similar to: [PATCH v1 0/4] GPU Direct RDMA (P2P DMA) for Device Private Pages

From: Yonatan Maman <Ymaman at Nvidia.com> This patch series addresses two critical issues in the Nouveau driver related to device channels, error handling, and sensitive data leaks. - Vulnerability in migrate_to_ram: The migrate_to_ram function might return a dirty HIGH_USER page when a copy push command (FW channel) fails, potentially exposing sensitive data and posing a security

From: Yonatan Maman <Ymaman at Nvidia.com> This patch series addresses two critical issues in the Nouveau driver related to device channels, error handling, and sensitive data leaks. - Vulnerability in migrate_to_ram: The migrate_to_ram function might return a dirty HIGH_USER page when a copy push command (FW channel) fails, potentially exposing sensitive data and posing a security

On 16/10/2024 7:23, Christoph Hellwig wrote: > On Tue, Oct 15, 2024 at 06:23:44PM +0300, Yonatan Maman wrote: >> From: Yonatan Maman <Ymaman at Nvidia.com> >> >> This patch series aims to enable Peer-to-Peer (P2P) DMA access in >> GPU-centric applications that utilize RDMA and private device pages. This >> enhancement is crucial for minimizing data transfer

From: Yonatan Maman <Ymaman at Nvidia.com> When `nouveau_dmem_copy_one` is called, the following error occurs: [272146.675156] nouveau 0000:06:00.0: fifo: PBDMA9: 00000004 [HCE_PRIV] ch 1 00000300 00003386 This indicates that a copy push command triggered a Host Copy Engine Privileged error on channel 1 (Copy Engine channel). To address this issue, modify the Copy Engine channel to allow

This patch series addresses two critical issues in the Nouveau driver related to device channels, error handling and memory leaking. - Memory Leak in migrate_to_ram - the migrate_to_ram function was identified as leaking memory when a copy push command fails. This results in the function returning a dirty HIGH_USER page, which can expose sensitive information and pose a security risk. To mitigate

A copy push command might fail, causing `migrate_to_ram` to return a dirty HIGH_USER page to the user. This exposes a security vulnerability in the nouveau driver. To prevent memory leaks in `migrate_to_ram` upon a copy error, allocate a zero page for the destination page. Signed-off-by: Yonatan Maman <Ymaman at Nvidia.com> Signed-off-by: Gal Shalom <GalShalom at Nvidia.com> ---

? 2024/10/16 17:16, Yonatan Maman ??: > > > On 16/10/2024 7:23, Christoph Hellwig wrote: >> On Tue, Oct 15, 2024 at 06:23:44PM +0300, Yonatan Maman wrote: >>> From: Yonatan Maman <Ymaman at Nvidia.com> >>> >>> This patch series aims to enable Peer-to-Peer (P2P) DMA access in >>> GPU-centric applications that utilize RDMA and private device

When `nouveau_dmem_copy_one` is called, the following error occurs: [272146.675156] nouveau 0000:06:00.0: fifo: PBDMA9: 00000004 [HCE_PRIV] ch 1 00000300 00003386 This indicates that a copy push command triggered a Host Copy Engine Privileged error on channel 1 (Copy Engine channel). To address this issue, modify the Copy Engine channel to allow privileged push commands Fixes: 6de125383a5cc

From: Jason Gunthorpe <jgg at mellanox.com> 8 of the mmu_notifier using drivers (i915_gem, radeon_mn, umem_odp, hfi1, scif_dma, vhost, gntdev, hmm) drivers are using a common pattern where they only use invalidate_range_start/end and immediately check the invalidating range against some driver data structure to tell if the driver is interested. Half of them use an interval_tree, the others

From: Jason Gunthorpe <jgg at mellanox.com> 8 of the mmu_notifier using drivers (i915_gem, radeon_mn, umem_odp, hfi1, scif_dma, vhost, gntdev, hmm) drivers are using a common pattern where they only use invalidate_range_start/end and immediately check the invalidating range against some driver data structure to tell if the driver is interested. Half of them use an interval_tree, the others

