[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#1039622: marked as done (bookworm-pu: package nvidia-cuda-toolkit/11.8.0-5~deb12u1)



Your message dated Sat, 22 Jul 2023 13:19:42 +0000
with message-id <E1qNCWM-005rqB-1j@coccia.debian.org>
and subject line Released with 12.1
has caused the Debian Bug report #1039622,
regarding bookworm-pu: package nvidia-cuda-toolkit/11.8.0-5~deb12u1
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@bugs.debian.org
immediately.)


-- 
1039622: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1039622
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems
--- Begin Message ---
Package: release.debian.org
Severity: normal
Tags: bookworm
User: release.debian.org@packages.debian.org
Usertags: pu

[ Reason ]
I'd like to update the openjdk-8 bundled with nvidia-cuda-toolkit to the
latest version from sid, fixing a bunch of CVEs.

[ Impact ]
reduces the number of open CVEs in stable/non-free

[ Tests ]
not really testable due to hardware and driver requirements

[ Risks ]
low, openjdk-8 is ony needed by nvidia-visual-profiler which is (nearly)
a leaf package

[ Checklist ]
  [x] *all* changes are documented in the d/changelog
  [x] I reviewed all changes and I approve them
  [x] attach debdiff against the package in (old)stable
      (does not include the changed component tarballs)
  [x] the issue is verified as fixed in unstable

[ Changes ]
update some component tarballs, they have versions in their name,
therefore that is possible without replacing the whole source package

trivial Linux 6.4 patch for nvidia-fs-dkms

[ Other info ]
n/a


Andreas
diff --git a/debian/changelog b/debian/changelog
index 273dc54..7225656 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,16 @@
+nvidia-cuda-toolkit (11.8.0-5~deb12u1) bookworm; urgency=medium
+
+  * Rebuild for bookworm.
+
+ -- Andreas Beckmann <anbe@debian.org>  Tue, 27 Jun 2023 21:33:39 +0200
+
+nvidia-cuda-toolkit (11.8.0-5) unstable; urgency=medium
+
+  * Fix nvidia-fs kernel module build for Linux 6.4.
+  * Use a snapshot of openjdk-8-jre (8u372-ga-1).
+
+ -- Andreas Beckmann <anbe@debian.org>  Tue, 27 Jun 2023 18:17:09 +0200
+
 nvidia-cuda-toolkit (11.8.0-4) unstable; urgency=medium
 
   * nsight-systems: Fix symlinks to executables.
@@ -10,10 +23,10 @@ nvidia-cuda-toolkit (11.8.0-3) unstable; urgency=medium
 
   * Add clang-14 as supported compiler alternative.
   * Use vendor-specific alternatives for the driver dkms package.
-  * Backport nvfs_io_complete changes from nvidia-fs 2.15 (CUDA 12.1) to fix
-    kernel module build for Linux 5.16.
-  * Backport nvfs_rw_verify_area changes from nvidia-fs 2.15 (CUDA 12.1) to
-    fix kernel module build for Linux 5.18.
+  * Backport nvfs_io_complete changes from nvidia-fs 2.14.14 (CUDA 12.0.1) to
+    fix kernel module build for Linux 5.16.
+  * Backport nvfs_rw_verify_area changes from nvidia-fs 2.14.14 (CUDA 12.0.1)
+    to fix kernel module build for Linux 5.18.
   * Switch from prandom_u32() to get_random_u32() to fix kernel module build
     for Linux 6.1.  (Closes: #1033288)
 
