* [PATCH] PCI: mobiveil: integer overflow in IB_WIN_SIZE
@ 2018-07-05 9:45 Dan Carpenter
2018-07-06 11:26 ` Lorenzo Pieralisi
2018-07-26 10:56 ` Lorenzo Pieralisi
0 siblings, 2 replies; 8+ messages in thread
From: Dan Carpenter @ 2018-07-05 9:45 UTC (permalink / raw)
To: Subrahmanya Lingappa
Cc: Lorenzo Pieralisi, Bjorn Helgaas, linux-pci, kernel-janitors
IB_WIN_SIZE is larger than INT_MAX so we need to cast it to u64.
Fixes: 9af6bcb11e12 ("PCI: mobiveil: Add Mobiveil PCIe Host Bridge IP driver")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
---
Untested... Possibly the problem is that there are too many multiplies
by * 1024.
diff --git a/drivers/pci/controller/pcie-mobiveil.c b/drivers/pci/controller/pcie-mobiveil.c
index 4d6c20e47bed..cf0aa7cee5b0 100644
--- a/drivers/pci/controller/pcie-mobiveil.c
+++ b/drivers/pci/controller/pcie-mobiveil.c
@@ -107,7 +107,7 @@
#define CFG_WINDOW_TYPE 0
#define IO_WINDOW_TYPE 1
#define MEM_WINDOW_TYPE 2
-#define IB_WIN_SIZE (256 * 1024 * 1024 * 1024)
+#define IB_WIN_SIZE ((u64)256 * 1024 * 1024 * 1024)
#define MAX_PIO_WINDOWS 8
/* Parameters for the waiting for link up routine */
^ permalink raw reply related [flat|nested] 8+ messages in thread* Re: [PATCH] PCI: mobiveil: integer overflow in IB_WIN_SIZE 2018-07-05 9:45 [PATCH] PCI: mobiveil: integer overflow in IB_WIN_SIZE Dan Carpenter @ 2018-07-06 11:26 ` Lorenzo Pieralisi 2018-07-09 10:25 ` Subrahmanya Lingappa 2018-07-26 10:56 ` Lorenzo Pieralisi 1 sibling, 1 reply; 8+ messages in thread From: Lorenzo Pieralisi @ 2018-07-06 11:26 UTC (permalink / raw) To: Dan Carpenter, Subrahmanya Lingappa Cc: Bjorn Helgaas, linux-pci, kernel-janitors On Thu, Jul 05, 2018 at 12:45:39PM +0300, Dan Carpenter wrote: > IB_WIN_SIZE is larger than INT_MAX so we need to cast it to u64. > > Fixes: 9af6bcb11e12 ("PCI: mobiveil: Add Mobiveil PCIe Host Bridge IP driver") > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> > --- > Untested... Possibly the problem is that there are too many multiplies > by * 1024. Hi Subrahmanya, I always thought that that macro value was questionable, please can you comment on this ? Thanks, Lorenzo > diff --git a/drivers/pci/controller/pcie-mobiveil.c b/drivers/pci/controller/pcie-mobiveil.c > index 4d6c20e47bed..cf0aa7cee5b0 100644 > --- a/drivers/pci/controller/pcie-mobiveil.c > +++ b/drivers/pci/controller/pcie-mobiveil.c > @@ -107,7 +107,7 @@ > #define CFG_WINDOW_TYPE 0 > #define IO_WINDOW_TYPE 1 > #define MEM_WINDOW_TYPE 2 > -#define IB_WIN_SIZE (256 * 1024 * 1024 * 1024) > +#define IB_WIN_SIZE ((u64)256 * 1024 * 1024 * 1024) > #define MAX_PIO_WINDOWS 8 > > /* Parameters for the waiting for link up routine */ ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] PCI: mobiveil: integer overflow in IB_WIN_SIZE 2018-07-06 11:26 ` Lorenzo Pieralisi @ 2018-07-09 10:25 ` Subrahmanya Lingappa 2018-07-09 11:42 ` Lorenzo Pieralisi 0 siblings, 1 reply; 8+ messages in thread From: Subrahmanya Lingappa @ 2018-07-09 10:25 UTC (permalink / raw) To: Lorenzo Pieralisi Cc: Dan Carpenter, Bjorn Helgaas, linux-pci, kernel-janitors Hi Lorenzo, Our reference IP's hardware registers require the to support 256GB inbound window size and this driver did exactly that I believe. May I ask, why do you think this size