linux-kernel - Re: [PATCH 1/7] scsi/bfa: use pcie_capability

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAErSpo4d2s06Lb8U9TaBR+Niq0EaT8zK5dRaj=WSV3PpMpM0pQ@mail.gmail.com>
Date:	Tue, 3 Sep 2013 20:37:54 -0600
From:	Bjorn Helgaas <bhelgaas@...gle.com>
To:	Yijing Wang <wangyijing@...wei.com>
Cc:	Benjamin Herrenschmidt <benh@...nel.crashing.org>,
	Gavin Shan <shangw@...ux.vnet.ibm.com>,
	"James E.J. Bottomley" <JBottomley@...allels.com>,
	"David S. Miller" <davem@...emloft.net>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
	Hanjun Guo <guohanjun@...wei.com>,
	Jiang Liu <jiang.liu@...wei.com>,
	Anil Gurumurthy <agurumur@...cade.com>,
	Vijaya Mohan Guvva <vmohan@...cade.com>,
	"linux-scsi@...r.kernel.org" <linux-scsi@...r.kernel.org>
Subject: Re: [PATCH 1/7] scsi/bfa: use pcie_capability_xxx to simplify code

On Tue, Sep 3, 2013 at 5:34 PM, Bjorn Helgaas <bhelgaas@...gle.com> wrote:
> On Tue, Sep 03, 2013 at 03:35:09PM +0800, Yijing Wang wrote:
>> Pcie_capability_xxx() interfaces were introudced to
>> simplify code to access PCIe Cap config space. And
>> because PCI core saves the PCIe Cap offset in
>> set_pcie_port_type() when device is enumerated.
>> So we can use pci_is_pcie() instead.
>>
>> Signed-off-by: Yijing Wang <wangyijing@...wei.com>
>> Cc: Jiang Liu <jiang.liu@...wei.com>
>> Cc: Anil Gurumurthy <agurumur@...cade.com>
>> Cc: Vijaya Mohan Guvva <vmohan@...cade.com>
>> Cc: "James E.J. Bottomley" <JBottomley@...allels.com>
>> Cc: linux-scsi@...r.kernel.org
>> Cc: linux-kernel@...r.kernel.org
>> ---
>>  drivers/scsi/bfa/bfad.c |    9 +++------
>>  1 files changed, 3 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/scsi/bfa/bfad.c b/drivers/scsi/bfa/bfad.c
>> index 9611195..d726b81 100644
>> --- a/drivers/scsi/bfa/bfad.c
>> +++ b/drivers/scsi/bfa/bfad.c
>> @@ -767,7 +767,6 @@ bfad_pci_init(struct pci_dev *pdev, struct bfad_s *bfad)
>>
>>       /* Adjust PCIe Maximum Read Request Size */
>>       if (pcie_max_read_reqsz > 0) {
>> -             int pcie_cap_reg;
>>               u16 pcie_dev_ctl;
>>               u16 mask = 0xffff;
>>
>> @@ -794,10 +793,8 @@ bfad_pci_init(struct pci_dev *pdev, struct bfad_s *bfad)
>>                       break;
>>               }
>>
>> -             pcie_cap_reg = pci_find_capability(pdev, PCI_CAP_ID_EXP);
>> -             if (mask != 0xffff && pcie_cap_reg) {
>> -                     pcie_cap_reg += 0x08;
>> -                     pci_read_config_word(pdev, pcie_cap_reg, &pcie_dev_ctl);
>> +             if (mask != 0xffff && pci_is_pcie(pdev)) {
>
> Please move the pci_is_pcie() test up to the
> "if (pcie_mas_read_reqsz ..." statement.  There's no point in doing
> the switch statement if this isn't a PCIe device.
>
>> +                     pcie_capability_read_word(pdev, PCI_EXP_DEVCTL, &pcie_dev_ctl);
>>                       if ((pcie_dev_ctl & 0x7000) != mask) {

I forgot to mention that this 0x7000 constant should be
PCI_EXP_DEVCTL_READRQ instead.

>>                               printk(KERN_WARNING "BFA[%s]: "
>>                               "pcie_max_read_request_size is %d, "
>> @@ -806,7 +803,7 @@ bfad_pci_init(struct pci_dev *pdev, struct bfad_s *bfad)
>>                               pcie_max_read_reqsz);
>>
>>                               pcie_dev_ctl &= ~0x7000;
>> -                             pci_write_config_word(pdev, pcie_cap_reg,
>> +                             pcie_capability_write_word(pdev, PCI_EXP_DEVCTL,
>>                                               pcie_dev_ctl | mask);
>
> Please rework this to use pcie_set_readrq() instead of writing
> the capability directly.  If we write the capability directly, we
> risk writing a value that is incompatible with the MPS
> configuration done by the PCI core.
>
>>                       }
>>               }
>> --
>> 1.7.1
>>
>>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/