diff --git a/debian/gbp.conf b/debian/gbp.conf
index 37bb286..2f6cb15 100644
--- a/debian/gbp.conf
+++ b/debian/gbp.conf
@@ -10,8 +10,8 @@ component = [
 	'amd64',
 	'ppc64el',
 	'arm64',
-	'openjdk-8-source-8u362-ga-4',
-	'openjdk-8-jre-amd64-8u362-ga-4',
-	'openjdk-8-jre-ppc64el-8u362-ga-4',
+	'openjdk-8-source-8u372-ga-1',
+	'openjdk-8-jre-amd64-8u372-ga-1',
+	'openjdk-8-jre-ppc64el-8u372-ga-1',
 	]
 debian-branch = main
diff --git a/debian/patches/nvidia-fs.patch b/debian/patches/nvidia-fs.patch
index 8951951..44c20a3 100644
--- a/debian/patches/nvidia-fs.patch
+++ b/debian/patches/nvidia-fs.patch
@@ -1,9 +1,9 @@
 Author: Andreas Beckmann <anbe@debian.org>
 Description: backport support for newer kernels
- * backport nvfs_io_complete changes from nvidia-fs 2.15 (CUDA 12.1) to fix
-   kernel module build for Linux 5.16
- * backport nvfs_rw_verify_area changes from nvidia-fs 2.15 (CUDA 12.1) to
-   fix kernel module build for Linux 5.18
+ * backport nvfs_io_complete changes from nvidia-fs 2.14.14 (CUDA 12.0.1) to
+   fix kernel module build for Linux 5.16
+ * backport nvfs_rw_verify_area changes from nvidia-fs 2.14.14 (CUDA 12.0.1)
+   to fix kernel module build for Linux 5.18
  * switch from prandom_u32() to get_random_u32() to fix kernel module build
    for Linux 6.1
 
@@ -86,7 +86,7 @@ Description: backport support for newer kernels
          .owner = THIS_MODULE,
  };
  
-+#if LINUX_VERSION_CODE <  KERNEL_VERSION(6,2,0)
++#if LINUX_VERSION_CODE < KERNEL_VERSION(6,2,0)
  static char *nvfs_devnode(struct device *dev, umode_t *mode)
 +#else
 +static char *nvfs_devnode(const struct device *dev, umode_t *mode)
@@ -94,13 +94,25 @@ Description: backport support for newer kernels
  {
          if (!mode)
                  return NULL;
+@@ -2297,7 +2309,11 @@ static int __init nvfs_init(void)
+ 	pr_info("nvidia_fs: registered correctly with major number %d\n",
+ 			major_number);
+ 
++#if LINUX_VERSION_CODE < KERNEL_VERSION(6,4,0)
+ 	nvfs_class = class_create(THIS_MODULE, CLASS_NAME);
++#else
++	nvfs_class = class_create(CLASS_NAME);
++#endif
+ 
+ 	if (IS_ERR(nvfs_class)) {
+ 		unregister_chrdev(major_number, DEVICE_NAME);
 --- a/nvidia-cuda/nvidia_fs/usr/src/nvidia-fs/nvfs-mmap.c
 +++ b/nvidia-cuda/nvidia_fs/usr/src/nvidia-fs/nvfs-mmap.c
 @@ -598,7 +598,11 @@ static int nvfs_mgroup_mmap_internal(str
          }
  
          /* dont allow mremap to expand and dont allow copy on fork */
-+#if LINUX_VERSION_CODE <  KERNEL_VERSION(6,3,0)
++#if LINUX_VERSION_CODE < KERNEL_VERSION(6,3,0)
          vma->vm_flags |= VM_IO | VM_MIXEDMAP | VM_DONTEXPAND | VM_DONTDUMP | VM_DONTCOPY;
 +#else
 +        vm_flags_set(vma, VM_IO | VM_MIXEDMAP | VM_DONTEXPAND | VM_DONTDUMP | VM_DONTCOPY);
diff --git a/debian/rules.defs b/debian/rules.defs
index e2a3f1f..ad372d0 100644
--- a/debian/rules.defs
+++ b/debian/rules.defs
@@ -111,7 +111,7 @@ nvidia-fs-dkms.VERSION			 = $(CUDA_VERSION_nvidia_fs)
 
 DEFAULT_GCC_VERSION			 = 11
 
-OPENJDK_VERSION.default			 = 8u362-ga-4
+OPENJDK_VERSION.default			 = 8u372-ga-1
 
 UPSTREAM_HOST_ARCH.default		 = $(DEB_HOST_GNU_CPU)
 UPSTREAM_HOST_ARCH.ppc64el		 = ppc64le

--- End Message ---
--- Begin Message ---
Version: 12.1

The upload requested in this bug has been released as part of 12.1.

--- End Message ---

Reply to: