* [PATCH v3 0/3] Fix find_first_zero_bit() usage
@ 2017-11-27 15:49 Niklas Cassel
2017-11-27 15:49 ` [PATCH v3 1/3] PCI: designware-ep: " Niklas Cassel
` (2 more replies)
0 siblings, 3 replies; 10+ messages in thread
From: Niklas Cassel @ 2017-11-27 15:49 UTC (permalink / raw)
To: linux-pci; +Cc: kishon, Niklas Cassel, linux-kernel
find_first_zero_bit()'s parameter 'size' is defined in bits,
not in bytes.
Calling find_first_zero_bit() with the wrong size unit
will lead to insidious bugs.
Fix all uses of find_first_zero_bit() called with
sizeof() as size argument in drivers/pci.
Also had to fix broken error handling in pci_epc_epf_link()
in order to do proper error handling for find_first_zero_bit().
Niklas Cassel (3):
PCI: designware-ep: Fix find_first_zero_bit() usage
PCI: endpoint: Fix error handling in pci_epc_epf_link()
PCI: endpoint: Fix find_first_zero_bit() usage
drivers/pci/dwc/pcie-designware-ep.c | 22 ++++++++++++++--------
drivers/pci/dwc/pcie-designware.h | 7 +++++--
drivers/pci/endpoint/pci-ep-cfs.c | 14 +++++++++-----
3 files changed, 28 insertions(+), 15 deletions(-)
--
2.14.2
^ permalink raw reply [flat|nested] 10+ messages in thread
* [PATCH v3 1/3] PCI: designware-ep: Fix find_first_zero_bit() usage
2017-11-27 15:49 [PATCH v3 0/3] Fix find_first_zero_bit() usage Niklas Cassel
@ 2017-11-27 15:49 ` Niklas Cassel
2017-11-27 16:55 ` Lorenzo Pieralisi
2017-11-27 15:49 ` [PATCH v3 2/3] PCI: endpoint: Fix error handling in pci_epc_epf_link() Niklas Cassel
2017-11-27 15:49 ` [PATCH v3 3/3] PCI: endpoint: Fix find_first_zero_bit() usage Niklas Cassel
2 siblings, 1 reply; 10+ messages in thread
From: Niklas Cassel @ 2017-11-27 15:49 UTC (permalink / raw)
To: Jingoo Han, Joao Pinto, Lorenzo Pieralisi, Bjorn Helgaas
Cc: kishon, Niklas Cassel, linux-pci, linux-kernel
find_first_zero_bit()'s parameter 'size' is defined in bits,
not in bytes.
find_first_zero_bit() was called with size in 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.
Fix this by using bitmaps with a static upper limit (MAX_IATUS).
256 was chosen since according to the databook, that is the
maximum number of iATUs supported by the IP.
Additionally, make sure that ep->num_ob_windows and ep->num_ib_windows,
which are obtained from device tree, are less than the maximum number of
iATUs (MAX_IATUS).
Fixes: f8aed6ec624f ("PCI: dwc: designware: Add EP mode support")
Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
---
drivers/pci/dwc/pcie-designware-ep.c | 22 ++++++++++++++--------
drivers/pci/dwc/pcie-designware.h | 7 +++++--
2 files changed, 19 insertions(+), 10 deletions(-)
diff --git a/drivers/pci/dwc/pcie-designware-ep.c b/drivers/pci/dwc/pcie-designware-ep.c
index d53d5f168363..8b14e7db5487 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 = find_first_zero_bit(ep->ib_window_map, ep->num_ib_windows);
if (free_win >= ep->num_ib_windows) {
dev_err(pci->dev, "no free inbound window\n");
return -EINVAL;
@@ -85,7 +84,7 @@ static int dw_pcie_ep_inbound_atu(struct dw_pcie_ep *ep, enum pci_barno bar,
}
ep->bar_to_atu[bar] = free_win;
- set_bit(free_win, &ep->ib_window_map);
+ set_bit(free_win, ep->ib_window_map);
return 0;
}
@@ -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 = find_first_zero_bit(ep->ob_window_map, ep->num_ob_windows);
if (free_win >= ep->num_ob_windows) {
