This patch series contains fixes and improvements to pcie iproc driver.
This patch set is based on Linux-5.5-rc1.
Bharat Gooty (1):
PCI: iproc: fix out of bound array access
Roman Bacik (1):
PCI: iproc: fix invalidating PAXB address mapping
Srinath Mannam (1):
PCI: iproc: Display PCIe Link information
drivers/pci/controller/pcie-iproc.c | 23 +++++++++++++++++------
1 file changed, 17 insertions(+), 6 deletions(-)
--
2.7.4
From: Bharat Gooty <[email protected]>
Declare the full size array for all revisions of PAX register sets
to avoid potentially out of bound access of the register array
when they are being initialized in the 'iproc_pcie_rev_init'
function.
Fixes: 06324ede76cdf ("PCI: iproc: Improve core register population")
Signed-off-by: Bharat Gooty <[email protected]>
---
drivers/pci/controller/pcie-iproc.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/pci/controller/pcie-iproc.c b/drivers/pci/controller/pcie-iproc.c
index 0a468c7..6972ca4 100644
--- a/drivers/pci/controller/pcie-iproc.c
+++ b/drivers/pci/controller/pcie-iproc.c
@@ -307,7 +307,7 @@ enum iproc_pcie_reg {
};
/* iProc PCIe PAXB BCMA registers */
-static const u16 iproc_pcie_reg_paxb_bcma[] = {
+static const u16 iproc_pcie_reg_paxb_bcma[IPROC_PCIE_MAX_NUM_REG] = {
[IPROC_PCIE_CLK_CTRL] = 0x000,
[IPROC_PCIE_CFG_IND_ADDR] = 0x120,
[IPROC_PCIE_CFG_IND_DATA] = 0x124,
@@ -318,7 +318,7 @@ static const u16 iproc_pcie_reg_paxb_bcma[] = {
};
/* iProc PCIe PAXB registers */
-static const u16 iproc_pcie_reg_paxb[] = {
+static const u16 iproc_pcie_reg_paxb[IPROC_PCIE_MAX_NUM_REG] = {
[IPROC_PCIE_CLK_CTRL] = 0x000,
[IPROC_PCIE_CFG_IND_ADDR] = 0x120,
[IPROC_PCIE_CFG_IND_DATA] = 0x124,
@@ -334,7 +334,7 @@ static const u16 iproc_pcie_reg_paxb[] = {
};
/* iProc PCIe PAXB v2 registers */
-static const u16 iproc_pcie_reg_paxb_v2[] = {
+static const u16 iproc_pcie_reg_paxb_v2[IPROC_PCIE_MAX_NUM_REG] = {
[IPROC_PCIE_CLK_CTRL] = 0x000,
[IPROC_PCIE_CFG_IND_ADDR] = 0x120,
[IPROC_PCIE_CFG_IND_DATA] = 0x124,
@@ -363,7 +363,7 @@ static const u16 iproc_pcie_reg_paxb_v2[] = {
};
/* iProc PCIe PAXC v1 registers */
-static const u16 iproc_pcie_reg_paxc[] = {
+static const u16 iproc_pcie_reg_paxc[IPROC_PCIE_MAX_NUM_REG] = {
[IPROC_PCIE_CLK_CTRL] = 0x000,
[IPROC_PCIE_CFG_IND_ADDR] = 0x1f0,
[IPROC_PCIE_CFG_IND_DATA] = 0x1f4,
@@ -372,7 +372,7 @@ static const u16 iproc_pcie_reg_paxc[] = {
};
/* iProc PCIe PAXC v2 registers */
-static const u16 iproc_pcie_reg_paxc_v2[] = {
+static const u16 iproc_pcie_reg_paxc_v2[IPROC_PCIE_MAX_NUM_REG] = {
[IPROC_PCIE_MSI_GIC_MODE] = 0x050,
[IPROC_PCIE_MSI_BASE_ADDR] = 0x074,
[IPROC_PCIE_MSI_WINDOW_SIZE] = 0x078,
--
2.7.4
From: Roman Bacik <[email protected]>
Second stage bootloader prior to Linux boot may use all inbound windows
including IARR1/IMAP1. We need to ensure all previous configuration of
inbound windows are invalidated during the initialization stage of the
Linux iProc PCIe driver. Add fix to invalidate IARR1/IMAP1 because it was
missed in previous patch.
