[edk2-devel] [PATCH] ArmPkg/ArmGicLib: fix maximum interrupts supported

[Adrián Herrera]

Hi Sami,

That patch provides the same solution. I'll be sure to check the mailing
list before sending the next time.
Any benefit on marking this as duplicate? If so, how could I do it?

Thank you,
Adrián.

On Mon, 26 Apr 2021 at 08:56, Sami Mujawar <Sami.Mujawar at arm.com> wrote:

> Hi Adrian,
>
>
>
> I believe there is already a similar patch on the mailing list at
> https://edk2.groups.io/g/devel/message/72596. This patch is pending
> review and tested-by.
>
> Can you check if this patch covers the problems you describe, please?
>
>
>
> Regards,
>
>
>
> Sami Mujawar
>
>
>
> *From: *devel at edk2.groups.io <devel at edk2.groups.io> on behalf of Adrián
> Herrera via groups.io <adr.her.arc.95=gmail.com at groups.io>
> *Date: *Saturday, 24 April 2021 at 03:57
> *To: *devel at edk2.groups.io <devel at edk2.groups.io>
> *Cc: *Adrián Herrera <adr.her.arc.95 at gmail.com>
> *Subject: *[edk2-devel] [PATCH] ArmPkg/ArmGicLib: fix maximum interrupts
> supported
>
> The maximum number of interrupts supported is determined as
> 32 * (GICD_TYPER.ITLinesNumber + 1).
>
> When GICD_TYPER.ITLinesNumber = 0b11111, the maximum number of
> interrupts supported is 1024. However, both GICv2 and GICv3 reserve
> INTIDs 1020-1023 for special purposes.
>
> This results in runtime crashes because:
>   (1) ArmGicLib functions do not guard against special interrupts.
>   (2) ArmGicGetMaxNumInterrupts number includes special interrupts.
>   (2) ArmGicV*Dxe relies on ArmGicGetMaxNumInterrupts, and thus
>       programs special interrupts through ArmGicLib.
>
> ArmGicGetMaxNumInterrupts now does not include special interrupts, that
> is, it reports 1020 instead of 1024 when GICD_TYPER.ITLinesNumber =
> 0b11111.
> This avoids the overhead of guarding ArmGicLib functions while not
> requiring to modify the drivers code.
>
> Signed-off-by: Adrián Herrera <adr.her.arc.95 at gmail.com>
> ---
>  ArmPkg/Drivers/ArmGic/ArmGicLib.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/ArmPkg/Drivers/ArmGic/ArmGicLib.c
> b/ArmPkg/Drivers/ArmGic/ArmGicLib.c
> index 6b01c88206..dff1401e9c 100644
> --- a/ArmPkg/Drivers/ArmGic/ArmGicLib.c
> +++ b/ArmPkg/Drivers/ArmGic/ArmGicLib.c
> @@ -120,7 +120,10 @@ ArmGicGetMaxNumInterrupts (
>    IN  INTN          GicDistributorBase
>
>
>    )
>
>
>  {
>
>
> -  return 32 * ((MmioRead32 (GicDistributorBase + ARM_GIC_ICDICTR) & 0x1F)
> + 1);
>
>
> +  UINT32 ITLinesNumber;
>
>
> +
>
>
> +  ITLinesNumber = MmioRead32 (GicDistributorBase + ARM_GIC_ICDICTR) &
> 0x1F;
>
>
> +  return MIN (32 * (ITLinesNumber+ 1), 1020);
>
>
>  }
>
>
>
>
>
>  VOID
>
>
> --
> 2.30.0
>
>
>
> 
>
>


-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#74445): https://edk2.groups.io/g/devel/message/74445
Mute This Topic: https://groups.io/mt/82327316/1813853
Group Owner: devel+owner at edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [edk2-devel-archive at redhat.com]
-=-=-=-=-=-=-=-=-=-=-=-


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/edk2-devel-archive/attachments/20210427/f3dcae2b/attachment.htm>