dev_err(pci->dev, "no free outbound window\n");
return -EINVAL;
@@ -106,7 +104,7 @@ static int dw_pcie_ep_outbound_atu(struct dw_pcie_ep *ep, phys_addr_t phys_addr,
dw_pcie_prog_outbound_atu(pci, free_win, PCIE_ATU_TYPE_MEM,
phys_addr, pci_addr, size);
- set_bit(free_win, &ep->ob_window_map);
+ set_bit(free_win, ep->ob_window_map);
ep->outbound_addr[free_win] = phys_addr;
return 0;
@@ -121,7 +119,7 @@ static void dw_pcie_ep_clear_bar(struct pci_epc *epc, enum pci_barno bar)
dw_pcie_ep_reset_bar(pci, bar);
dw_pcie_disable_atu(pci, atu_index, DW_PCIE_REGION_INBOUND);
- clear_bit(atu_index, &ep->ib_window_map);
+ clear_bit(atu_index, ep->ib_window_map);
}
static int dw_pcie_ep_set_bar(struct pci_epc *epc, enum pci_barno bar,
@@ -175,7 +173,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, phys_addr_t addr)
return;
dw_pcie_disable_atu(pci, atu_index, DW_PCIE_REGION_OUTBOUND);
- clear_bit(atu_index, &ep->ob_window_map);
+ clear_bit(atu_index, ep->ob_window_map);
}
static int dw_pcie_ep_map_addr(struct pci_epc *epc, phys_addr_t addr,
@@ -298,12 +296,20 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
dev_err(dev, "unable to read *num-ib-windows* property\n");
return ret;
}
+ if (ep->num_ib_windows > MAX_IATUS) {
+ dev_err(dev, "invalid *num-ib-windows*\n");
+ return -EINVAL;
+ }
ret = of_property_read_u32(np, "num-ob-windows", &ep->num_ob_windows);
if (ret < 0) {
dev_err(dev, "unable to read *num-ob-windows* property\n");
return ret;
}
+ if (ep->num_ob_windows > MAX_IATUS) {
+ dev_err(dev, "invalid *num-ob-windows*\n");
+ return -EINVAL;
+ }
addr = devm_kzalloc(dev, sizeof(phys_addr_t) * ep->num_ob_windows,
GFP_KERNEL);
diff --git a/drivers/pci/dwc/pcie-designware.h b/drivers/pci/dwc/pcie-designware.h
index e5d9d77b778e..189f45264c2a 100644
--- a/drivers/pci/dwc/pcie-designware.h
+++ b/drivers/pci/dwc/pcie-designware.h
@@ -113,6 +113,9 @@
#define MAX_MSI_IRQS 32
#define MAX_MSI_CTRLS (MAX_MSI_IRQS / 32)
+/* Maximum number of inbound/outbound iATUs */
+#define MAX_IATUS 256
+
struct pcie_port;
struct dw_pcie;
struct dw_pcie_ep;
@@ -192,8 +195,8 @@ struct dw_pcie_ep {
size_t page_size;
u8 bar_to_atu[6];
phys_addr_t *outbound_addr;
- unsigned long ib_window_map;
- unsigned long ob_window_map;
+ DECLARE_BITMAP(ib_window_map, MAX_IATUS);
+ DECLARE_BITMAP(ob_window_map, MAX_IATUS);
u32 num_ib_windows;
u32 num_ob_windows;
};
--
2.14.2
^ permalink raw reply related [flat|nested] 10+ messages in thread
* [PATCH v3 2/3] PCI: endpoint: Fix error handling in pci_epc_epf_link()
2017-11-27 15:49 [PATCH v3 0/3] Fix find_first_zero_bit() usage Niklas Cassel
2017-11-27 15:49 ` [PATCH v3 1/3] PCI: designware-ep: " Niklas Cassel
@ 2017-11-27 15:49 ` Niklas Cassel
2017-11-27 16:16 ` Lorenzo Pieralisi
2017-11-27 15:49 ` [PATCH v3 3/3] PCI: endpoint: Fix find_first_zero_bit() usage Niklas Cassel
2 siblings, 1 reply; 10+ messages in thread
From: Niklas Cassel @ 2017-11-27 15:49 UTC (permalink / raw)
To: Kishon Vijay Abraham I, Lorenzo Pieralisi, Bjorn Helgaas
Cc: Niklas Cassel, linux-pci, linux-kernel
The error handling in pci_epc_epf_link() was broken,
since the clean up code for pci_epc_add_epf() did clear_bit(),
even though the matching set_bit() is done after pci_epc_add_epf().
Also, clear_bit() should be done before pci_epc_remove_epf(),
since clean up code should normally do things in the reverse order.
