Message ID | 1406270593-17193-1-git-send-email-bisson.gary@gmail.com |
---|---|
State | Accepted |
Headers | show |
Gary, All, On 2014-07-24 23:43 -0700, Gary Bisson spake thusly: > Same as similar patches for imx-vpu, the returned value from > IOGetVirtMem was not properly checked. That's the same problem as seen in imx-vpu and fixed by: package/freescale-imx/imx-vpu/imx-vpu-0001-vpu-io-fix-IOSystemInit-failure.patch right? > Signed-off-by: Gary Bisson <bisson.gary@gmail.com> > > --- > Yann, Eric, all, > This patch is just a copy of what Eric pushed for this package in Yocto. > Thanks, > Gary > --- > ...-fix-tests-of-return-value-from-IOGetVirt.patch | 41 ++++++++++++++++++++++ > 1 file changed, 41 insertions(+) > create mode 100644 package/libfslvpuwrap/libfslvpuwrap-0001-vpu_wrapper-fix-tests-of-return-value-from-IOGetVirt.patch > > diff --git a/package/libfslvpuwrap/libfslvpuwrap-0001-vpu_wrapper-fix-tests-of-return-value-from-IOGetVirt.patch b/package/libfslvpuwrap/libfslvpuwrap-0001-vpu_wrapper-fix-tests-of-return-value-from-IOGetVirt.patch > new file mode 100644 > index 0000000..ccd2d9e > --- /dev/null > +++ b/package/libfslvpuwrap/libfslvpuwrap-0001-vpu_wrapper-fix-tests-of-return-value-from-IOGetVirt.patch > @@ -0,0 +1,41 @@ > +From 7ca66c10c05168c7d342df7c7a70d4a1ae0629f7 Mon Sep 17 00:00:00 2001 > +From: Eric Nelson <eric.nelson@boundarydevices.com> > +Date: Sat, 28 Jun 2014 09:45:09 -0700 > +Subject: [PATCH] vpu_wrapper: fix tests of return value from IOGetVirtMem > + > +IOGetVirtMem() returns a pointer or specifically -1 (MAP_FAILED) > +in the case of failure. > + > +Upstream-Status: Pending Is there a public URL to the bug report or the patch submission? If so, can you paste the URL here, so we know when we update that we need to remove that patch? With the URL added, you can add my: Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Regards, Yann E. MORIN. > + > +Signed-off-by: Laci Tele <laci@boundarydevices.com> > +Signed-off-by: Eric Nelson <eric.nelson@boundarydevices.com> > +--- > + vpu_wrapper.c | 4 ++-- > + 1 file changed, 2 insertions(+), 2 deletions(-) > + > +diff --git a/vpu_wrapper.c b/vpu_wrapper.c > +index 9249174..148c5df 100755 > +--- a/vpu_wrapper.c > ++++ b/vpu_wrapper.c > +@@ -6653,7 +6653,7 @@ VpuDecRetCode VPU_DecGetMem(VpuMemDesc* pInOutMem) > + return VPU_DEC_RET_FAILURE; > + } > + ret=IOGetVirtMem(&buff); > +- if(ret==0) //if(ret!=RETCODE_SUCCESS) > ++ if(ret == -1) //if(ret==MAP_FAILED) > + { > + VPU_ERROR("%s: get virtual memory failure: size=%d, ret=%d \r\n",__FUNCTION__,buff.size,ret); > + return VPU_DEC_RET_FAILURE; > +@@ -8277,7 +8277,7 @@ VpuEncRetCode VPU_EncGetMem(VpuMemDesc* pInOutMem) > + return VPU_ENC_RET_FAILURE; > + } > + ret=IOGetVirtMem(&buff); > +- if(ret==0) //if(ret!=RETCODE_SUCCESS) > ++ if(ret == -1) //if(ret==MAP_FAILED) > + { > + VPU_ENC_ERROR("%s: get virtual memory failure: size=%d, ret=%d \r\n",__FUNCTION__,buff.size,(UINT32)ret); > + return VPU_ENC_RET_FAILURE; > +-- > +1.9.1 > + > -- > 2.0.1 >
Hi Yann, On 07/25/2014 09:55 AM, Yann E. MORIN wrote: > Gary, All, > > <snip> > >> --- /dev/null >> +++ b/package/libfslvpuwrap/libfslvpuwrap-0001-vpu_wrapper-fix-tests-of-return-value-from-IOGetVirt.patch >> @@ -0,0 +1,41 @@ >> +From 7ca66c10c05168c7d342df7c7a70d4a1ae0629f7 Mon Sep 17 00:00:00 2001 >> +From: Eric Nelson <eric.nelson@boundarydevices.com> >> +Date: Sat, 28 Jun 2014 09:45:09 -0700 >> +Subject: [PATCH] vpu_wrapper: fix tests of return value from IOGetVirtMem >> + >> +IOGetVirtMem() returns a pointer or specifically -1 (MAP_FAILED) >> +in the case of failure. >> + >> +Upstream-Status: Pending > > Is there a public URL to the bug report or the patch submission? > If so, can you paste the URL here, so we know when we update that we > need to remove that patch? > Since Freescale is now basing their releases on Yocto, the meta-freescale list is the best equivalent of "upstream" for these Freescale components. Unfortunately, patches there are in a different form, and I pushed the equivalent forms and they were acked and promises made by Freescale (Lauren) to include in the next release: https://lists.yoctoproject.org/pipermail/meta-freescale/2014-June/thread.html#8881 Regards, Eric