I haven't had enough time to fully understand the deferred logic in this change. I spotted one problem, see comments inline. On 2019-10-28 4:10 p.m., Jason Gunthorpe wrote: > From: Jason Gunthorpe <jgg at mellanox.com> > > Of the 13 users of mmu_notifiers, 8 of them use only > invalidate_range_start/end() and immediately intersect the > mmu_notifier_range with some kind

On 10/28/19 1:10 PM, Jason Gunthorpe wrote: > From: Jason Gunthorpe <jgg at mellanox.com> > > 8 of the mmu_notifier using drivers (i915_gem, radeon_mn, umem_odp, hfi1, > scif_dma, vhost, gntdev, hmm) drivers are using a common pattern where > they only use invalidate_range_start/end and immediately check the > invalidating range against some driver data structure to tell

I'm attempting to build an R package for distribution and am working from the directions found at http://www.maths.bris.ac.uk/~maman/computerstuff/Rhelp/Rpackages.html#Wi n-Win I've read through Writing R Extensions and various other "helpful" web sites. I've installed all relevant software (perl, cygwin, mingwin, hhc.exe). Here is what I have done so far: 1) Sourced data

Many DRM drivers implement an encoder with an empty implementation. This patchset adds drm_simple_encoder_init() and drm_simple_encoder_create(), which can be used by drivers instead. Except for the destroy callback, the simple encoder's implementation is empty. The patchset also converts 4 encoder instances to use the simple-encoder helpers. But there are at least 11 other drivers which can

Hi all, I am trying to install a custom package from its source using windows. Using this guide http://www.maths.bris.ac.uk/~maman/computerstuff/Rhelp/Rpackages.html I've gotten to the fifth step, but I get the following error in the command window: "'sh' is not recognized as an internal or external command, operable program or batch file." I saw in the help archives that

hello. I have a query about the Matrix package for R. I wrote some code a while ago using the Matrix package version 1.6.2 with an early version of R, to do some linear least squares for regression: xn [,1] [,2] [,3] [,4] [1,] 1 0.7429352 0.5519528 0.4100652 [2,] 1 0.7443713 0.5540886 0.4124477 [3,] 1 0.7447385 0.5546355 0.4130584 [4,] 1 0.7459597 0.5564558

I'm creating my own package for personal and I'm having trouble getting it to a point where R (v 2.3.1) will recognise it. I've followed two different tutorials for how to create the package structure and the DESCRIPTION file ( http://web.maths.unsw.edu.au/~wand/webcpdg/rpack.html , http://www.maths.bris.ac.uk/~maman/computerstuff/Rhelp/Rpackages.html#Lin-Lin ). I'm still getting

From: Jason Gunthorpe <jgg at mellanox.com> Of the 13 users of mmu_notifiers, 8 of them use only invalidate_range_start/end() and immediately intersect the mmu_notifier_range with some kind of internal list of VAs. 4 use an interval tree (i915_gem, radeon_mn, umem_odp, hfi1). 4 use a linked list of some kind (scif_dma, vhost, gntdev, hmm) And the remaining 5 either don't use

From: Jason Gunthorpe <jgg at mellanox.com> Of the 13 users of mmu_notifiers, 8 of them use only invalidate_range_start/end() and immediately intersect the mmu_notifier_range with some kind of internal list of VAs. 4 use an interval tree (i915_gem, radeon_mn, umem_odp, hfi1). 4 use a linked list of some kind (scif_dma, vhost, gntdev, hmm) And the remaining 5 either don't use

Hello, I wanted to know if it is possible to specify tftp block size > 2k in pxelinux.asm ? In the source of pxelinux.asm, it is said : "The code assumes TFTP_LARGEBLK <= 2K.". What is the limitation ? Would it be possible to increase this value ? Thanks in advance greg Yahoo! Mail : votre e-mail personnel et gratuit qui vous suit partout ! Cr?ez votre Yahoo!