Fixes: d74679911610 ("PCI: endpoint: Introduce configfs entry for configuring EP functions")
Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
---
drivers/pci/endpoint/pci-ep-cfs.c | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/drivers/pci/endpoint/pci-ep-cfs.c b/drivers/pci/endpoint/pci-ep-cfs.c
index 4f74386c1ced..e1f5adc9e113 100644
--- a/drivers/pci/endpoint/pci-ep-cfs.c
+++ b/drivers/pci/endpoint/pci-ep-cfs.c
@@ -106,7 +106,7 @@ static int pci_epc_epf_link(struct config_item *epc_item,
epf = epf_group->epf;
ret = pci_epc_add_epf(epc, epf);
if (ret)
- goto err_add_epf;
+ return ret;
func_no = find_first_zero_bit(&epc_group->function_num_map,
sizeof(epc_group->function_num_map));
@@ -120,10 +120,8 @@ static int pci_epc_epf_link(struct config_item *epc_item,
return 0;
err_epf_bind:
- pci_epc_remove_epf(epc, epf);
-
-err_add_epf:
clear_bit(func_no, &epc_group->function_num_map);
+ pci_epc_remove_epf(epc, epf);
return ret;
}
--
2.14.2
^ permalink raw reply related [flat|nested] 10+ messages in thread
* [PATCH v3 3/3] PCI: endpoint: Fix find_first_zero_bit() usage
2017-11-27 15:49 [PATCH v3 0/3] Fix find_first_zero_bit() usage Niklas Cassel
2017-11-27 15:49 ` [PATCH v3 1/3] PCI: designware-ep: " Niklas Cassel
2017-11-27 15:49 ` [PATCH v3 2/3] PCI: endpoint: Fix error handling in pci_epc_epf_link() Niklas Cassel
@ 2017-11-27 15:49 ` Niklas Cassel
2017-11-27 17:44 ` Lorenzo Pieralisi
2017-11-28 9:53 ` David Laight
2 siblings, 2 replies; 10+ messages in thread
From: Niklas Cassel @ 2017-11-27 15:49 UTC (permalink / raw)
To: Kishon Vijay Abraham I, Lorenzo Pieralisi, Bjorn Helgaas
Cc: Niklas Cassel, linux-pci, linux-kernel
find_first_zero_bit()'s parameter 'size' is defined in bits,
not in bytes.
Calling find_first_zero_bit() with the wrong size unit
will lead to insidious bugs.
Fix this by calling find_first_zero_bit() with size
BITS_PER_LONG, rather than sizeof().
Also add proper error handling for find_first_zero_bit(),
since this was missing.
Fixes: d74679911610 ("PCI: endpoint: Introduce configfs entry for configuring EP functions")
Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
---
drivers/pci/endpoint/pci-ep-cfs.c | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/drivers/pci/endpoint/pci-ep-cfs.c b/drivers/pci/endpoint/pci-ep-cfs.c
index e1f5adc9e113..0a22a7976580 100644
--- a/drivers/pci/endpoint/pci-ep-cfs.c
+++ b/drivers/pci/endpoint/pci-ep-cfs.c
@@ -109,7 +109,12 @@ static int pci_epc_epf_link(struct config_item *epc_item,
return ret;
func_no = find_first_zero_bit(&epc_group->function_num_map,
- sizeof(epc_group->function_num_map));
+ BITS_PER_LONG);
+ if (func_no >= BITS_PER_LONG) {
+ ret = -EINVAL;
+ dev_err(&epc->dev, "failed to link endpoint function device\n");
+ goto err_func_no;
+ }
set_bit(func_no, &epc_group->function_num_map);
epf->func_no = func_no;
@@ -121,6 +126,7 @@ static int pci_epc_epf_link(struct config_item *epc_item,
err_epf_bind:
clear_bit(func_no, &epc_group->function_num_map);
+err_func_no:
pci_epc_remove_epf(epc, epf);
return ret;
--
2.14.2
^ permalink raw reply related [flat|nested] 10+ messages in thread
* Re: [PATCH v3 2/3] PCI: endpoint: Fix error handling in pci_epc_epf_link()
2017-11-27 15:49 ` [PATCH v3 2/3] PCI: endpoint: Fix error handling in pci_epc_epf_link() Niklas Cassel
@ 2017-11-27 16:16 ` Lorenzo Pieralisi
0 siblings, 0 replies; 10+ messages in thread
From: Lorenzo Pieralisi @ 2017-11-27 16:16 UTC (permalink / raw)
To: Niklas Cassel
Cc: Kishon Vijay Abraham I, Bjorn Helgaas, Niklas Cassel, linux-pci,
linux-kernel
On Mon, Nov 27, 2017 at 04:49:54PM +0100, Niklas Cassel wrote:
> The error handling in pci_epc_epf_link() was broken,
s/was/is
(ie commit logs should be written in the present tense)
> since the clean up code for pci_epc_add_epf() did clear_bit(),
s/did/calls (see above)
> even though the matching set_bit() is done after pci_epc_add_epf().
>
> Also, clear_bit() should be done before pci_epc_remove_epf(),
> since clean up code should normally do things in the reverse order.
