Re: panic on mfii(4) vd removal

[Edgar Fuß]

> I get a panic if I remove a virtual disk from an mfii(4) device.
That's another blunder in mfii(4). Patch (including the last) attached.
Index: sys/dev/pci/mfii.c
===
RCS file: /cvsroot/src/sys/dev/pci/mfii.c,v
retrieving revision 1.3.2.7
diff -u -p -r1.3.2.7 mfii.c
--- sys/dev/pci/mfii.c  29 Sep 2022 14:41:43 -  1.3.2.7
+++ sys/dev/pci/mfii.c  22 Sep 2023 12:16:42 -
@@ -503,6 +503,8 @@ static const char *mfi_bbu_indicators[] 
 };
 #endif
 
+#define MFI_BBU_SENSORS 4
+
 static voidmfii_init_ld_sensor(struct mfii_softc *, envsys_data_t *, int);
 static voidmfii_refresh_ld_sensor(struct mfii_softc *, envsys_data_t *);
 static voidmfii_attach_sensor(struct mfii_softc *, envsys_data_t *);
@@ -1373,18 +1375,20 @@ mfii_aen_ld_update(struct mfii_softc *sc
if (old == -1 && nld != -1) {
printf("%s: logical drive %d added (target %d)\n",
DEVNAME(sc), i, nld);
+   sc->sc_ld[i].ld_present = 1;
 
// XXX scsi_probe_target(sc->sc_scsibus, i);
 
-   mfii_init_ld_sensor(sc, >sc_sensors[i], i);
-   mfii_attach_sensor(sc, >sc_sensors[i]);
+   mfii_init_ld_sensor(sc, >sc_sensors[i + 
MFI_BBU_SENSORS], i);
+   mfii_attach_sensor(sc, >sc_sensors[i + 
MFI_BBU_SENSORS]);
} else if (nld == -1 && old != -1) {
printf("%s: logical drive %d removed (target %d)\n",
DEVNAME(sc), i, old);
+   sc->sc_ld[i].ld_present = 0;
 
scsipi_target_detach(>sc_chan, i, 0, DETACH_FORCE);
sysmon_envsys_sensor_detach(sc->sc_sme,
-   >sc_sensors[i]);
+   >sc_sensors[i + MFI_BBU_SENSORS]);
}
}
 
@@ -3716,8 +3720,6 @@ freeme:
 
 #endif /* NBIO > 0 */
 
-#define MFI_BBU_SENSORS 4
-
 static void
 mfii_bbu(struct mfii_softc *sc, envsys_data_t *edata)
 {

Re: Adding a virtual disk to mpii(4)

[Christos Zoulas]

In article ,
Edgar Fuß   wrote:
>-=-=-=-=-=-
>
>After adding a virtual disk to an mfii(4) device (racadm createvirtualdisk, 
>in my case), you get a nice
>   mfii0: logical drive 2 added (target 2)
>message, but
>   scsictl scsibus0 scan 2 0
>doesn't find any drive.
>
>That's because sc_ld[i].ld_present is still unset from mfii_attach() and so 
>mfii_scsipi_request() will return early (with xs->error = XS_SELTIMEOUT).
>
>The attached fix seems pretty dammn obvious and appears to work.
>
>Shall I file a PR?

Committed, thanks!

christos