x86/sgx: Do not update sgx_nr_free_pages in sgx_setup_epc_section()
Jarkko Sakkinen
maintainerPATCH 1/1 x86/sgx: Do not update sgx_nr_free_pages in sgx_setup_epc_section()
NUMA patches introduced this change to __sgx_sanitize_pages():
- if (!ret)
- list_move(&page->list, §ion->page_list);
- else
+ if (!ret) {
+ /*
+ * page is now sanitized. Make it available via the SGX
+ * page allocator:
+ */
+ list_del(&page->list);
+ sgx_free_epc_page(page);
+ } else {
+ /* The page is not yet clean - move to the dirty list. */
list_move_tail(&page->list, &dirty);
-
- spin_unlock(§ion->lock);
+ }
This was done for the reason that it is best to keep the logic to assign available-for-use EPC pages to the correct NUMA lists in a single location.
The problem is that the sgx_nr_free_pages is also incremented by sgx_free_epc_pages(), and thus it ends up having double the number of pages available.
The count was even before NUMA patches kind of out-of-sync, i.e. free pages count was incremented before putting them to the free list, but it didn't matter that much, because sanitization is fairly fast and it only prevented ksgxd to trigger small time after the system had powered on.
---
v2:
* Wrote more verbose and detailed description what is going on.
* Split out from the patches. This is urgent - the attributes can wait.
arch/x86/kernel/cpu/sgx/main.c | 1 - 1 file changed, 1 deletion(-)
arch/x86/kernel/cpu/sgx/main.c +0 -1
@@ -657,7 +657,6 @@ static bool __init sgx_setup_epc_section(u64 phys_addr, u64 size,list_add_tail (§ion->pages[i].list, &sgx_dirty_page_list); } - sgx_nr_free_pages += nr_pages;return true; }
Signature · 1 line
2.31.1
The following commit has been merged into the x86/sgx branch of tip:
The following commit has been merged into the x86/sgx branch of tip:
Commit-ID: ae40aaf6bdbf0354a75b8284a0de453fcf5f4d32 Gitweb: https://git.kernel.org/tip/ae40aaf6bdbf0354a75b8284a0de453fcf5f4d32 Author: Jarkko Sakkinen <[email protected]> AuthorDate: Thu, 08 Apr 2021 12:29:24 +03:00 Committer: Borislav Petkov <[email protected]> CommitterDate: Thu, 08 Apr 2021 17:24:42 +02:00
x86/sgx: Do not update sgx_nr_free_pages in sgx_setup_epc_section()
The commit in Fixes: changed the SGX EPC page sanitization to end up in sgx_free_epc_page() which puts clean and sanitized pages on the free list.
This was done for the reason that it is best to keep the logic to assign available-for-use EPC pages to the correct NUMA lists in a single location.
sgx_nr_free_pages is also incremented by sgx_free_epc_pages() but those pages which are being added there per EPC section do not belong to the free list yet because they haven't been sanitized yet - they land on the dirty list first and the sanitization happens later when ksgxd starts massaging them.
So remove that addition there and have sgx_free_epc_page() do that solely.
[ bp: Sanitize commit message too. ]
--- arch/x86/kernel/cpu/sgx/main.c | 1 - 1 file changed, 1 deletion(-)
arch/x86/kernel/cpu/sgx/main.c +0 -1
@@ -656,7 +656,6 @@ static bool __init sgx_setup_epc_section(u64 phys_addr, u64 size,list_add_tail (§ion->pages[i].list, &sgx_dirty_page_list); } - sgx_nr_free_pages += nr_pages;return true; }