Fixes: 9415743e4c8a ("PCI: iproc: Invalidate PAXB address mapping")
Signed-off-by: Roman Bacik <[email protected]>
---
drivers/pci/controller/pcie-iproc.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/pci/controller/pcie-iproc.c b/drivers/pci/controller/pcie-iproc.c
index 6972ca4..e7f0d58 100644
--- a/drivers/pci/controller/pcie-iproc.c
+++ b/drivers/pci/controller/pcie-iproc.c
@@ -351,6 +351,8 @@ static const u16 iproc_pcie_reg_paxb_v2[IPROC_PCIE_MAX_NUM_REG] = {
[IPROC_PCIE_OMAP3] = 0xdf8,
[IPROC_PCIE_IARR0] = 0xd00,
[IPROC_PCIE_IMAP0] = 0xc00,
+ [IPROC_PCIE_IARR1] = 0xd08,
+ [IPROC_PCIE_IMAP1] = 0xd70,
[IPROC_PCIE_IARR2] = 0xd10,
[IPROC_PCIE_IMAP2] = 0xcc0,
[IPROC_PCIE_IARR3] = 0xe00,
--
2.7.4
Add more comprehensive information to show PCIe link speed and link
width to the console.
Signed-off-by: Srinath Mannam <[email protected]>
---
drivers/pci/controller/pcie-iproc.c | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/drivers/pci/controller/pcie-iproc.c b/drivers/pci/controller/pcie-iproc.c
index e7f0d58..ed41357 100644
--- a/drivers/pci/controller/pcie-iproc.c
+++ b/drivers/pci/controller/pcie-iproc.c
@@ -823,6 +823,8 @@ static int iproc_pcie_check_link(struct iproc_pcie *pcie)
#define PCI_TARGET_LINK_SPEED_MASK 0xf
#define PCI_TARGET_LINK_SPEED_GEN2 0x2
#define PCI_TARGET_LINK_SPEED_GEN1 0x1
+#define PCI_TARGET_LINK_WIDTH_MASK 0x3f
+#define PCI_TARGET_LINK_WIDTH_OFFSET 0x4
iproc_pci_raw_config_read32(pcie, 0,
IPROC_PCI_EXP_CAP + PCI_EXP_LNKCTL2,
4, &link_ctrl);
@@ -843,7 +845,14 @@ static int iproc_pcie_check_link(struct iproc_pcie *pcie)
}
}
- dev_info(dev, "link: %s\n", link_is_active ? "UP" : "DOWN");
+ if (link_is_active) {
+ dev_info(dev, "link UP @ Speed Gen-%d and width-x%d\n",
+ link_status & PCI_TARGET_LINK_SPEED_MASK,
+ (link_status >> PCI_TARGET_LINK_WIDTH_OFFSET) &
+ PCI_TARGET_LINK_WIDTH_MASK);
+ } else {
+ dev_info(dev, "link DOWN\n");
+ }
return link_is_active ? 0 : -ENODEV;
}
--
2.7.4
On Thu, Mar 26, 2020 at 12:37:27PM +0530, Srinath Mannam wrote:
> Add more comprehensive information to show PCIe link speed and link
> width to the console.
>
> Signed-off-by: Srinath Mannam <[email protected]>
> ---
> drivers/pci/controller/pcie-iproc.c | 11 ++++++++++-
> 1 file changed, 10 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/pci/controller/pcie-iproc.c b/drivers/pci/controller/pcie-iproc.c
> index e7f0d58..ed41357 100644
> --- a/drivers/pci/controller/pcie-iproc.c
> +++ b/drivers/pci/controller/pcie-iproc.c
> @@ -823,6 +823,8 @@ static int iproc_pcie_check_link(struct iproc_pcie *pcie)
> #define PCI_TARGET_LINK_SPEED_MASK 0xf
> #define PCI_TARGET_LINK_SPEED_GEN2 0x2
> #define PCI_TARGET_LINK_SPEED_GEN1 0x1
> +#define PCI_TARGET_LINK_WIDTH_MASK 0x3f
> +#define PCI_TARGET_LINK_WIDTH_OFFSET 0x4
> iproc_pci_raw_config_read32(pcie, 0,
> IPROC_PCI_EXP_CAP + PCI_EXP_LNKCTL2,
> 4, &link_ctrl);
> @@ -843,7 +845,14 @@ static int iproc_pcie_check_link(struct iproc_pcie *pcie)
> }
> }
>
> - dev_info(dev, "link: %s\n", link_is_active ? "UP" : "DOWN");
> + if (link_is_active) {
> + dev_info(dev, "link UP @ Speed Gen-%d and width-x%d\n",
> + link_status & PCI_TARGET_LINK_SPEED_MASK,
> + (link_status >> PCI_TARGET_LINK_WIDTH_OFFSET) &
> + PCI_TARGET_LINK_WIDTH_MASK);
Can you use pcie_print_link_status() or some variant here instead of
rolling your own?
