mpt fusion: Avoid racing when mptsas and mptcl module are loaded in parallel
Subsystems (2)
PATCH 3/3 mpt fusion: Avoid racing when mptsas and mptcl module are loaded in parallel
This patch sets the IOC pointer in drvrdata of pcidev before adding the IOC into the list of IOCs. Without this patch the driver oops when the mptsas and mptctl modules are loaded in parallel.
Signed-off-by Sathya Prakash <[email protected]>
---
drivers/message/fusion/mptbase.c +3 -1
@@ -1704,6 +1704,9 @@ mpt_attach(struct pci_dev *pdev, const struct pci_device_id *id) ioc->active =0 ;CHIPREG_WRITE32 (&ioc->chip->IntStatus,0 ); +/* Set IOC ptr in the pcidev's driver data. */ +pci_set_drvdata (ioc->pcidev, ioc); +/* Set lookup ptr. */ list_add_tail (&ioc->list, &ioc_list);
@@ -2074,7 +2077,6 @@ mpt_do_ioc_recovery(MPT_ADAPTER *ioc, u32 reason, int sleepFlag) irq_allocated =1 ; ioc->pci_irq = ioc->pcidev->irq;pci_set_master (ioc->pcidev);/* ?? */ -pci_set_drvdata (ioc->pcidev, ioc);dprintk (ioc,printk (MYIOC_s_INFO_FMT"installed at interrupt " "%d\n" , ioc->name, ioc->pcidev->irq)); }
James Bottomley
maintainerQuoted 3 lines · Sathya
This patch sets the IOC pointer in drvrdata of pcidev before adding the IOC into the list of IOCs. Without this patch the driver oops when the mptsas and mptctl modules are loaded in parallel.
Much better, thanks.
Would I be correct in assuming that even though this email says [PATCH 3/3] there aren't two missing precursor patches?
James