Re: kern/52767: The wm driver cannot receive traffic on i82583V (rev. 0x00)

[Masanobu SAITOH <msaitoh%execsw.org@localhost>]

The following reply was made to PR kern/52767; it has been noted by GNATS.

From: Masanobu SAITOH <msaitoh%execsw.org@localhost>
To: gnats-bugs%NetBSD.org@localhost, kern-bug-people%netbsd.org@localhost,
 gnats-admin%netbsd.org@localhost, netbsd-bugs%netbsd.org@localhost, clare%csel.org@localhost
Cc: msaitoh%execsw.org@localhost, Hisashi T Fujinaka <htodd%twofifty.com@localhost>
Subject: Re: kern/52767: The wm driver cannot receive traffic on i82583V (rev.
 0x00)
Date: Tue, 28 Nov 2017 18:21:33 +0900

 Hi, all.
 
 On 2017/11/28 16:45, clare%csel.org@localhost wrote:
 > The following reply was made to PR kern/52767; it has been noted by GNATS.
 > 
 > From: clare%csel.org@localhost
 > To: gnats-bugs%NetBSD.org@localhost
 > Cc: kern-bug-people%netbsd.org@localhost, gnats-admin%netbsd.org@localhost,
 >   netbsd-bugs%netbsd.org@localhost
 > Subject: Re: kern/52767: The wm driver cannot receive traffic on i82583V
 >   (rev. 0x00)
 > Date: Tue, 28 Nov 2017 16:43:07 +0900
 > 
 >   >  >  Does it work on Linux?
 >   >  >
 >   >  >  I'd also check the NVM. That OUI isn't valid.
 >   >
 >   >  Fedora Linux live DVD runs on the machine looks okay.
 >   >  Linux localhost-live 4.11.8-300.fc26.x86_64 #1 SMP Thu Jun 29 20:09:48 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
 >   >
 >   >  Windows 10 is also okay.
 >   >  FreeBSD or pfSense is not tested yet.
 >   
 >   FreeBSD em driver works on the machine.
 >   The problem may be in the wm driver...
 >   
 >   pcib1: <ACPI PCI-PCI bridge> irq 16 at device 28.0 on pci0
 >   pci1: <ACPI PCI bus> on pcib1
 >   em0: <Intel(R) PRO/1000 Network Connection 7.6.1-k> port 0xe000-0xe01f mem 0xf7d00000-0xf7d1ffff,0xf7d20000-0xf7d23fff irq 16 at device 0.0 on pci1
 >   em0: Using an MSI interrupt
 >   em0: Ethernet address: 0c:e8:5c:**:**:**
 >   em0: netmap queues/slots: TX 1/1024, RX 1/1024
 >   
 
 My 82583V machine has an MSI-X cap:
 
 >   Capability register at 0xc8
 >     type: 0x01 (Power Management)
 >   Capability register at 0xd0
 >     type: 0x05 (MSI)
 >   Capability register at 0xe0
 >     type: 0x10 (PCI Express)
 >   Capability register at 0xa0
 >     type: 0x11 (MSI-X)
 (snip)
 >   MSI-X Capability Register
 >     Message Control register: 0x0000
 >       Table Size: 1
 >       Function Mask: off
 >       MSI-X Enable: off
 >     Table offset register: 0x00000003
 >       Table offset: 0x00000000
 >       BIR: 0x3
 >     Pending bit array register: 0x00002003
 >       Pending bit array offset: 0x00002000
 >       BIR: 0x3
 
 but the document doesn't say anything about MSI-X.
 The description in the document says 0xa0 is reserved.
 
 > Address Item			Next Pointer
 > 0xC8-CF PCI power management	0xD0
 > 0xD0-DF MSI			0xE0
 > 0xA0-AB Reserved		0x00   <======= this!
 > 0xE0-F3 PCIe Capabilities	0xA0
 
 And more, neither linux nor FreeBSD don't use MSI-X.
 
 Could you test the following patch?
 
 -------------------------
 Index: if_wm.c
 ===================================================================
 RCS file: /cvsroot/src/sys/dev/pci/if_wm.c,v
 retrieving revision 1.544
 diff -u -p -r1.544 if_wm.c
 --- if_wm.c	22 Nov 2017 02:36:52 -0000	1.544
 +++ if_wm.c	28 Nov 2017 09:19:54 -0000
 @@ -1853,7 +1853,13 @@ wm_attach(device_t parent, device_t self
   
   	/* Allocation settings */
   	max_type = PCI_INTR_TYPE_MSIX;
 -	counts[PCI_INTR_TYPE_MSIX] = sc->sc_nqueues + 1;
 +	/*
 +	 * 82583 has a MSI-X capability in the PCI configuration space but
 +	 * it doesn't support it. At least the document doesn't say anything
 +	 * about MSI-X.
 +	 */
 +	counts[PCI_INTR_TYPE_MSIX]
 +	    = (sc->sc_type == WM_T_82583) ? 0 : sc->sc_nqueues + 1;
   	counts[PCI_INTR_TYPE_MSI] = 1;
   	counts[PCI_INTR_TYPE_INTX] = 1;
   	/* overridden by disable flags */
 -------------------------
 
 -- 
 -----------------------------------------------
                  SAITOH Masanobu (msaitoh%execsw.org@localhost
                                   msaitoh%netbsd.org@localhost)