is questionable ? Thanks. On Fri, Jul 6, 2018 at 4:56 PM, Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> wrote: > On Thu, Jul 05, 2018 at 12:45:39PM +0300, Dan Carpenter wrote: >> IB_WIN_SIZE is larger than INT_MAX so we need to cast it to u64. >> >> Fixes: 9af6bcb11e12 ("PCI: mobiveil: Add Mobiveil PCIe Host Bridge IP driver") >> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> >> --- >> Untested... Possibly the problem is that there are too many multiplies >> by * 1024. > > Hi Subrahmanya, > > I always thought that that macro value was questionable, please > can you comment on this ? > > Thanks, > Lorenzo > >> diff --git a/drivers/pci/controller/pcie-mobiveil.c b/drivers/pci/controller/pcie-mobiveil.c >> index 4d6c20e47bed..cf0aa7cee5b0 100644 >> --- a/drivers/pci/controller/pcie-mobiveil.c >> +++ b/drivers/pci/controller/pcie-mobiveil.c >> @@ -107,7 +107,7 @@ >> #define CFG_WINDOW_TYPE 0 >> #define IO_WINDOW_TYPE 1 >> #define MEM_WINDOW_TYPE 2 >> -#define IB_WIN_SIZE (256 * 1024 * 1024 * 1024) >> +#define IB_WIN_SIZE ((u64)256 * 1024 * 1024 * 1024) >> #define MAX_PIO_WINDOWS 8 >> >> /* Parameters for the waiting for link up routine */ ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] PCI: mobiveil: integer overflow in IB_WIN_SIZE 2018-07-09 10:25 ` Subrahmanya Lingappa @ 2018-07-09 11:42 ` Lorenzo Pieralisi 2018-07-16 6:40 ` Subrahmanya Lingappa 0 siblings, 1 reply; 8+ messages in thread From: Lorenzo Pieralisi @ 2018-07-09 11:42 UTC (permalink / raw) To: Subrahmanya Lingappa Cc: Dan Carpenter, Bjorn Helgaas, linux-pci, kernel-janitors On Mon, Jul 09, 2018 at 03:55:00PM +0530, Subrahmanya Lingappa wrote: > Hi Lorenzo, > > Our reference IP's hardware registers require the to support 256GB > inbound window size and this driver did exactly that I believe. No it does not, please check what this patch does, the window size is not currently configured as you wish. More importantly, nobody can even test it given that this driver lacks a Kconfig and a Makefile entry unless I am missing something, I presume they were lost in patch reviews but you ought to fix this as soon as possible please since this was something I should not have missed. > May I ask, why do you think this size is questionable ? The macro is not questionable, it is buggy, the size is what you tell me it should be, I agree. Lorenzo > > Thanks. > > > On Fri, Jul 6, 2018 at 4:56 PM, Lorenzo Pieralisi > <lorenzo.pieralisi@arm.com> wrote: > > On Thu, Jul 05, 2018 at 12:45:39PM +0300, Dan Carpenter wrote: > >> IB_WIN_SIZE is larger than INT_MAX so we need to cast it to u64. > >> > >> Fixes: 9af6bcb11e12 ("PCI: mobiveil: Add Mobiveil PCIe Host Bridge IP driver") > >> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> > >> --- > >> Untested... Possibly the problem is that there are too many multiplies > >> by * 1024. > > > > Hi Subrahmanya, > > > > I always thought that that macro value was questionable, please > > can you comment on this ? > > > > Thanks, > > Lorenzo > > > >> diff --git a/drivers/pci/controller/pcie-mobiveil.c b/drivers/pci/controller/pcie-mobiveil.c > >> index 4d6c20e47bed..cf0aa7cee5b0 100644 > >> --- a/drivers/pci/controller/pcie-mobiveil.c > >> +++ b/drivers/pci/controller/pcie-mobiveil.c > >> @@ -107,7 +107,7 @@ > >> #define CFG_WINDOW_TYPE 0 > >> #define IO_WINDOW_TYPE 1 > >> #define MEM_WINDOW_TYPE 2 > >> -#define IB_WIN_SIZE (256 * 1024 * 1024 * 1024) > >> +#define IB_WIN_SIZE ((u64)256 * 1024 * 1024 * 1024) > >> #define MAX_PIO_WINDOWS 8 > >> > >> /* Parameters for the waiting for link up routine */ ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] PCI: mobiveil: integer overflow in IB_WIN_SIZE 2018-07-09 11:42 ` Lorenzo Pieralisi @ 2018-07-16 6:40 ` Subrahmanya Lingappa 2018-07-26 9:46 ` Lorenzo Pieralisi 0 siblings, 1 reply; 8+ messages in thread From: Subrahmanya Lingappa @ 2018-07-16 6:40 UTC (permalink / raw) To: Lorenzo Pieralisi Cc: Dan Carpenter, Bjorn Helgaas, linux-pci, kernel-janitors Lorenzo, looks like I missed it during the review process; I'll submit the Kconfig and makefile as soon as possible. Thanks, On Mon, Jul 9, 2018 at 5:12 PM, Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> wrote: > On Mon, Jul 09, 2018 at 03:55:00PM +0530, Subrahmanya Lingappa wrote: >> Hi Lorenzo, >> >> Our reference IP's hardware registers require the to support 256GB >> inbound window size and this driver did exactly that I believe. > > No it does not, please check what this patch does, the window size > is not currently configured as you wish. > > More importantly, nobody can even test it given that this driver lacks a > Kconfig and a Makefile entry unless I am missing something, I presume > they were lost in patch reviews but you ought to fix this as soon as > possible please since this was something I should not have missed. > >> May I ask, why do you think this size is questionable ? > > The macro is not questionable, it is buggy, the size is what you > tell me it should be, I agree. > > Lorenzo > >> >> Thanks. >> >> >> On Fri, Jul 6, 2018 at 4:56 PM, Lorenzo Pieralisi >> <lorenzo.pieralisi@arm.com> wrote: >> > On Thu, Jul 05, 2018 at 12:45:39PM +0300, Dan Carpenter wrote: >> >> IB_WIN_SIZE is larger than INT_MAX so we need to cast it to u64. >> >> >> >> Fixes: 9af6bcb11e12 ("PCI: mobiveil: Add Mobiveil PCIe Host Bridge IP driver") >> >> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> >> >> --- >> >> Untested... Possibly the problem is that there are too many multiplies >> >> by * 1024. >> > >> > Hi Subrahmanya, >> > >> > I always thought that that macro value was questionable, please >> > can you comment on this ? >> > >> > Thanks, >> > Lorenzo >> > >> >> diff --git a/drivers/pci/controller/pcie-mobiveil.c b/drivers/pci/controller/pcie-mobiveil.c >> >> index 4d6c20e47bed..cf0aa7cee5b0 100644 >> >> --- a/drivers/pci/controller/pcie-mobiveil.c >> >> +++ b/drivers/pci/controller/pcie-mobiveil.c >> >> @@ -107,7 +107,7 @@ >> >> #define CFG_WINDOW_TYPE 0 >> >> #define IO_WINDOW_TYPE 1 >> >> #define MEM_WINDOW_TYPE 2 >> >> -#define IB_WIN_SIZE (256 * 1024 * 1024 * 1024) >> >> +#define IB_WIN_SIZE ((u64)256 * 1024 * 1024 * 1024) >> >> #define MAX_PIO_WINDOWS 8 >> >> >> >> /* Parameters for the waiting for link up routine */ ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] PCI: mobiveil: integer overflow in IB_WIN_SIZE 2018-07-16 6:40 ` Subrahmanya Lingappa @ 2018-07-26 9:46 ` Lorenzo Pieralisi 0 siblings, 0 replies; 8+ messages in thread From: Lorenzo Pieralisi @ 2018-07-26 9:46 UTC (permalink / raw) To: Subrahmanya Lingappa Cc: Dan Carpenter, Bjorn Helgaas, linux-pci, kernel-janitors On Mon, Jul 16, 2018 at 12:10:46PM +0530, Subrahmanya Lingappa wrote: > Lorenzo, > > looks like I missed it during the review process; > I'll submit the Kconfig and makefile as soon as possible. Hi Subrahmanya, Will you have time to submit a patch soon ? I would like to queue this fix too. Thanks, Lorenzo > Thanks, > > On Mon, Jul 9, 2018 at 5:12 PM, Lorenzo Pieralisi > <lorenzo.pieralisi@arm.com> wrote: > > On Mon, Jul 09, 2018 at 03:55:00PM +0530, Subrahmanya Lingappa wrote: > >> Hi Lorenzo, > >> > >> Our reference IP's hardware registers require the to support 256GB > >> inbound window size and this driver did exactly that I believe. > > > > No it does not, please check what this patch does, the window size > > is not currently configured as you wish. > > > > More importantly, nobody can even test it given that this driver lacks a > > Kconfig and a Makefile entry unless I am missing something, I presume > > they were lost in patch reviews but you ought to fix this as soon as > > possible please since this was something I should not have missed. > > > >> May I ask, why do you think this size is questionable ? > > > > The macro is not questionable, it is buggy, the size is what you > > tell me it should be, I agree. > > > > Lorenzo > > > >> > >> Thanks. > >> > >> > >> On Fri, Jul 6, 2018 at 4:56 PM, Lorenzo Pieralisi > >> <lorenzo.pieralisi@arm.com> wrote: > >> > On Thu, Jul 05, 2018 at 12:45:39PM +0300, Dan Carpenter wrote: > >> >> IB_WIN_SIZE is larger than INT_MAX so we need to cast it to u64. > >> >> > >> >> Fixes: 9af6bcb11e12 ("PCI: mobiveil: Add Mobiveil PCIe Host Bridge IP driver") > >> >> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> > >> >> --- > >> >> Untested... Possibly the problem is that there are too many multiplies > >> >> by * 1024. > >> > > >> > Hi Subrahmanya, > >> > > >> > I always thought that that macro value was questionable, please > >> > can you comment on this ? > >> > > >> > Thanks, > >> > Lorenzo > >> > > >> >> diff --git a/drivers/pci/controller/pcie-mobiveil.c b/drivers/pci/controller/pcie-mobiveil.c > >> >> index 4d6c20e47bed..cf0aa7cee5b0 100644 > >> >> --- a/drivers/pci/controller/pcie-mobiveil.c > >> >> +++ b/drivers/pci/controller/pcie-mobiveil.c > >> >> @@ -107,7 +107,7 @@ > >> >> #define CFG_WINDOW_TYPE 0 > >> >> #define IO_WINDOW_TYPE 1 > >> >> #define MEM_WINDOW_TYPE 2 > >> >> -#define IB_WIN_SIZE (256 * 1024 * 1024 * 1024) > >> >> +#define IB_WIN_SIZE ((u64)256 * 1024 * 1024 * 1024) > >> >> #define MAX_PIO_WINDOWS 8 > >> >> > >> >> /* Parameters for the waiting for link up routine */ ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] PCI: mobiveil: integer overflow in IB_WIN_SIZE 2018-07-05 9:45 [PATCH] PCI: mobiveil: integer overflow in IB_WIN_SIZE Dan Carpenter 2018-07-06 11:26 ` Lorenzo Pieralisi @ 2018-07-26 10:56 ` Lorenzo Pieralisi 2018-08-01 7:07 ` Subrahmanya Lingappa 1 sibling, 1 reply; 8+ messages in thread From: Lorenzo Pieralisi @ 2018-07-26 10:56 UTC (permalink / raw) To: Dan Carpenter Cc: Subrahmanya Lingappa, Bjorn Helgaas, linux-pci, kernel-janitors On Thu, Jul 05, 2018 at 12:45:39PM +0300, Dan Carpenter wrote: > IB_WIN_SIZE is larger than INT_MAX so we need to cast it to u64. > > Fixes: 9af6bcb11e12 ("PCI: mobiveil: Add Mobiveil PCIe Host Bridge IP driver") > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> > --- > Untested... Possibly the problem is that there are too many multiplies > by * 1024. > > diff --git a/drivers/pci/controller/pcie-mobiveil.c b/drivers/pci/controller/pcie-mobiveil.c Applied to pci/mobiveil for v4.19 (it is still not fixing anything given that the driver lacks a Kconfig and Makefile entries - that must be fixed with urgency), thanks. Lorenzo > index 4d6c20e47bed..cf0aa7cee5b0 100644 > --- a/drivers/pci/controller/pcie-mobiveil.c > +++ b/drivers/pci/controller/pcie-mobiveil.c > @@ -107,7 +107,7 @@ > #define CFG_WINDOW_TYPE 0 > #define IO_WINDOW_TYPE 1 > #define MEM_WINDOW_TYPE 2 > -#define IB_WIN_SIZE (256 * 1024 * 1024 * 1024) > +#define IB_WIN_SIZE ((u64)256 * 1024 * 1024 * 1024) > #define MAX_PIO_WINDOWS 8 > > /* Parameters for the waiting for link up routine */ ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] PCI: mobiveil: integer overflow in IB_WIN_SIZE 2018-07-26 10:56 ` Lorenzo Pieralisi @ 2018-08-01 7:07 ` Subrahmanya Lingappa 0 siblings, 0 replies; 8+ messages in thread From: Subrahmanya Lingappa @ 2018-08-01 7:07 UTC (permalink / raw) To: Lorenzo Pieralisi Cc: Dan Carpenter, Bjorn Helgaas, linux-pci, kernel-janitors Lorenzo, Added the compilation support in the next patch submitted at : https://patchwork.kernel.org/patch/10551771/ please review. Thanks, Subrahmanya On Thu, Jul 26, 2018 at 4:26 PM, Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> wrote: > On Thu, Jul 05, 2018 at 12:45:39PM +0300, Dan Carpenter wrote: >> IB_WIN_SIZE is larger than INT_MAX so we need to cast it to u64. >> >> Fixes: 9af6bcb11e12 ("PCI: mobiveil: Add Mobiveil PCIe Host Bridge IP driver") >> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> >> --- >> Untested... Possibly the problem is that there are too many multiplies >> by * 1024. >> >> diff --git a/drivers/pci/controller/pcie-mobiveil.c b/drivers/pci/controller/pcie-mobiveil.c > > Applied to pci/mobiveil for v4.19 (it is still not fixing anything > given that the driver lacks a Kconfig and Makefile entries - that > must be fixed with urgency), thanks. > > Lorenzo > >> index 4d6c20e47bed..cf0aa7cee5b0 100644 >> --- a/drivers/pci/controller/pcie-mobiveil.c >> +++ b/drivers/pci/controller/pcie-mobiveil.c >> @@ -107,7 +107,7 @@ >> #define CFG_WINDOW_TYPE 0 >> #define IO_WINDOW_TYPE 1 >> #define MEM_WINDOW_TYPE 2 >> -#define IB_WIN_SIZE (256 * 1024 * 1024 * 1024) >> +#define IB_WIN_SIZE ((u64)256 * 1024 * 1024 * 1024) >> #define MAX_PIO_WINDOWS 8 >> >> /* Parameters for the waiting for link up routine */ ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2018-08-01 8:51 UTC | newest] Thread overview: 8+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2018-07-05 9:45 [PATCH] PCI: mobiveil: integer overflow in IB_WIN_SIZE Dan Carpenter 2018-07-06 11:26 ` Lorenzo Pieralisi 2018-07-09 10:25 ` Subrahmanya Lingappa 2018-07-09 11:42 ` Lorenzo Pieralisi 2018-07-16 6:40 ` Subrahmanya Lingappa 2018-07-26 9:46 ` Lorenzo Pieralisi 2018-07-26 10:56 ` Lorenzo Pieralisi 2018-08-01 7:07 ` Subrahmanya Lingappa
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).