> + } else {
> + dev_info(dev, "link DOWN\n");
> + }
>
> return link_is_active ? 0 : -ENODEV;
> }
> --
> 2.7.4
>
[fixed Andrew's email addr]
Change subject to match convention, e.g.,
PCI: iproc: Invalidate correct PAXB inbound windows
On Thu, Mar 26, 2020 at 12:37:26PM +0530, Srinath Mannam wrote:
> From: Roman Bacik <[email protected]>
>
> Second stage bootloader prior to Linux boot may use all inbound windows
> including IARR1/IMAP1. We need to ensure all previous configuration of
> inbound windows are invalidated during the initialization stage of the
> Linux iProc PCIe driver. Add fix to invalidate IARR1/IMAP1 because it was
> missed in previous patch.
>
> Fixes: 9415743e4c8a ("PCI: iproc: Invalidate PAXB address mapping")
> Signed-off-by: Roman Bacik <[email protected]>
> ---
> drivers/pci/controller/pcie-iproc.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/pci/controller/pcie-iproc.c b/drivers/pci/controller/pcie-iproc.c
> index 6972ca4..e7f0d58 100644
> --- a/drivers/pci/controller/pcie-iproc.c
> +++ b/drivers/pci/controller/pcie-iproc.c
> @@ -351,6 +351,8 @@ static const u16 iproc_pcie_reg_paxb_v2[IPROC_PCIE_MAX_NUM_REG] = {
> [IPROC_PCIE_OMAP3] = 0xdf8,
> [IPROC_PCIE_IARR0] = 0xd00,
> [IPROC_PCIE_IMAP0] = 0xc00,
> + [IPROC_PCIE_IARR1] = 0xd08,
> + [IPROC_PCIE_IMAP1] = 0xd70,
Wow, this is a little too subtle for my taste.
9415743e4c8a added this loop:
+ for (idx = 0; idx < ib->nr_regions; idx++) {
+ iproc_pcie_write_reg(pcie,
+ MAP_REG(IPROC_PCIE_IARR0, idx), 0);
This patch doesn't change the loop or the limit, so I *guess*
previously we invalidated IPROC_PCIE_IARR0, IPROC_PCIE_IARR2, ...,
(skipping IPROC_PCIE_IARR1), and now we will invalidate
IPROC_PCIE_IARR0, IPROC_PCIE_IARR1, ... instead?
> [IPROC_PCIE_IARR2] = 0xd10,
> [IPROC_PCIE_IMAP2] = 0xcc0,
> [IPROC_PCIE_IARR3] = 0xe00,
On Thu, Mar 26, 2020 at 12:37:26PM +0530, Srinath Mannam wrote:
> From: Roman Bacik <[email protected]>
>
> Second stage bootloader prior to Linux boot may use all inbound windows
> including IARR1/IMAP1. We need to ensure all previous configuration of
> inbound windows are invalidated during the initialization stage of the
> Linux iProc PCIe driver. Add fix to invalidate IARR1/IMAP1 because it was
> missed in previous patch.
>
> Fixes: 9415743e4c8a ("PCI: iproc: Invalidate PAXB address mapping")
> Signed-off-by: Roman Bacik <[email protected]>
> ---
> drivers/pci/controller/pcie-iproc.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/pci/controller/pcie-iproc.c b/drivers/pci/controller/pcie-iproc.c
> index 6972ca4..e7f0d58 100644
> --- a/drivers/pci/controller/pcie-iproc.c
> +++ b/drivers/pci/controller/pcie-iproc.c
> @@ -351,6 +351,8 @@ static const u16 iproc_pcie_reg_paxb_v2[IPROC_PCIE_MAX_NUM_REG] = {
> [IPROC_PCIE_OMAP3] = 0xdf8,
> [IPROC_PCIE_IARR0] = 0xd00,
> [IPROC_PCIE_IMAP0] = 0xc00,
> + [IPROC_PCIE_IARR1] = 0xd08,
> + [IPROC_PCIE_IMAP1] = 0xd70,
And paxb_v2_ib_map[] has a comment saying "IARR1/IMAP1 (currently
unused)". Is that comment now wrong?