Eric, All, On 2014-07-25 10:01 -0700, Eric Nelson spake thusly: > On 07/25/2014 09:55 AM, Yann E. MORIN wrote: [--SNIP--] > >> --- /dev/null > >> +++ b/package/libfslvpuwrap/libfslvpuwrap-0001-vpu_wrapper-fix-tests-of-return-value-from-IOGetVirt.patch > >> @@ -0,0 +1,41 @@ > >> +From 7ca66c10c05168c7d342df7c7a70d4a1ae0629f7 Mon Sep 17 00:00:00 2001 > >> +From: Eric Nelson <eric.nelson@boundarydevices.com> > >> +Date: Sat, 28 Jun 2014 09:45:09 -0700 > >> +Subject: [PATCH] vpu_wrapper: fix tests of return value from IOGetVirtMem > >> + > >> +IOGetVirtMem() returns a pointer or specifically -1 (MAP_FAILED) > >> +in the case of failure. > >> + > >> +Upstream-Status: Pending > > > > Is there a public URL to the bug report or the patch submission? > > If so, can you paste the URL here, so we know when we update that we > > need to remove that patch? > > > > Since Freescale is now basing their releases on Yocto, the > meta-freescale list is the best equivalent of "upstream" > for these Freescale components. > > Unfortunately, patches there are in a different form, and > I pushed the equivalent forms and they were acked and > promises made by Freescale (Lauren) to include in the next > release: > > https://lists.yoctoproject.org/pipermail/meta-freescale/2014-June/thread.html#8881 Great! We can add this to the patch description: https://lists.yoctoproject.org/pipermail/meta-freescale/2014-June/008993.html where Freescale said "The patch is good. Team said they'll push it into our upcoming beta release." I guess this is easy enough that it does not warrant a resend of the patch, so the two URLs above can be added by whoever commits that to the tree. Thanks! ;-) Regards, Yann E. MORIN.
Dear Gary Bisson, On Thu, 24 Jul 2014 23:43:13 -0700, Gary Bisson wrote: > Same as similar patches for imx-vpu, the returned value from > IOGetVirtMem was not properly checked. > > Signed-off-by: Gary Bisson <bisson.gary@gmail.com> Thanks, patch applied, after adding the reference to the upstream discussion, as suggested by Yann. Thomas
diff --git a/package/libfslvpuwrap/libfslvpuwrap-0001-vpu_wrapper-fix-tests-of-return-value-from-IOGetVirt.patch b/package/libfslvpuwrap/libfslvpuwrap-0001-vpu_wrapper-fix-tests-of-return-value-from-IOGetVirt.patch new file mode 100644 index 0000000..ccd2d9e --- /dev/null +++ b/package/libfslvpuwrap/libfslvpuwrap-0001-vpu_wrapper-fix-tests-of-return-value-from-IOGetVirt.patch @@ -0,0 +1,41 @@ +From 7ca66c10c05168c7d342df7c7a70d4a1ae0629f7 Mon Sep 17 00:00:00 2001 +From: Eric Nelson <eric.nelson@boundarydevices.com> +Date: Sat, 28 Jun 2014 09:45:09 -0700 +Subject: [PATCH] vpu_wrapper: fix tests of return value from IOGetVirtMem + +IOGetVirtMem() returns a pointer or specifically -1 (MAP_FAILED) +in the case of failure. + +Upstream-Status: Pending + +Signed-off-by: Laci Tele <laci@boundarydevices.com> +Signed-off-by: Eric Nelson <eric.nelson@boundarydevices.com> +--- + vpu_wrapper.c | 4 ++-- + 1 file changed, 2 insertions(+), 2 deletions(-) + +diff --git a/vpu_wrapper.c b/vpu_wrapper.c +index 9249174..148c5df 100755 +--- a/vpu_wrapper.c ++++ b/vpu_wrapper.c +@@ -6653,7 +6653,7 @@ VpuDecRetCode VPU_DecGetMem(VpuMemDesc* pInOutMem) + return VPU_DEC_RET_FAILURE; + } + ret=IOGetVirtMem(&buff); +- if(ret==0) //if(ret!=RETCODE_SUCCESS) ++ if(ret == -1) //if(ret==MAP_FAILED) + { + VPU_ERROR("%s: get virtual memory failure: size=%d, ret=%d \r\n",__FUNCTION__,buff.size,ret); + return VPU_DEC_RET_FAILURE; +@@ -8277,7 +8277,7 @@ VpuEncRetCode VPU_EncGetMem(VpuMemDesc* pInOutMem) + return VPU_ENC_RET_FAILURE; + } + ret=IOGetVirtMem(&buff); +- if(ret==0) //if(ret!=RETCODE_SUCCESS) ++ if(ret == -1) //if(ret==MAP_FAILED) + { + VPU_ENC_ERROR("%s: get virtual memory failure: size=%d, ret=%d \r\n",__FUNCTION__,buff.size,(UINT32)ret); + return VPU_ENC_RET_FAILURE; +-- +1.9.1 +
Same as similar patches for imx-vpu, the returned value from IOGetVirtMem was not properly checked. Signed-off-by: Gary Bisson <bisson.gary@gmail.com> --- Yann, Eric, all, This patch is just a copy of what Eric pushed for this package in Yocto. Thanks, Gary --- ...-fix-tests-of-return-value-from-IOGetVirt.patch | 41 ++++++++++++++++++++++ 1 file changed, 41 insertions(+) create mode 100644 package/libfslvpuwrap/libfslvpuwrap-0001-vpu_wrapper-fix-tests-of-return-value-from-IOGetVirt.patch