| Message ID | 20171116183438.12613-1-niklas.cassel@axis.com (mailing list archive) |
|---|---|
| State | New, archived |
| Delegated to: | Bjorn Helgaas |
| Headers | show |
Please follow Bjorn's request for patch formatting: https://marc.info/?l=linux-pci&m=150905742808166&w=2 I would mention in the commit title that this is a fix. eg "PCI: designware-ep: Fix find_first_zero_bit() usage" On Thu, Nov 16, 2017 at 07:34:38PM +0100, Niklas Cassel wrote: > find_first_zero_bit()'s parameter 'size' is defined in bits, > not in bytes. > > find_first_zero_bit() was called with bytes rather than bits, "was called with size in bytes"... > which thus defined a too low upper limit, causing > dw_pcie_ep_inbound_atu() to assign iatu index #4 to both bar 4 > and bar 5, which made bar 5 overwrite the settings set by bar 4. > > ./pcitest.sh > BAR tests > BAR0: OKAY > BAR1: OKAY > BAR2: OKAY > BAR3: OKAY > BAR4: NOT OKAY > BAR5: OKAY Sorry but this example is pretty much useless - either you add a self-explanatory kernel log or you just remove it. > Fix this by using replacing find_first_zero_bit() with ffz(), > since ffz() only works on a single 'unsigned long' and therefore > does not need a size argument. > > Signed-off-by: Niklas Cassel <niklas.cassel@axis.com> We need a Fixes: tag at least, see Bjorn's post above. Thanks, Lorenzo > --- > drivers/pci/dwc/pcie-designware-ep.c | 6 ++---- > 1 file changed, 2 insertions(+), 4 deletions(-) > > diff --git a/drivers/pci/dwc/pcie-designware-ep.c b/drivers/pci/dwc/pcie-designware-ep.c > index d53d5f168363..ab9a9e160daf 100644 > --- a/drivers/pci/dwc/pcie-designware-ep.c > +++ b/drivers/pci/dwc/pcie-designware-ep.c > @@ -70,8 +70,7 @@ static int dw_pcie_ep_inbound_atu(struct dw_pcie_ep *ep, enum pci_barno bar, > u32 free_win; > struct dw_pcie *pci = to_dw_pcie_from_ep(ep); > > - free_win = find_first_zero_bit(&ep->ib_window_map, > - sizeof(ep->ib_window_map)); > + free_win = ffz(ep->ib_window_map); > if (free_win >= ep->num_ib_windows) { > dev_err(pci->dev, "no free inbound window\n"); > return -EINVAL; > @@ -96,8 +95,7 @@ static int dw_pcie_ep_outbound_atu(struct dw_pcie_ep *ep, phys_addr_t phys_addr, > u32 free_win; > struct dw_pcie *pci = to_dw_pcie_from_ep(ep); > > - free_win = find_first_zero_bit(&ep->ob_window_map, > - sizeof(ep->ob_window_map)); > + free_win = ffz(ep->ob_window_map); > if (free_win >= ep->num_ob_windows) { > dev_err(pci->dev, "no free outbound window\n"); > return -EINVAL; > -- > 2.14.2 >
On Fri, 2017-11-17 at 11:38 +0000, Lorenzo Pieralisi wrote: > Please follow Bjorn's request for patch formatting: > > https://marc.info/?l=linux-pci&m=150905742808166&w=2 > > I would mention in the commit title that this is a fix. > > eg "PCI: designware-ep: Fix find_first_zero_bit() usage" > > On Thu, Nov 16, 2017 at 07:34:38PM +0100, Niklas Cassel wrote: > > find_first_zero_bit()'s parameter 'size' is defined in bits, > > not in bytes. > > > > find_first_zero_bit() was called with bytes rather than bits, > > "was called with size in bytes"... This code also uses set_bit/clear_bit. I suggest not intermixing set_bit/clear_bit calls and ffz on the same object. Use either (&/~/ffz) or (set_bit/clear_bit/find_first_zero_bit) BITS_PER_LONG would work fine for find_first_zero_bit > > which thus defined a too low upper limit, causing > > dw_pcie_ep_inbound_atu() to assign iatu index #4 to both bar 4 > > and bar 5, which made bar 5 overwrite