> [IPROC_PCIE_IARR2] = 0xd10,
> [IPROC_PCIE_IMAP2] = 0xcc0,
> [IPROC_PCIE_IARR3] = 0xe00,
> --
> 2.7.4
>
On Thu, Mar 26, 2020 at 8:33 AM Bjorn Helgaas <[email protected]> wrote:
>
> On Thu, Mar 26, 2020 at 12:37:26PM +0530, Srinath Mannam wrote:
> > From: Roman Bacik <[email protected]>
> >
> > Second stage bootloader prior to Linux boot may use all inbound windows
> > including IARR1/IMAP1. We need to ensure all previous configuration of
> > inbound windows are invalidated during the initialization stage of the
> > Linux iProc PCIe driver. Add fix to invalidate IARR1/IMAP1 because it was
> > missed in previous patch.
> >
> > Fixes: 9415743e4c8a ("PCI: iproc: Invalidate PAXB address mapping")
> > Signed-off-by: Roman Bacik <[email protected]>
> > ---
> > drivers/pci/controller/pcie-iproc.c | 2 ++
> > 1 file changed, 2 insertions(+)
> >
> > diff --git a/drivers/pci/controller/pcie-iproc.c b/drivers/pci/controller/pcie-iproc.c
> > index 6972ca4..e7f0d58 100644
> > --- a/drivers/pci/controller/pcie-iproc.c
> > +++ b/drivers/pci/controller/pcie-iproc.c
> > @@ -351,6 +351,8 @@ static const u16 iproc_pcie_reg_paxb_v2[IPROC_PCIE_MAX_NUM_REG] = {
> > [IPROC_PCIE_OMAP3] = 0xdf8,
> > [IPROC_PCIE_IARR0] = 0xd00,
> > [IPROC_PCIE_IMAP0] = 0xc00,
> > + [IPROC_PCIE_IARR1] = 0xd08,
> > + [IPROC_PCIE_IMAP1] = 0xd70,
>
> And paxb_v2_ib_map[] has a comment saying "IARR1/IMAP1 (currently
> unused)". Is that comment now wrong?
>
The comment is still correct, IARR1/IMAP1 is unused in Linux. But it
may need to be invalidated in case it was modified by bootloaders.
> > [IPROC_PCIE_IARR2] = 0xd10,
> > [IPROC_PCIE_IMAP2] = 0xcc0,
> > [IPROC_PCIE_IARR3] = 0xe00,
> > --
> > 2.7.4
> >
Change subject to match convention, e.g.,
PCI: iproc: Fix out-of-bound array accesses
On Thu, Mar 26, 2020 at 12:37:25PM +0530, Srinath Mannam wrote:
> From: Bharat Gooty <[email protected]>
>
> Declare the full size array for all revisions of PAX register sets
> to avoid potentially out of bound access of the register array
> when they are being initialized in the 'iproc_pcie_rev_init'
> function.
s/the 'iproc_pcie_rev_init' function/iproc_pcie_rev_init()/
It's outside the scope of this patch, but I'm not really a fan of the
pcie->reg_offsets[] scheme this driver uses to deal with these
differences. There usually seems to be *something* that keeps the
driver from referencing registers that don't exist, but it doesn't
seem like the mechanism is very consistent or robust:
- IPROC_PCIE_LINK_STATUS is implemented by PAXB but not PAXC.
iproc_pcie_check_link() avoids using it if "ep_is_internal", which
is set for PAXC and PAXC_V2. Not an obvious connection.
- IPROC_PCIE_CLK_CTRL is implemented for PAXB and PAXC_V1, but not
PAXC_V2. iproc_pcie_perst_ctrl() avoids using it ep_is_internal",
so it *doesn't* use it for PAXC_V1, which does implement it.
Maybe a bug, maybe intentional; I can't tell.
- IPROC_PCIE_INTX_EN is only implemented by PAXB (not PAXC), but
AFAICT, we always call iproc_pcie_enable() and rely on
iproc_pcie_write_reg() silently drop the write to it on PAXC.
- IPROC_PCIE_OARR0 is implemented by PAXB and PAXB_V2 and used by
iproc_pcie_map_ranges(), which is called if "need_ob_cfg", which
is set if there's a "brcm,pcie-ob" DT property. No clear
connection to PAXB.
I think it would be more readable if we used a single variant
identifier consistently, e.g., the "pcie->type" already used in
iproc_pcie_msi_steer(), or maybe a set of variant-specific function
pointers as pcie-qcom.c does.
> Fixes: 06324ede76cdf ("PCI: iproc: Improve core register population")
> Signed-off-by: Bharat Gooty <[email protected]>
> ---
> drivers/pci/controller/pcie-iproc.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/pci/controller/pcie-iproc.c b/drivers/pci/controller/pcie-iproc.c
> index 0a468c7..6972ca4 100644
> --- a/drivers/pci/controller/pcie-iproc.c
> +++ b/drivers/pci/controller/pcie-iproc.c
> @@ -307,7 +307,7 @@ enum iproc_pcie_reg {
> };
>
> /* iProc PCIe PAXB BCMA registers */
> -static const u16 iproc_pcie_reg_paxb_bcma[] = {
> +static const u16 iproc_pcie_reg_paxb_bcma[IPROC_PCIE_MAX_NUM_REG] = {
> [IPROC_PCIE_CLK_CTRL] = 0x000,
> [IPROC_PCIE_CFG_IND_ADDR] = 0x120,
> [IPROC_PCIE_CFG_IND_DATA] = 0x124,
> @@ -318,7 +318,7 @@ static const u16 iproc_pcie_reg_paxb_bcma[] = {
> };
>
> /* iProc PCIe PAXB registers */
> -static const u16 iproc_pcie_reg_paxb[] = {
> +static const u16 iproc_pcie_reg_paxb[IPROC_PCIE_MAX_NUM_REG] = {
> [IPROC_PCIE_CLK_CTRL] = 0x000,
> [IPROC_PCIE_CFG_IND_ADDR] = 0x120,
> [IPROC_PCIE_CFG_IND_DATA] = 0x124,
> @@ -334,7 +334,7 @@ static const u16 iproc_pcie_reg_paxb[] = {
> };
>
> /* iProc PCIe PAXB v2 registers */
> -static const u16 iproc_pcie_reg_paxb_v2[] = {
> +static const u16 iproc_pcie_reg_paxb_v2[IPROC_PCIE_MAX_NUM_REG] = {
> [IPROC_PCIE_CLK_CTRL] = 0x000,
> [IPROC_PCIE_CFG_IND_ADDR] = 0x120,
> [IPROC_PCIE_CFG_IND_DATA] = 0x124,
> @@ -363,7 +363,7 @@ static const u16 iproc_pcie_reg_paxb_v2[] = {
> };
>
> /* iProc PCIe PAXC v1 registers */
> -static const u16 iproc_pcie_reg_paxc[] = {
> +static const u16 iproc_pcie_reg_paxc[IPROC_PCIE_MAX_NUM_REG] = {
> [IPROC_PCIE_CLK_CTRL] = 0x000,
> [IPROC_PCIE_CFG_IND_ADDR] = 0x1f0,
> [IPROC_PCIE_CFG_IND_DATA] = 0x1f4,
> @@ -372,7 +372,7 @@ static const u16 iproc_pcie_reg_paxc[] = {
> };
>
> /* iProc PCIe PAXC v2 registers */
> -static const u16 iproc_pcie_reg_paxc_v2[] = {
> +static const u16 iproc_pcie_reg_paxc_v2[IPROC_PCIE_MAX_NUM_REG] = {
> [IPROC_PCIE_MSI_GIC_MODE] = 0x050,
> [IPROC_PCIE_MSI_BASE_ADDR] = 0x074,
> [IPROC_PCIE_MSI_WINDOW_SIZE] = 0x078,
> --
> 2.7.4
>
Hi Bjorn,
On 3/26/2020 12:48 PM, Bjorn Helgaas wrote:
> Change subject to match convention, e.g.,
>
> PCI: iproc: Fix out-of-bound array accesses
>
> On Thu, Mar 26, 2020 at 12:37:25PM +0530, Srinath Mannam wrote:
>> From: Bharat Gooty <[email protected]>
>>
>> Declare the full size array for all revisions of PAX register sets
>> to avoid potentially out of bound access of the register array
>> when they are being initialized in the 'iproc_pcie_rev_init'
>> function.
>
> s/the 'iproc_pcie_rev_init' function/iproc_pcie_rev_init()/
>
> It's outside the scope of this patch, but I'm not really a fan of the
> pcie->reg_offsets[] scheme this driver uses to deal with these
> differences. There usually seems to be *something* that keeps the
> driver from referencing registers that don't exist, but it doesn't
> seem like the mechanism is very consistent or robust:
>
> - IPROC_PCIE_LINK_STATUS is implemented by PAXB but not PAXC.
