Hello, time for some necro reply! :D

Did I understand right that I can try to disable MSI and other things by
flags in config/boot -c ?

If yes, then I tried to use 0x8 (and other combinations) with no success :-(

#define AHCI_F_NO_NCQ                   (1<<0)
#define AHCI_F_IPMS_PROBE               (1<<1)  /* IPMS on failed PMP probe
*/
#define AHCI_F_NO_PMP                   (1<<2)  /* ignore PMP capability */
#define AHCI_F_NO_MSI                   (1<<3)  /* disable MSI */



сб, 22 янв. 2022 г. в 22:45, Christopher Zimmermann <[email protected]>:

> On Sat, Jan 22, 2022 at 06:25:08PM +0300, S V wrote:
> >Hello, I continue my hard way of launching OpenBSD on Baikal-M board, I
> was
> >able to get to usb/uart console and successfully launched bsd.rd and
> >ctrl-C'ed to shell, BUT
> >
> >I'm stuck with extremely slow SATA access, literally slow "time fdisk sd0"
> >takes 3 minutes!
> >
> >related dmesg lines
> >
> >mainbus0 at root: Elpitech Baikal-M mitx board
> >
> >simplebus0 at mainbus0: "soc"
> >ahci0 at simplebus0: AHCI 1.3
> >
> >ahci0: port 0: 6.0Gb/s
> >scsibus0 at ahci0: 32 targets
> >sd0 at scsibus0 targ 0 lun 0: <ATA, TEAM T253X2128G, U040>
> >t10.ATA_TEAM_T253X2128G_TPBF2108200090201893
> >
> >sd0: 122104MB, 512 bytes/sector, 250069680 sectors, thin
> >
> >
> >and full dmesg from ramdisk attached
>
> Hi,
>
> I suffered from similar problems caused by a JMicron JMB582 controller
> put into a RockPro64. It did not manage to properly deliver interrupts
> over msi. Forcing it to intx solved the problem for me. After adding a
> workaround to the driver I sent the sata card to kettenis@, so that he
> can try it on a different board. Your dmesg doesn't identify the ahci
> chipset and interrupt mode.
> This is what I did:
> https://marc.info/?l=openbsd-arm&m=163735423105422&w=2
>
> maybe something similar will help your controller deliver interrupts?
>
>
> Christopher
>


-- 
Nerfur Dragon
-==(UDIC)==-

Reply via email to