>
> Fixes: d74679911610 ("PCI: endpoint: Introduce configfs entry for configuring EP functions")
> Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
> ---
> drivers/pci/endpoint/pci-ep-cfs.c | 6 ++----
> 1 file changed, 2 insertions(+), 4 deletions(-)
Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
> diff --git a/drivers/pci/endpoint/pci-ep-cfs.c b/drivers/pci/endpoint/pci-ep-cfs.c
> index 4f74386c1ced..e1f5adc9e113 100644
> --- a/drivers/pci/endpoint/pci-ep-cfs.c
> +++ b/drivers/pci/endpoint/pci-ep-cfs.c
> @@ -106,7 +106,7 @@ static int pci_epc_epf_link(struct config_item *epc_item,
> epf = epf_group->epf;
> ret = pci_epc_add_epf(epc, epf);
> if (ret)
> - goto err_add_epf;
> + return ret;
>
> func_no = find_first_zero_bit(&epc_group->function_num_map,
> sizeof(epc_group->function_num_map));
> @@ -120,10 +120,8 @@ static int pci_epc_epf_link(struct config_item *epc_item,
> return 0;
>
> err_epf_bind:
> - pci_epc_remove_epf(epc, epf);
> -
> -err_add_epf:
> clear_bit(func_no, &epc_group->function_num_map);
> + pci_epc_remove_epf(epc, epf);
>
> return ret;
> }
> --
> 2.14.2
>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v3 1/3] PCI: designware-ep: Fix find_first_zero_bit() usage
2017-11-27 15:49 ` [PATCH v3 1/3] PCI: designware-ep: " Niklas Cassel
@ 2017-11-27 16:55 ` Lorenzo Pieralisi
2017-12-08 11:21 ` Lorenzo Pieralisi
0 siblings, 1 reply; 10+ messages in thread
From: Lorenzo Pieralisi @ 2017-11-27 16:55 UTC (permalink / raw)
To: Niklas Cassel
Cc: Jingoo Han, Joao Pinto, Bjorn Helgaas, kishon, Niklas Cassel,
linux-pci, linux-kernel
On Mon, Nov 27, 2017 at 04:49:53PM +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 size in 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.
Same comments about past tense usage.
> Fix this by using bitmaps with a static upper limit (MAX_IATUS).
I think that given that we have the size from FW (num_ib/ob_windows),
you should allocate them dynamically (sorry for asking another respin).
> 256 was chosen since according to the databook, that is the
Nit: I think "datasheet" is more appropriate than "databook" here.
> maximum number of iATUs supported by the IP.
>
> Additionally, make sure that ep->num_ob_windows and ep->num_ib_windows,
> which are obtained from device tree, are less than the maximum number of
s/less/smaller
> iATUs (MAX_IATUS).
>
> Fixes: f8aed6ec624f ("PCI: dwc: designware: Add EP mode support")
> Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
> ---
> drivers/pci/dwc/pcie-designware-ep.c | 22 ++++++++++++++--------
> drivers/pci/dwc/pcie-designware.h | 7 +++++--
> 2 files changed, 19 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/pci/dwc/pcie-designware-ep.c b/drivers/pci/dwc/pcie-designware-ep.c
> index d53d5f168363..8b14e7db5487 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 = find_first_zero_bit(ep->ib_window_map, ep->num_ib_windows);
> if (free_win >= ep->num_ib_windows) {
> dev_err(pci->dev, "no free inbound window\n");
> return -EINVAL;
> @@ -85,7 +84,7 @@ static int dw_pcie_ep_inbound_atu(struct dw_pcie_ep *ep, enum pci_barno bar,
> }
>
> ep->bar_to_atu[bar] = free_win;
> - set_bit(free_win, &ep->ib_window_map);
> + set_bit(free_win, ep->ib_window_map);
>
> return 0;
> }
> @@ -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 = find_first_zero_bit(ep->ob_window_map, ep->num_ob_windows);
> if (free_win >= ep->num_ob_windows) {
> dev_err(pci->dev, "no free outbound window\n");
> return -EINVAL;
> @@ -106,7 +104,7 @@ static int dw_pcie_ep_outbound_atu(struct dw_pcie_ep *ep, phys_addr_t phys_addr,
> dw_pcie_prog_outbound_atu(pci, free_win, PCIE_ATU_TYPE_MEM,
> phys_addr, pci_addr, size);
>
> - set_bit(free_win, &ep->ob_window_map);
