search for: _va

At the bottom of page 40 of the Interface manual, an "update_va_mapping" prototype is shown, but it should be an "update_va_mapping_otherdomain" prototype. (This prototype introduces the "update_va_mapping_otherdomain" hypercall documentation, so the association between hypercall and documentation is wrong.) Patch below. - bruce d...

Hi everyone, lately I've been getting familiar with library and working on slight re-layering of the library. It's about having locking layer in public API and tracing one layer below that (let's call it __t_ layer. I'm not very good at making up names, so this is temporary:) ). Then making sure that all generated public stuff call __t_ layer and all other internal stuff

...struct *mm) pmd_t *pmd = kmem_cache_alloc(pmd_cache, GFP_KERNEL); if (!pmd) goto out_oom; + SetPagePDE(virt_to_page(pmd)); set_pgd(&pgd[i], __pgd(1 + __pa(pmd))); } return pgd; out_oom: - for (i--; i >= 0; i--) + for (i--; i >= 0; i--) { + ClearPagePDE(pfn_to_page(pgd_val(pgd[i]) >> PAGE_SHIFT)); kmem_cache_free(pmd_cache, (void *)__va(pgd_val(pgd[i])-1)); + } kmem_cache_free(pgd_cache, pgd); return NULL; } @@ -261,8 +266,10 @@ void pgd_free(pgd_t *pgd) /* in the PAE case user pgd entries are overwritten before usage */ if (PTRS_PER_PMD > 1)...

...struct *mm) pmd_t *pmd = kmem_cache_alloc(pmd_cache, GFP_KERNEL); if (!pmd) goto out_oom; + SetPagePDE(virt_to_page(pmd)); set_pgd(&pgd[i], __pgd(1 + __pa(pmd))); } return pgd; out_oom: - for (i--; i >= 0; i--) + for (i--; i >= 0; i--) { + ClearPagePDE(pfn_to_page(pgd_val(pgd[i]) >> PAGE_SHIFT)); kmem_cache_free(pmd_cache, (void *)__va(pgd_val(pgd[i])-1)); + } kmem_cache_free(pgd_cache, pgd); return NULL; } @@ -261,8 +266,10 @@ void pgd_free(pgd_t *pgd) /* in the PAE case user pgd entries are overwritten before usage */ if (PTRS_PER_PMD > 1)...

...--- linux-2.6.13.orig/include/asm-i386/mach-default/mach_pgalloc.h 2005-08-10 17:11:38.000000000 -0700 +++ linux-2.6.13/include/asm-i386/mach-default/mach_pgalloc.h 2005-08-10 17:11:38.000000000 -0700 @@ -0,0 +1,7 @@ +#ifndef __ASM_MACH_PGALLOC_H +#define __ASM_MACH_PGALLOC_H + +#define SetPagesLDT(_va, _pages) +#define ClearPagesLDT(_va, _pages) + +#endif

...--- linux-2.6.13.orig/include/asm-i386/mach-default/mach_pgalloc.h 2005-08-10 17:11:38.000000000 -0700 +++ linux-2.6.13/include/asm-i386/mach-default/mach_pgalloc.h 2005-08-10 17:11:38.000000000 -0700 @@ -0,0 +1,7 @@ +#ifndef __ASM_MACH_PGALLOC_H +#define __ASM_MACH_PGALLOC_H + +#define SetPagesLDT(_va, _pages) +#define ClearPagesLDT(_va, _pages) + +#endif

Is there any real use in having a ctor/dtor for the pgd cache? Given that all pgd allocation happens via pgd_alloc/pgd_free, why not just fold the [cd]tor in? I'm asking because Xen wants pgd[3] to be unshared in the PAE case, and it looks to me like the easiest way to handle that is by making pgd_alloc/free pv-ops and doing the appropriate thing in the Xen code. Would need to sort out the

Is there any real use in having a ctor/dtor for the pgd cache? Given that all pgd allocation happens via pgd_alloc/pgd_free, why not just fold the [cd]tor in? I'm asking because Xen wants pgd[3] to be unshared in the PAE case, and it looks to me like the easiest way to handle that is by making pgd_alloc/free pv-ops and doing the appropriate thing in the Xen code. Would need to sort out the

This rather complex set of patches allow non-optargs functions to gain optional arguments, while preserving source and binary backwards compatibility. The problem is that we cannot add an optional argument to an existing function. For example, we might want to add flags to the 'lvresize' API which currently has no optional arguments.

