. 10 nov. 2019 16:25:22 +0100, a ecrit:
> Samuel Thibault, le dim. 10 nov. 2019 16:22:04 +0100, a ecrit:
> > Ok, I have restored some of the mask/unmask_irq, so that curr_pic_mask
> > will still exist, but without the notion of spl: it will just store
> > which IRQs have been
Samuel Thibault, le dim. 10 nov. 2019 16:22:04 +0100, a ecrit:
> Ok, I have restored some of the mask/unmask_irq, so that curr_pic_mask
> will still exist, but without the notion of spl: it will just store
> which IRQs have been effectively disabled. That'll already be good for a
> first step :)
A
Ok, I have restored some of the mask/unmask_irq, so that curr_pic_mask
will still exist, but without the notion of spl: it will just store
which IRQs have been effectively disabled. That'll already be good for a
first step :)
Samuel
Damien Zammit, le sam. 09 nov. 2019 15:46:33 +1100, a ecrit:
> diff --git a/i386/i386/pic.h b/i386/i386/pic.h
> index 80bf65d..abf0993 100644
> --- a/i386/i386/pic.h
> +++ b/i386/i386/pic.h
> @@ -176,11 +176,18 @@ WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
> #define READ_IR_ONRD 0x00
>
Hello,
Damien Zammit, le sam. 09 nov. 2019 15:46:33 +1100, a ecrit:
> The first patch combines all spls 1-7 into a single spl that disables all
> interrupts.
> It boots with just this patch.
I have applied this one, thanks!
> The second patch removes all concept of a cached PIC mask, and allows
Hi there,
Please find attached 2x patches for gnumach (apply cleanly on debian gnumach).
The first patch combines all spls 1-7 into a single spl that disables all
interrupts.
It boots with just this patch.
The second patch removes all concept of a cached PIC mask, and allows the PIC
to remain
14 matches
Mail list logo