> + set_bit(free_win, ep->ob_window_map);
> ep->outbound_addr[free_win] = phys_addr;
>
> return 0;
> @@ -121,7 +119,7 @@ static void dw_pcie_ep_clear_bar(struct pci_epc *epc, enum pci_barno bar)
> dw_pcie_ep_reset_bar(pci, bar);
>
> dw_pcie_disable_atu(pci, atu_index, DW_PCIE_REGION_INBOUND);
> - clear_bit(atu_index, &ep->ib_window_map);
> + clear_bit(atu_index, ep->ib_window_map);
> }
>
> static int dw_pcie_ep_set_bar(struct pci_epc *epc, enum pci_barno bar,
> @@ -175,7 +173,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, phys_addr_t addr)
> return;
>
> dw_pcie_disable_atu(pci, atu_index, DW_PCIE_REGION_OUTBOUND);
> - clear_bit(atu_index, &ep->ob_window_map);
> + clear_bit(atu_index, ep->ob_window_map);
> }
>
> static int dw_pcie_ep_map_addr(struct pci_epc *epc, phys_addr_t addr,
> @@ -298,12 +296,20 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> dev_err(dev, "unable to read *num-ib-windows* property\n");
> return ret;
> }
> + if (ep->num_ib_windows > MAX_IATUS) {
> + dev_err(dev, "invalid *num-ib-windows*\n");
> + return -EINVAL;
> + }
>
> ret = of_property_read_u32(np, "num-ob-windows", &ep->num_ob_windows);
> if (ret < 0) {
> dev_err(dev, "unable to read *num-ob-windows* property\n");
> return ret;
> }
> + if (ep->num_ob_windows > MAX_IATUS) {
> + dev_err(dev, "invalid *num-ob-windows*\n");
> + return -EINVAL;
> + }
>
> addr = devm_kzalloc(dev, sizeof(phys_addr_t) * ep->num_ob_windows,
> GFP_KERNEL);
> diff --git a/drivers/pci/dwc/pcie-designware.h b/drivers/pci/dwc/pcie-designware.h
> index e5d9d77b778e..189f45264c2a 100644
> --- a/drivers/pci/dwc/pcie-designware.h
> +++ b/drivers/pci/dwc/pcie-designware.h
> @@ -113,6 +113,9 @@
> #define MAX_MSI_IRQS 32
> #define MAX_MSI_CTRLS (MAX_MSI_IRQS / 32)
>
> +/* Maximum number of inbound/outbound iATUs */
> +#define MAX_IATUS 256
> +
> struct pcie_port;
> struct dw_pcie;
> struct dw_pcie_ep;
> @@ -192,8 +195,8 @@ struct dw_pcie_ep {
> size_t page_size;
> u8 bar_to_atu[6];
> phys_addr_t *outbound_addr;
> - unsigned long ib_window_map;
> - unsigned long ob_window_map;
> + DECLARE_BITMAP(ib_window_map, MAX_IATUS);
> + DECLARE_BITMAP(ob_window_map, MAX_IATUS);
Only comment I have on the patch is, make the bitmap dynamic ie allocate
them using eg:
ib_window_map = kzalloc(BITS_TO_LONG(num_ib_windows) * sizeof(long),
GFP_KERNEL);
it is a common pattern in the kernel. I just do not like preallocating
static arrays if we know the real size.
It is correct to check the number of ib/ob windows parsed from DT
against MAX_IATUS, since that's a HW limitation therefore a FW (DT)
bug if they are exceeded.
find_first_zero_bit() should use num_ib/ob_windows as size parameter
(ie current patch is correct).
Thanks,
Lorenzo
> u32 num_ib_windows;
> u32 num_ob_windows;
> };
> --
> 2.14.2
>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v3 3/3] PCI: endpoint: Fix find_first_zero_bit() usage
2017-11-27 15:49 ` [PATCH v3 3/3] PCI: endpoint: Fix find_first_zero_bit() usage Niklas Cassel
@ 2017-11-27 17:44 ` Lorenzo Pieralisi
2017-11-28 9:53 ` David Laight
1 sibling, 0 replies; 10+ messages in thread
From: Lorenzo Pieralisi @ 2017-11-27 17:44 UTC (permalink / raw)
To: Niklas Cassel
Cc: Kishon Vijay Abraham I, Bjorn Helgaas, Niklas Cassel, linux-pci,
linux-kernel
On Mon, Nov 27, 2017 at 04:49:55PM +0100, Niklas Cassel wrote:
> find_first_zero_bit()'s parameter 'size' is defined in bits,
> not in bytes.
>
> Calling find_first_zero_bit() with the wrong size unit
> will lead to insidious bugs.
>
> Fix this by calling find_first_zero_bit() with size BITS_PER_LONG,
> rather than sizeof().
>
> Also add proper error handling for find_first_zero_bit(), since this
> was missing.
"Fix this by calling find_first_zero_bit() with size BITS_PER_LONG,
rather than sizeof() and add missing find_first_zero_bit() return
handling"
> Fixes: d74679911610 ("PCI: endpoint: Introduce configfs entry for configuring EP functions")
> Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
> ---
> drivers/pci/endpoint/pci-ep-cfs.c | 8 +++++++-
> 1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/pci/endpoint/pci-ep-cfs.c b/drivers/pci/endpoint/pci-ep-cfs.c
> index e1f5adc9e113..0a22a7976580 100644
> --- a/drivers/pci/endpoint/pci-ep-cfs.c
> +++ b/drivers/pci/endpoint/pci-ep-cfs.c
> @@ -109,7 +109,12 @@ static int pci_epc_epf_link(struct config_item *epc_item,
> return ret;
>
> func_no = find_first_zero_bit(&epc_group->function_num_map,
> - sizeof(epc_group->function_num_map));
> + BITS_PER_LONG);
> + if (func_no >= BITS_PER_LONG) {
> + ret = -EINVAL;
> + dev_err(&epc->dev, "failed to link endpoint function device\n");
I would not log the error for the time being or make it more explicit:
dev_error(&epc->dev, "unsupported number of functions, failed to link
endpoint function device\n");
Side note: we _do_ know the max number of functions the epc supports
(epc->max_functions) when this code is called so we do know the
bitmap size - we can rework this code path based on that (@kishon,
thoughts ?).
Other than that:
Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
> + goto err_func_no;
> + }
> set_bit(func_no, &epc_group->function_num_map);
> epf->func_no = func_no;
>
> @@ -121,6 +126,7 @@ static int pci_epc_epf_link(struct config_item *epc_item,
>
> err_epf_bind:
> clear_bit(func_no, &epc_group->function_num_map);
> +err_func_no:
> pci_epc_remove_epf(epc, epf);
>
> return ret;
> --
> 2.14.2
>
^ permalink raw reply [flat|nested] 10+ messages in thread
* RE: [PATCH v3 3/3] PCI: endpoint: Fix find_first_zero_bit() usage
2017-11-27 15:49 ` [PATCH v3 3/3] PCI: endpoint: Fix find_first_zero_bit() usage Niklas Cassel
2017-11-27 17:44 ` Lorenzo Pieralisi
@ 2017-11-28 9:53 ` David Laight
2017-11-28 12:13 ` Lorenzo Pieralisi
1 sibling, 1 reply; 10+ messages in thread
From: David Laight @ 2017-11-28 9:53 UTC (permalink / raw)
To: 'Niklas Cassel',
Kishon Vijay Abraham I, Lorenzo Pieralisi, Bjorn Helgaas
Cc: Niklas Cassel, linux-pci, linux-kernel
From: Niklas Cassel
> Sent: 27 November 2017 15:50
> find_first_zero_bit()'s parameter 'size' is defined in bits,
> not in bytes.
>
> Calling find_first_zero_bit() with the wrong size unit
> will lead to insidious bugs.
>
> Fix this by calling find_first_zero_bit() with size
> BITS_PER_LONG, rather than sizeof().
>
> Also add proper error handling for find_first_zero_bit(),
> since this was missing.
>
> Fixes: d74679911610 ("PCI: endpoint: Introduce configfs entry for configuring EP functions")
> Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
> ---
> drivers/pci/endpoint/pci-ep-cfs.c | 8 +++++++-
> 1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/pci/endpoint/pci-ep-cfs.c b/drivers/pci/endpoint/pci-ep-cfs.c
> index e1f5adc9e113..0a22a7976580 100644
> --- a/drivers/pci/endpoint/pci-ep-cfs.c
> +++ b/drivers/pci/endpoint/pci-ep-cfs.c
> @@ -109,7 +109,12 @@ static int pci_epc_epf_link(struct config_item *epc_item,
> return ret;
>
> func_no = find_first_zero_bit(&epc_group->function_num_map,
> - sizeof(epc_group->function_num_map));
> + BITS_PER_LONG);
Surely this should be either 8 * sizeof() or you should use ffz()
that takes a numeric argument rather than the function designed
for arbitrary size bitmaps.
David
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v3 3/3] PCI: endpoint: Fix find_first_zero_bit() usage
2017-11-28 9:53 ` David Laight
@ 2017-11-28 12:13 ` Lorenzo Pieralisi
0 siblings, 0 replies; 10+ messages in thread
From: Lorenzo Pieralisi @ 2017-11-28 12:13 UTC (permalink / raw)
To: David Laight
Cc: 'Niklas Cassel',
Kishon Vijay Abraham I, Bjorn Helgaas, Niklas Cassel, linux-pci,
linux-kernel
On Tue, Nov 28, 2017 at 09:53:12AM +0000, David Laight wrote:
> From: Niklas Cassel
> > Sent: 27 November 2017 15:50
> > find_first_zero_bit()'s parameter 'size' is defined in bits,
> > not in bytes.
> >
> > Calling find_first_zero_bit() with the wrong size unit
> > will lead to insidious bugs.
> >
> > Fix this by calling find_first_zero_bit() with size
> > BITS_PER_LONG, rather than sizeof().
> >
> > Also add proper error handling for find_first_zero_bit(),
> > since this was missing.