...file descriptor). Thought this is mostly usefull when sharing same bo accross different process. I think my radeon virtual address ioclt are nice design but other might disagree. If you want to look at the code : drivers/gpu/drm/radeon/radeon_vm.c drivers/gpu/drm/radeon/radeon_gem.c Grep for _va (virtual address per bo) or _vm (virtual address manager per file descriptor) function name and structure name. On the command stream and bo eviction side everything is as usual on radeon. So a bo can be evicted btw 2 command stream to make room for another one. Either its mapping is invalidated o...

From: "Richard W.M. Jones" <rjones at redhat.com> --- configure.ac | 4 ++++ 1 files changed, 4 insertions(+), 0 deletions(-) diff --git a/configure.ac b/configure.ac index fa97479..6e42423 100644 --- a/configure.ac +++ b/configure.ac @@ -141,6 +141,10 @@ if test "$gl_gcc_warnings" = yes; then # Work around warning in src/inspect.c. This seems to be a bug in gcc

Hello, I am currently looking into ways to support fixed virtual address allocations and sparse mappings in nouveau, as a step towards supporting CUDA. CUDA requires that the GPU virtual address for a given buffer match the CPU virtual address. Therefore, when mapping a CUDA buffer, we have to have a way of specifying a particular virtual address to map to (we would ask that the CPU virtual

Previous discussion here: https://www.redhat.com/archives/libguestfs/2015-June/thread.html#00048 v2: - Use a cleanup handler to release the lock. - Rebase to upstream. Note I have not fixed the problem(s) with error handling (patch 3).

Previously posted in 2015: v1: https://www.redhat.com/archives/libguestfs/2015-June/msg00048.html v2: https://www.redhat.com/archives/libguestfs/2015-June/msg00118.html I have rebased and tidied up the patches, fixing a few spelling mistakes, but they are broadly the same as before. I also ran all the tests, which pass. As with the previous versions, this makes a change to the API, where you

By splitting up large C files we can make parallel compiles a bit faster. Rich.

...or each function. *) - let generate_daemon_stub { name = name; c_name = c_name; + let generate_daemon_stub { name; c_name; style = ret, args, optargs as style } = let errcode = match errcode_of_ret ret with @@ -2072,7 +2072,7 @@ and generate_client_actions_variants () = "; - let generate_va_variants { name = name; c_name = c_name; + let generate_va_variants { name; c_name; style = ret, args, optargs as style } = assert (optargs <> []); (* checked by caller *) @@ -2175,7 +2175,7 @@ and generate_clien...

Per my previous email: https://www.redhat.com/archives/libguestfs/2017-September/msg00203.html I'd like to talk about requiring a more modern version of the OCaml compiler. These commits show some of the code changes which would be possible with OCaml >= 3.12 [which it turns out we already require by accident] and also with OCaml >= 4.02. The latter is my favoured option. Rich.

Previously posted in 2015: v1: https://www.redhat.com/archives/libguestfs/2015-June/msg00048.html v2: https://www.redhat.com/archives/libguestfs/2015-June/msg00118.html This series was posted about 4 weeks ago: v3: https://www.redhat.com/archives/libguestfs/2017-June/msg00287.html There is no change in this series except I rebased it against current upstream head and retested. Last time there

The existing mechanism was clunky, slow and used ~ 10 MB of local disk. Rich.

Hello list. I'm trying to rebuild the 2.6.9.67.0.20.EL kernel, but it fails even without modifications. How did I try it? Created a (non-root) build environment (not a mock ) Installed the kernel.scr.rpm and did a rpmbuild -ba --target=`uname -m` kernel-2.6.spec 2> prep-err.log | tee prep-out.log The build failed at the end: Processing files: kernel-xenU-devel-2.6.9-67.0.20.EL Checking