Messages in this thread | | | Date | Tue, 5 Sep 2023 14:24:50 -0600 | Subject | Re: [PATCH v6 2/5] PCI: brcmstb: Configure HW CLKREQ# mode appropriate for downstream device | From | Sam Edwards <> |
| |
On 6/23/23 08:40, Jim Quinlan wrote:
Hi Jim,
As I understand, the issue that this patch fixes can be difficult to trigger reliably. So I'll report that the abort occurs 100% of the time on a Compute Module 4, installed in slot 3 of a Turing Pi 2 cluster board, when a bootloader (U-Boot, in my case) is in the boot path. The third slot of the TP2 board connects to an ASM1061 chip with the CLKREQ# line left floating. I do not know why the bootloader is necessary for the crash to occur; I suspect it's timing-related.
With this patch applied to my kernel, my observed crash rate goes from 100% to 0%, so:
> Signed-off-by: Jim Quinlan <james.quinlan@broadcom.com> > Tested-by: Florian Fainelli <f.fainelli@gmail.com> > Tested-by: Cyril Brulebois <cyril@debamax.com> Tested-by: Sam Edwards <CFSworks@gmail.com>
Many thanks, Sam
| |