> >
> > Fixes: d74679911610 ("PCI: endpoint: Introduce configfs entry for configuring EP functions")
> > Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
> > ---
> > drivers/pci/endpoint/pci-ep-cfs.c | 8 +++++++-
> > 1 file changed, 7 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/pci/endpoint/pci-ep-cfs.c b/drivers/pci/endpoint/pci-ep-cfs.c
> > index e1f5adc9e113..0a22a7976580 100644
> > --- a/drivers/pci/endpoint/pci-ep-cfs.c
> > +++ b/drivers/pci/endpoint/pci-ep-cfs.c
> > @@ -109,7 +109,12 @@ static int pci_epc_epf_link(struct config_item *epc_item,
> > return ret;
> >
> > func_no = find_first_zero_bit(&epc_group->function_num_map,
> > - sizeof(epc_group->function_num_map));
> > + BITS_PER_LONG);
>
> Surely this should be either 8 * sizeof() or you should use ffz()
> that takes a numeric argument rather than the function designed
> for arbitrary size bitmaps.
Do you see a problem with this code ? It can be made a bitmap by
allocating it using epc->max_functions as bitmap number of bits
and that can be made on top of the code above that would go in
as a fix unless you strongly object to it, as Joe mentioned
both ffz+| and find_first_zero_bit()+set_bit() are ok with me as
long as we choose one and fix the issue.
Thanks,
Lorenzo
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v3 1/3] PCI: designware-ep: Fix find_first_zero_bit() usage
2017-11-27 16:55 ` Lorenzo Pieralisi
@ 2017-12-08 11:21 ` Lorenzo Pieralisi
0 siblings, 0 replies; 10+ messages in thread
From: Lorenzo Pieralisi @ 2017-12-08 11:21 UTC (permalink / raw)
To: Niklas Cassel
Cc: Jingoo Han, Joao Pinto, Bjorn Helgaas, kishon, Niklas Cassel,
linux-pci, linux-kernel
On Mon, Nov 27, 2017 at 04:55:05PM +0000, Lorenzo Pieralisi wrote:
> On Mon, Nov 27, 2017 at 04:49:53PM +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 size in 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.
>
> Same comments about past tense usage.
>
> > Fix this by using bitmaps with a static upper limit (MAX_IATUS).
>
> I think that given that we have the size from FW (num_ib/ob_windows),
> you should allocate them dynamically (sorry for asking another respin).
>
> > 256 was chosen since according to the databook, that is the
>
> Nit: I think "datasheet" is more appropriate than "databook" here.
>
> > maximum number of iATUs supported by the IP.
> >
> > Additionally, make sure that ep->num_ob_windows and ep->num_ib_windows,
> > which are obtained from device tree, are less than the maximum number of
>
> s/less/smaller
>
> > iATUs (MAX_IATUS).
> >
> > Fixes: f8aed6ec624f ("PCI: dwc: designware: Add EP mode support")
> > Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
> > ---
> > drivers/pci/dwc/pcie-designware-ep.c | 22 ++++++++++++++--------
> > drivers/pci/dwc/pcie-designware.h | 7 +++++--
> > 2 files changed, 19 insertions(+), 10 deletions(-)
> >
> > diff --git a/drivers/pci/dwc/pcie-designware-ep.c b/drivers/pci/dwc/pcie-designware-ep.c
> > index d53d5f168363..8b14e7db5487 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 = find_first_zero_bit(ep->ib_window_map, ep->num_ib_windows);
> > if (free_win >= ep->num_ib_windows) {
> > dev_err(pci->dev, "no free inbound window\n");
> > return -EINVAL;
> > @@ -85,7 +84,7 @@ static int dw_pcie_ep_inbound_atu(struct dw_pcie_ep *ep, enum pci_barno bar,
> > }
> >
> > ep->bar_to_atu[bar] = free_win;
> > - set_bit(free_win, &ep->ib_window_map);
> > + set_bit(free_win, ep->ib_window_map);
> >
> > return 0;
> > }
> > @@ -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 = find_first_zero_bit(ep->ob_window_map, ep->num_ob_windows);
> > if (free_win >= ep->num_ob_windows) {
> > dev_err(pci->dev, "no free outbound window\n");
> > return -EINVAL;
> > @@ -106,7 +104,7 @@ static int dw_pcie_ep_outbound_atu(struct dw_pcie_ep *ep, phys_addr_t phys_addr,
> > dw_pcie_prog_outbound_atu(pci, free_win, PCIE_ATU_TYPE_MEM,