the settings set by bar 4. > > > > ./pcitest.sh > > BAR tests > > BAR0: OKAY > > BAR1: OKAY > > BAR2: OKAY > > BAR3: OKAY > > BAR4: NOT OKAY > > BAR5: OKAY > > Sorry but this example is pretty much useless - either you add > a self-explanatory kernel log or you just remove it. > > > Fix this by using replacing find_first_zero_bit() with ffz(), > > since ffz() only works on a single 'unsigned long' and therefore > > does not need a size argument. > > > > Signed-off-by: Niklas Cassel <niklas.cassel@axis.com> > > We need a Fixes: tag at least, see Bjorn's post above. > > Thanks, > Lorenzo > > > --- > > drivers/pci/dwc/pcie-designware-ep.c | 6 ++---- > > 1 file changed, 2 insertions(+), 4 deletions(-) > > > > diff --git a/drivers/pci/dwc/pcie-designware-ep.c b/drivers/pci/dwc/pcie-designware-ep.c > > index d53d5f168363..ab9a9e160daf 100644 > > --- a/drivers/pci/dwc/pcie-designware-ep.c > > +++ b/drivers/pci/dwc/pcie-designware-ep.c > > @@ -70,8 +70,7 @@ static int dw_pcie_ep_inbound_atu(struct dw_pcie_ep *ep, enum pci_barno bar, > > u32 free_win; > > struct dw_pcie *pci = to_dw_pcie_from_ep(ep); > > > > - free_win = find_first_zero_bit(&ep->ib_window_map, > > - sizeof(ep->ib_window_map)); > > + free_win = ffz(ep->ib_window_map); > > if (free_win >= ep->num_ib_windows) { > > dev_err(pci->dev, "no free inbound window\n"); > > return -EINVAL; > > @@ -96,8 +95,7 @@ static int dw_pcie_ep_outbound_atu(struct dw_pcie_ep *ep, phys_addr_t phys_addr, > > u32 free_win; > > struct dw_pcie *pci = to_dw_pcie_from_ep(ep); > > > > - free_win = find_first_zero_bit(&ep->ob_window_map, > > - sizeof(ep->ob_window_map)); > > + free_win = ffz(ep->ob_window_map); > > if (free_win >= ep->num_ob_windows) { > > dev_err(pci->dev, "no free outbound window\n"); > > return -EINVAL; > > -- > > 2.14.2 > >
On Tue, Nov 21, 2017 at 08:12:10AM -0800, Joe Perches wrote: > On Fri, 2017-11-17 at 11:38 +0000, Lorenzo Pieralisi wrote: > > Please follow Bjorn's request for patch formatting: > > > > https://marc.info/?l=linux-pci&m=150905742808166&w=2 > > > > I would mention in the commit title that this is a fix. > > > > eg "PCI: designware-ep: Fix find_first_zero_bit() usage" > > > > On Thu, Nov 16, 2017 at 07:34:38PM +0100, Niklas Cassel wrote: > > > find_first_zero_bit()'s parameter 'size' is defined in bits, > > > not in bytes. > > > > > > find_first_zero_bit() was called with bytes rather than bits, > > > > "was called with size in bytes"... > > This code also uses set_bit/clear_bit. > > I suggest not intermixing set_bit/clear_bit calls and ffz > on the same object. > > Use either (&/~/ffz) or (set_bit/clear_bit/find_first_zero_bit) > > BITS_PER_LONG would work fine for find_first_zero_bit That would do as a fix and I agree with you that we should not mix the APIs usage (a check for return value < BITS_PER_LONG should be added too though). Actually we can just allocate a bitmap given that we know the bitmap required size (num_{ib/ob}_windows) and update the find_first_zero_bit() call with the appropriate size in bits parameter but that can be a subsequent refactoring patch. Thanks, Lorenzo > > > which thus defined a too low upper limit, causing > > > dw_pcie_ep_inbound_atu() to assign iatu index #4 to both bar 4 > > > and bar 5, which made bar 5 overwrite the settings set by bar 4. > > > > > > ./pcitest.sh > > > BAR tests > > > BAR0: OKAY > > > BAR1: OKAY > > > BAR2: OKAY > > > BAR3: OKAY > > > BAR4: NOT OKAY > > > BAR5: OKAY > > > > Sorry but this example is pretty much useless - either you add > > a self-explanatory kernel log or you just remove it. > > > > > Fix this by using replacing find_first_zero_bit() with ffz(), > > > since ffz() only works on a single 'unsigned long' and therefore > > > does not need a size argument. > > > > > > Signed-off-by: Niklas Cassel <niklas.cassel@axis.com> > > > > We need a Fixes: tag at least, see Bjorn's post above. > > > > Thanks, > > Lorenzo > > > > > --- > > > drivers/pci/dwc/pcie-designware-ep.c | 6 ++---- > > > 1 file changed, 2 insertions(+), 4 deletions(-) > > > > > > diff --git a/drivers/pci/dwc/pcie-designware-ep.c b/drivers/pci/dwc/pcie-designware-ep.c > > > index d53d5f168363..ab9a9e160daf 100644 > > > --- a/drivers/pci/dwc/pcie-designware-ep.c > > > +++ b/drivers/pci/dwc/pcie-designware-ep.c > > > @@ -70,8 +70,7 @@ static int dw_pcie_ep_inbound_atu(struct dw_pcie_ep *ep, enum pci_barno bar, > > > u32 free_win; > > > struct dw_pcie *pci = to_dw_pcie_from_ep(ep); > > > > > > - free_win = find_first_zero_bit(&ep->ib_window_map, > > > - sizeof(ep->ib_window_map)); > > > + free_win = ffz(ep->ib_window_map); > > > if (free_win >= ep->num_ib_windows) { > > > dev_err(pci->dev, "no free inbound window\n"); > > > return -EINVAL; > > > @@ -96,8 +95,7 @@ static int dw_pcie_ep_outbound_atu(struct dw_pcie_ep *ep, phys_addr_t phys_addr, > > > u32 free_win; > > > struct dw_pcie *pci = to_dw_pcie_from_ep(ep); > > > > > > - free_win = find_first_zero_bit(&ep->ob_window_map, > > > - sizeof(ep->ob_window_map)); > > > + free_win = ffz(ep->ob_window_map); > > > if (free_win >= ep->num_ob_windows) { > > > dev_err(pci->dev, "no free outbound window\n"); > > > return -EINVAL; > > > -- > > > 2.14.2 > > >
diff --git a/drivers/pci/dwc/pcie-designware-ep.c b/drivers/pci/dwc/pcie-designware-ep.c index d53d5f168363..ab9a9e160daf 100644 --- a/drivers/pci/dwc/pcie-designware-ep.c +++ b/drivers/pci/dwc/pcie-designware-ep.c @@ -70,8 +70,7 @@ static int dw_pcie_ep_inbound_atu(struct dw_pcie_ep *ep, enum pci_barno bar, u32 free_win; struct dw_pcie *pci = to_dw_pcie_from_ep(ep); - free_win = find_first_zero_bit(&ep->ib_window_map, - sizeof(ep->ib_window_map)); + free_win = ffz(ep->ib_window_map); if (free_win >= ep->num_ib_windows) { dev_err(pci->dev, "no free inbound window\n"); return -EINVAL; @@ -96,8 +95,7 @@ static int dw_pcie_ep_outbound_atu(struct dw_pcie_ep *ep, phys_addr_t phys_addr, u32 free_win; struct dw_pcie *pci = to_dw_pcie_from_ep(ep); - free_win = find_first_zero_bit(&ep->ob_window_map, - sizeof(ep->ob_window_map)); + free_win = ffz(ep->ob_window_map); if (free_win >= ep->num_ob_windows) { dev_err(pci->dev, "no free outbound window\n"); return -EINVAL;
find_first_zero_bit()'s parameter 'size' is defined in bits, not in bytes. find_first_zero_bit() was called with bytes rather than bits, which thus defined a too low upper limit, causing dw_pcie_ep_inbound_atu() to assign iatu index #4 to both bar 4 and bar 5, which made bar 5 overwrite the settings set by bar 4. ./pcitest.sh BAR tests BAR0: OKAY BAR1: OKAY BAR2: OKAY BAR3: OKAY BAR4: NOT OKAY BAR5: OKAY Fix this by using replacing find_first_zero_bit() with ffz(), since ffz() only works on a single 'unsigned long' and therefore does not need a size argument. Signed-off-by: Niklas Cassel <niklas.cassel@axis.com> --- drivers/pci/dwc/pcie-designware-ep.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-)