> iproc_pcie_check_link() avoids using it if "ep_is_internal", which
> is set for PAXC and PAXC_V2. Not an obvious connection.
>
> - IPROC_PCIE_CLK_CTRL is implemented for PAXB and PAXC_V1, but not
> PAXC_V2. iproc_pcie_perst_ctrl() avoids using it ep_is_internal",
> so it *doesn't* use it for PAXC_V1, which does implement it.
> Maybe a bug, maybe intentional; I can't tell.
>
> - IPROC_PCIE_INTX_EN is only implemented by PAXB (not PAXC), but
> AFAICT, we always call iproc_pcie_enable() and rely on
> iproc_pcie_write_reg() silently drop the write to it on PAXC.
>
> - IPROC_PCIE_OARR0 is implemented by PAXB and PAXB_V2 and used by
> iproc_pcie_map_ranges(), which is called if "need_ob_cfg", which
> is set if there's a "brcm,pcie-ob" DT property. No clear
> connection to PAXB.
>
> I think it would be more readable if we used a single variant
> identifier consistently, e.g., the "pcie->type" already used in
> iproc_pcie_msi_steer(), or maybe a set of variant-specific function
> pointers as pcie-qcom.c does.
>
It is not possible to use a single variant identifier consistently,
i.e., 'pcie->type'. Many of these features are controller revision
specific, and certain revisions of the controllers may all have a
certain feature, while other revisions of the controllers do not. In
addition, there are overlap in features across different controllers.
IMO, it makes sense to have feature specific flags or booleans, and have
those features enabled or disabled based on 'pcie->type', which is what
the current driver does, but like you pointed out, what the driver
failed is to do this consistently.
The IPROC_PCIE_INTX_EN example you pointed out is a good example. I
agree with you that we shouldn't rely on iproc_pcie_write_reg to
silently drop the operation for PAXC. We should add code to make it
explictly obvious that legacy interrupt is not supported in all PAXC
controllers.
pcie->pcie->reg_offsets[] scheme was not intended to be used to silently
drop register access that are activated based on features. It's a
mistake that should be fixed if some code in the driver is done that
way, as you pointed out. The intention of reg_offsets[] is to allow many
of the code in this driver be made generic, and shared between different
revisions of the driver.
Thanks,
Ray
>> Fixes: 06324ede76cdf ("PCI: iproc: Improve core register population")
>> Signed-off-by: Bharat Gooty <[email protected]>
>> ---
>> drivers/pci/controller/pcie-iproc.c | 10 +++++-----
>> 1 file changed, 5 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/pci/controller/pcie-iproc.c b/drivers/pci/controller/pcie-iproc.c
>> index 0a468c7..6972ca4 100644
>> --- a/drivers/pci/controller/pcie-iproc.c
>> +++ b/drivers/pci/controller/pcie-iproc.c
>> @@ -307,7 +307,7 @@ enum iproc_pcie_reg {
>> };
>>
>> /* iProc PCIe PAXB BCMA registers */
>> -static const u16 iproc_pcie_reg_paxb_bcma[] = {
>> +static const u16 iproc_pcie_reg_paxb_bcma[IPROC_PCIE_MAX_NUM_REG] = {
>> [IPROC_PCIE_CLK_CTRL] = 0x000,
>> [IPROC_PCIE_CFG_IND_ADDR] = 0x120,
>> [IPROC_PCIE_CFG_IND_DATA] = 0x124,
>> @@ -318,7 +318,7 @@ static const u16 iproc_pcie_reg_paxb_bcma[] = {
>> };
>>
>> /* iProc PCIe PAXB registers */
>> -static const u16 iproc_pcie_reg_paxb[] = {
>> +static const u16 iproc_pcie_reg_paxb[IPROC_PCIE_MAX_NUM_REG] = {
>> [IPROC_PCIE_CLK_CTRL] = 0x000,
>> [IPROC_PCIE_CFG_IND_ADDR] = 0x120,
>> [IPROC_PCIE_CFG_IND_DATA] = 0x124,
>> @@ -334,7 +334,7 @@ static const u16 iproc_pcie_reg_paxb[] = {
>> };
>>
>> /* iProc PCIe PAXB v2 registers */
>> -static const u16 iproc_pcie_reg_paxb_v2[] = {
