From 47ee3f1dd93bcbe031539b1ecdaafb44b661c772 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Fri, 26 May 2023 12:34:20 -0700 Subject: x86: re-introduce support for ERMS copies for user space accesses I tried to streamline our user memory copy code fairly aggressively in commit adfcf4231b8c ("x86: don't use REP_GOOD or ERMS for user memory copies"), in order to then be able to clean up the code and inline the modern FSRM case in commit 577e6a7fd50d ("x86: inline the 'rep movs' in user copies for the FSRM case"). We had reports [1] of that causing regressions earlier with blogbench, but that turned out to be a horrible benchmark for that case, and not a sufficient reason for re-instating "rep movsb" on older machines. However, now Eric Dumazet reported [2] a regression in performance that seems to be a rather more real benchmark, where due to the removal of "rep movs" a TCP stream over a 100Gbps network no longer reaches line speed. And it turns out that with the simplified the calling convention for the non-FSRM case in commit 427fda2c8a49 ("x86: improve on the non-rep 'copy_user' function"), re-introducing the ERMS case is actually fairly simple. Of course, that "fairly simple" is glossing over several missteps due to having to fight our assembler alternative code. This code really wanted to rewrite a conditional branch to have two different targets, but that made objtool sufficiently unhappy that this instead just ended up doing a choice between "jump to the unrolled loop, or use 'rep movsb' directly". Let's see if somebody finds a case where the kernel memory copies also care (see commit 68674f94ffc9: "x86: don't use REP_GOOD or ERMS for small memory copies"). But Eric does argue that the user copies are special because networking tries to copy up to 32KB at a time, if order-3 pages allocations are possible. In-kernel memory copies are typically small, unless they are the special "copy pages at a time" kind that still use "rep movs". Link: https://lore.kernel.org/lkml/202305041446.71d46724-yujie.liu@intel.com/ [1] Link: https://lore.kernel.org/lkml/CANn89iKUbyrJ=r2+_kK+sb2ZSSHifFZ7QkPLDpAtkJ8v4WUumA@mail.gmail.com/ [2] Reported-and-tested-by: Eric Dumazet Fixes: adfcf4231b8c ("x86: don't use REP_GOOD or ERMS for user memory copies") Signed-off-by: Linus Torvalds --- arch/x86/lib/copy_user_64.S | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/x86/lib/copy_user_64.S b/arch/x86/lib/copy_user_64.S index 4fc5c2de2de4..01c5de4c279b 100644 --- a/arch/x86/lib/copy_user_64.S +++ b/arch/x86/lib/copy_user_64.S @@ -7,6 +7,8 @@ */ #include +#include +#include #include #include @@ -29,7 +31,7 @@ */ SYM_FUNC_START(rep_movs_alternative) cmpq $64,%rcx - jae .Lunrolled + jae .Llarge cmp $8,%ecx jae .Lword @@ -65,6 +67,12 @@ SYM_FUNC_START(rep_movs_alternative) _ASM_EXTABLE_UA( 2b, .Lcopy_user_tail) _ASM_EXTABLE_UA( 3b, .Lcopy_user_tail) +.Llarge: +0: ALTERNATIVE "jmp .Lunrolled", "rep movsb", X86_FEATURE_ERMS +1: RET + + _ASM_EXTABLE_UA( 0b, 1b) + .p2align 4 .Lunrolled: 10: movq (%rsi),%r8 -- cgit v1.2.3