Date: Mon, 4 Jun 2018 16:22:01 -0600 From: Scott Long <scottl@samsco.org> To: Harry Schmalzbauer <freebsd@omnilan.de> Cc: scsi@freebsd.org Subject: =?utf-8?Q?Re=3A_What_is_ENXIO_=E2=80=93_MSI_allocation_regression?= =?utf-8?Q?_in_=3A=5BWas_Re=3A_svn_commit=3A_r321714_-_in_head/sys/dev=3A_?= =?utf-8?Q?mpr_mps=5D?= Message-ID: <78611650-D7A4-4B1D-A254-DB058E1AC1C6@samsco.org> In-Reply-To: <e7d94e6a-89e8-ffa1-40da-7fb67e6bfc2b@omnilan.de> References: <201707300653.v6U6rwLN099096@repo.freebsd.org> <597DA578.6030101@omnilan.de> <597F56A8.1060603@omnilan.de> <D18DFAD4-6E93-4AE2-BE15-EFF4D8ABCB2A@samsco.org> <59804C8C.1020003@omnilan.de> <e7d94e6a-89e8-ffa1-40da-7fb67e6bfc2b@omnilan.de>
next in thread | previous in thread | raw e-mail | index | archive | help
> On Jun 4, 2018, at 4:51 AM, Harry Schmalzbauer <freebsd@omnilan.de> =
wrote:
>=20
> Am 01.08.2017 um 11:40 schrieb Harry Schmalzbauer:
>> Bez=C3=BCglich Scott Long's Nachricht vom 31.07.2017 18:56 =
(localtime):
>>=20
>> =E2=80=A6
>>>> I'd like to report one I hadn't expected:
>>>>=20
>>>> mps0: <Avago Technologies (LSI) SAS2008> port 0x4000-0x40ff mem =
0xc3bc0000-0xc3bc3fff,0xc3b80000-0xc3bbffff irq 19 at device 0.0 on pci7
>>>>=20
>>>> mps0: Firmware: 20.00.04.00, Driver: 21.02.00.00-fbsd
>>>> mps0: IOCCapabilities:
>>>> 185c<ScsiTaskFull,DiagTrace,SnapBuf,EEDP,TransRetry,IR>
>>>> mps0: Cannot allocate INTx interrupt
>>>> mps0: mps_iocfacts_allocate failed to setup interrupts
>>>> mps0: mps_attach IOC Facts based allocation failed with error 6
>>>> panic: resource_list_release: resource entry is not busy
>>>> cpuid =3D 6
>>>> KDB: stack backtrace:
>>>> #0 0xffffffff805e32d7 at kdb_backtrace+0x67
>>>> #1 0xffffffff805a1d26 at vpanic+0x186
>>>> #2 0xffffffff805a1b93 at panic+0x43
>>>> #3 0xffffffff805d71c6 at resource_list_release+0x1c6
>>>> #4 0xffffffff8040fef1 at mps_pci_free+0xe1
>>>> #5 0xffffffff8040fa23 at mps_pci_attach+0x1b3
>>>> #6 0xffffffff805d6594 at device_attach+0x3a4
>>>> #7 0xffffffff805d774d at bus_generic_attach+0x3d
>>>> #8 0xffffffff8044ac05 at pci_attach+0xd5
>>>> #9 0xffffffff805d6594 at device_attach+0x3a4
>>>> #10 0xffffffff805d774d at bus_generic_attach+0x3d
>>>> #11 0xffffffff80364761 at acpi_pcib_pci_attach+0xa1
>>>> #12 0xffffffff805d6594 at device_attach+0x3a4
>>>> #13 0xffffffff805d774d at bus_generic_attach+0x3d
>>>> #14 0xffffffff8044ac05 at pci_attach+0xd5
>>>> #15 0xffffffff805d6594 at device_attach+0x3a4
>>>> #16 0xffffffff805d774d at bus_generic_attach+0x3d
>>>> #17 0xffffffff80363e4d at acpi_pcib_acpi_attach+0x42d
>>>> Uptime: 1s
>> =E2=80=A6
>>=20
>>> Fixed in r321799, thanks for the report.
>> Fix confiremd; merged together with r321733 (and 321737) to 11.1 and
>> panic vanished.
>=20
> Late in the 11.2 phase, I identified this commit as a regression for =
MSI (non-x) alloctaion.
> I have an idea what probably causes the problem here (INTx allocation, =
although MSI (and MSI-x) capability):
> disable_msix is not 0 (I need to disable MSI-x because of =
ESXi-passthru=E2=80=A6).
>=20
> Corresponding lines:
> {
> device_t dev;
> int error, msgs;
>=20
> dev =3D sc->mps_dev;
> error =3D 0;
> msgs =3D 0;
>=20
> if ((sc->disable_msix =3D=3D 0) &&
> ((msgs =3D pci_msix_count(dev)) >=3D MPS_MSI_COUNT))
> error =3D mps_alloc_msix(sc, MPS_MSI_COUNT);
> if ((error !=3D 0) && (sc->disable_msi =3D=3D 0) &&
> ((msgs =3D pci_msi_count(dev)) >=3D MPS_MSI_COUNT))
> error =3D mps_alloc_msi(sc, MPS_MSI_COUNT);
> if (error !=3D 0)
> msgs =3D 0;
>=20
> sc->msi_msgs =3D msgs;
> return (error);
> }
>=20
> Before r321714, error was assigned ENXIO, which, if !=3D 0, could help =
make me understand the problem.
> Unfortunately I have no idea what ENXIO means, where it's defined and =
most important, how to find the place where the declaration/definition =
happens. Only joe and vi available here, any hints highly appreciated.
>=20
> I can confirm that MSI allocation works with =
mps.ko_21.02.00.00-fbsd-r321415 with my ESXi-passthru-non_msi-x setup.
> Although the dirver emits no message that an MSI was allocated, like =
toher drivers do. That's a cosmetic one though.
> But the MSI->INTx regression is a severe one for me, which I'd like to =
fix myself but I'm missing so many fundamental skills :-(
>=20
Hi Harry,
You are correct about the bug. Please change the line at the top of the =
function that reads
error =3D 0;
to
error =3D ENXIO;
Let me know if that fixes the MSI problem for you.
Scott
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?78611650-D7A4-4B1D-A254-DB058E1AC1C6>