> > phys_addr, pci_addr, size);
> >
> > - set_bit(free_win, &ep->ob_window_map);
> > + set_bit(free_win, ep->ob_window_map);
> > ep->outbound_addr[free_win] = phys_addr;
> >
> > return 0;
> > @@ -121,7 +119,7 @@ static void dw_pcie_ep_clear_bar(struct pci_epc *epc, enum pci_barno bar)
> > dw_pcie_ep_reset_bar(pci, bar);
> >
> > dw_pcie_disable_atu(pci, atu_index, DW_PCIE_REGION_INBOUND);
> > - clear_bit(atu_index, &ep->ib_window_map);
> > + clear_bit(atu_index, ep->ib_window_map);
> > }
> >
> > static int dw_pcie_ep_set_bar(struct pci_epc *epc, enum pci_barno bar,
> > @@ -175,7 +173,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, phys_addr_t addr)
> > return;
> >
> > dw_pcie_disable_atu(pci, atu_index, DW_PCIE_REGION_OUTBOUND);
> > - clear_bit(atu_index, &ep->ob_window_map);
> > + clear_bit(atu_index, ep->ob_window_map);
> > }
> >
> > static int dw_pcie_ep_map_addr(struct pci_epc *epc, phys_addr_t addr,
> > @@ -298,12 +296,20 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> > dev_err(dev, "unable to read *num-ib-windows* property\n");
> > return ret;
> > }
> > + if (ep->num_ib_windows > MAX_IATUS) {
> > + dev_err(dev, "invalid *num-ib-windows*\n");
> > + return -EINVAL;
> > + }
> >
> > ret = of_property_read_u32(np, "num-ob-windows", &ep->num_ob_windows);
> > if (ret < 0) {
> > dev_err(dev, "unable to read *num-ob-windows* property\n");
> > return ret;
> > }
> > + if (ep->num_ob_windows > MAX_IATUS) {
> > + dev_err(dev, "invalid *num-ob-windows*\n");
> > + return -EINVAL;
> > + }
> >
> > addr = devm_kzalloc(dev, sizeof(phys_addr_t) * ep->num_ob_windows,
> > GFP_KERNEL);
> > diff --git a/drivers/pci/dwc/pcie-designware.h b/drivers/pci/dwc/pcie-designware.h
> > index e5d9d77b778e..189f45264c2a 100644
> > --- a/drivers/pci/dwc/pcie-designware.h
> > +++ b/drivers/pci/dwc/pcie-designware.h
> > @@ -113,6 +113,9 @@
> > #define MAX_MSI_IRQS 32
> > #define MAX_MSI_CTRLS (MAX_MSI_IRQS / 32)
> >
> > +/* Maximum number of inbound/outbound iATUs */
> > +#define MAX_IATUS 256
> > +
> > struct pcie_port;
> > struct dw_pcie;
> > struct dw_pcie_ep;
> > @@ -192,8 +195,8 @@ struct dw_pcie_ep {
> > size_t page_size;
> > u8 bar_to_atu[6];
> > phys_addr_t *outbound_addr;
> > - unsigned long ib_window_map;
> > - unsigned long ob_window_map;
> > + DECLARE_BITMAP(ib_window_map, MAX_IATUS);
> > + DECLARE_BITMAP(ob_window_map, MAX_IATUS);
>
> Only comment I have on the patch is, make the bitmap dynamic ie allocate
> them using eg:
>
> ib_window_map = kzalloc(BITS_TO_LONG(num_ib_windows) * sizeof(long),
> GFP_KERNEL);
>
> it is a common pattern in the kernel. I just do not like preallocating
> static arrays if we know the real size.
>
> It is correct to check the number of ib/ob windows parsed from DT
> against MAX_IATUS, since that's a HW limitation therefore a FW (DT)
> bug if they are exceeded.
>
> find_first_zero_bit() should use num_ib/ob_windows as size parameter
> (ie current patch is correct).
Hi Niklas,
May I ask you please to update this series to the latest review comments
and send out a v4 so that Bjorn can pull it please ?
Thanks,
Lorenzo
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2017-12-08 11:20 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-11-27 15:49 [PATCH v3 0/3] Fix find_first_zero_bit() usage Niklas Cassel
2017-11-27 15:49 ` [PATCH v3 1/3] PCI: designware-ep: " Niklas Cassel
2017-11-27 16:55 ` Lorenzo Pieralisi
2017-12-08 11:21 ` Lorenzo Pieralisi
2017-11-27 15:49 ` [PATCH v3 2/3] PCI: endpoint: Fix error handling in pci_epc_epf_link() Niklas Cassel
2017-11-27 16:16 ` Lorenzo Pieralisi
2017-11-27 15:49 ` [PATCH v3 3/3] PCI: endpoint: Fix find_first_zero_bit() usage Niklas Cassel
2017-11-27 17:44 ` Lorenzo Pieralisi
2017-11-28 9:53 ` David Laight
2017-11-28 12:13 ` Lorenzo Pieralisi
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.