>> +static const u16 iproc_pcie_reg_paxb_v2[IPROC_PCIE_MAX_NUM_REG] = {
>> [IPROC_PCIE_CLK_CTRL] = 0x000,
>> [IPROC_PCIE_CFG_IND_ADDR] = 0x120,
>> [IPROC_PCIE_CFG_IND_DATA] = 0x124,
>> @@ -363,7 +363,7 @@ static const u16 iproc_pcie_reg_paxb_v2[] = {
>> };
>>
>> /* iProc PCIe PAXC v1 registers */
>> -static const u16 iproc_pcie_reg_paxc[] = {
>> +static const u16 iproc_pcie_reg_paxc[IPROC_PCIE_MAX_NUM_REG] = {
>> [IPROC_PCIE_CLK_CTRL] = 0x000,
>> [IPROC_PCIE_CFG_IND_ADDR] = 0x1f0,
>> [IPROC_PCIE_CFG_IND_DATA] = 0x1f4,
>> @@ -372,7 +372,7 @@ static const u16 iproc_pcie_reg_paxc[] = {
>> };
>>
>> /* iProc PCIe PAXC v2 registers */
>> -static const u16 iproc_pcie_reg_paxc_v2[] = {
>> +static const u16 iproc_pcie_reg_paxc_v2[IPROC_PCIE_MAX_NUM_REG] = {
>> [IPROC_PCIE_MSI_GIC_MODE] = 0x050,
>> [IPROC_PCIE_MSI_BASE_ADDR] = 0x074,
>> [IPROC_PCIE_MSI_WINDOW_SIZE] = 0x078,
>> --
>> 2.7.4
>>
On Thu, Mar 26, 2020 at 01:27:36PM -0700, Ray Jui wrote:
> On 3/26/2020 12:48 PM, Bjorn Helgaas wrote:
> > ...
> > It's outside the scope of this patch, but I'm not really a fan of the
> > pcie->reg_offsets[] scheme this driver uses to deal with these
> > differences. There usually seems to be *something* that keeps the
> > driver from referencing registers that don't exist, but it doesn't
> > seem like the mechanism is very consistent or robust:
> >
> > - IPROC_PCIE_LINK_STATUS is implemented by PAXB but not PAXC.
> > iproc_pcie_check_link() avoids using it if "ep_is_internal", which
> > is set for PAXC and PAXC_V2. Not an obvious connection.
> >
> > - IPROC_PCIE_CLK_CTRL is implemented for PAXB and PAXC_V1, but not
> > PAXC_V2. iproc_pcie_perst_ctrl() avoids using it ep_is_internal",
> > so it *doesn't* use it for PAXC_V1, which does implement it.
> > Maybe a bug, maybe intentional; I can't tell.
> >
> > - IPROC_PCIE_INTX_EN is only implemented by PAXB (not PAXC), but
> > AFAICT, we always call iproc_pcie_enable() and rely on
> > iproc_pcie_write_reg() silently drop the write to it on PAXC.
> >
> > - IPROC_PCIE_OARR0 is implemented by PAXB and PAXB_V2 and used by
> > iproc_pcie_map_ranges(), which is called if "need_ob_cfg", which
> > is set if there's a "brcm,pcie-ob" DT property. No clear
> > connection to PAXB.
> >
> > I think it would be more readable if we used a single variant
> > identifier consistently, e.g., the "pcie->type" already used in
> > iproc_pcie_msi_steer(), or maybe a set of variant-specific function
> > pointers as pcie-qcom.c does.
>
> It is not possible to use a single variant identifier consistently,
> i.e., 'pcie->type'. Many of these features are controller revision
> specific, and certain revisions of the controllers may all have a
> certain feature, while other revisions of the controllers do not. In
> addition, there are overlap in features across different controllers.
>
> IMO, it makes sense to have feature specific flags or booleans, and have
> those features enabled or disabled based on 'pcie->type', which is what
> the current driver does, but like you pointed out, what the driver
> failed is to do this consistently.
There are several drivers that have the same problem of dealing with
different revisions of hardware. It would be nice to do it in a
consistent style, whatever that is.
> The IPROC_PCIE_INTX_EN example you pointed out is a good example. I
> agree with you that we shouldn't rely on iproc_pcie_write_reg to
> silently drop the operation for PAXC. We should add code to make it
> explictly obvious that legacy interrupt is not supported in all PAXC
> controllers.
>
> pcie->pcie->reg_offsets[] scheme was not intended to be used to silently
> drop register access that are activated based on features. It's a
> mistake that should be fixed if some code in the driver is done that
> way, as you pointed out.
That's actually why I dug into this a bit -- the
iproc_pcie_reg_is_invalid() case is really a design-time error, so it
seemed like there should be a WARN() there instead of silently
returning 0 or ignoring a write.
Bjorn
On 3/26/2020 1:48 PM, Bjorn Helgaas wrote:
> On Thu, Mar 26, 2020 at 01:27:36PM -0700, Ray Jui wrote:
>> On 3/26/2020 12:48 PM, Bjorn Helgaas wrote:
>>> ...
>>> It's outside the scope of this patch, but I'm not really a fan of the
>>> pcie->reg_offsets[] scheme this driver uses to deal with these
>>> differences. There usually seems to be *something* that keeps the
>>> driver from referencing registers that don't exist, but it doesn't
>>> seem like the mechanism is very consistent or robust:
>>>
>>> - IPROC_PCIE_LINK_STATUS is implemented by PAXB but not PAXC.
>>> iproc_pcie_check_link() avoids using it if "ep_is_internal", which
>>> is set for PAXC and PAXC_V2. Not an obvious connection.
>>>
>>> - IPROC_PCIE_CLK_CTRL is implemented for PAXB and PAXC_V1, but not
>>> PAXC_V2. iproc_pcie_perst_ctrl() avoids using it ep_is_internal",
>>> so it *doesn't* use it for PAXC_V1, which does implement it.
>>> Maybe a bug, maybe intentional; I can't tell.
>>>
>>> - IPROC_PCIE_INTX_EN is only implemented by PAXB (not PAXC), but
>>> AFAICT, we always call iproc_pcie_enable() and rely on
>>> iproc_pcie_write_reg() silently drop the write to it on PAXC.
>>>
>>> - IPROC_PCIE_OARR0 is implemented by PAXB and PAXB_V2 and used by
>>> iproc_pcie_map_ranges(), which is called if "need_ob_cfg", which
>>> is set if there's a "brcm,pcie-ob" DT property. No clear
>>> connection to PAXB.
>>>
>>> I think it would be more readable if we used a single variant
>>> identifier consistently, e.g., the "pcie->type" already used in
>>> iproc_pcie_msi_steer(), or maybe a set of variant-specific function
>>> pointers as pcie-qcom.c does.
>>
>> It is not possible to use a single variant identifier consistently,
>> i.e., 'pcie->type'. Many of these features are controller revision
>> specific, and certain revisions of the controllers may all have a
>> certain feature, while other revisions of the controllers do not. In
>> addition, there are overlap in features across different controllers.
>>
>> IMO, it makes sense to have feature specific flags or booleans, and have
>> those features enabled or disabled based on 'pcie->type', which is what
>> the current driver does, but like you pointed out, what the driver
>> failed is to do this consistently.
>
> There are several drivers that have the same problem of dealing with
> different revisions of hardware. It would be nice to do it in a
> consistent style, whatever that is.
>
Sure, agree with you that it should be handled in a consistent way
within this driver, and the current driver is not handling this
consistently.
>> The IPROC_PCIE_INTX_EN example you pointed out is a good example. I
>> agree with you that we shouldn't rely on iproc_pcie_write_reg to
>> silently drop the operation for PAXC. We should add code to make it
>> explictly obvious that legacy interrupt is not supported in all PAXC
>> controllers.
>>
>> pcie->pcie->reg_offsets[] scheme was not intended to be used to silently
>> drop register access that are activated based on features. It's a
>> mistake that should be fixed if some code in the driver is done that
>> way, as you pointed out.
>
> That's actually why I dug into this a bit -- the
> iproc_pcie_reg_is_invalid() case is really a design-time error, so it
> seemed like there should be a WARN() there instead of silently
> returning 0 or ignoring a write.
>
I think 'iproc_pcie_reg_is_invalid' is a fall back protection. We should
aim to prevent this from happening in the first place using whatever
means we determined appropriate, and do that consistently. In addition,
I also agree with you that there should be a WARN instead of silently
returning zero (for reads) and dropping the writes.
We'll be looking into improving this as you suggested when we have a
chance. In the mean time, I think both of us agree this is out of the
scope of the issue that this patch is trying to fix, which is actually a
pretty critical issue that can cause potential corruption of memory and
the fix should be picked up ASAP (and for older LTS kernels too).
Thanks,
Ray
> Bjorn
>