summaryrefslogtreecommitdiff
path: root/arch/mips/kernel
diff options
context:
space:
mode:
authorAlex Smith <alex.smith@imgtec.com>2015-10-21 11:57:44 +0300
committerRalf Baechle <ralf@linux-mips.org>2015-11-11 10:36:41 +0300
commita7f4df4e21dd8a8dab96e88acd2c9c5017b83fc6 (patch)
tree36c00d0b92be2e2a39da15839f4bfe09be72b572 /arch/mips/kernel
parentc0a9f72c156baf1e88c33c6ba4450647af1b8804 (diff)
downloadlinux-a7f4df4e21dd8a8dab96e88acd2c9c5017b83fc6.tar.xz
MIPS: VDSO: Add implementations of gettimeofday() and clock_gettime()
Add user-mode implementations of gettimeofday() and clock_gettime() to the VDSO. This is currently usable with 2 clocksources: the CP0 count register, which is accessible to user-mode via RDHWR on R2 and later cores, or the MIPS Global Interrupt Controller (GIC) timer, which provides a "user-mode visible" section containing a mirror of its counter registers. This section must be mapped into user memory, which is done below the VDSO data page. When a supported clocksource is not in use, the VDSO functions will return -ENOSYS, which causes libc to fall back on the standard syscall path. When support for neither of these clocksources is compiled into the kernel at all, the VDSO still provides clock_gettime(), as the coarse realtime/monotonic clocks can still be implemented. However, gettimeofday() is not provided in this case as nothing can be done without a suitable clocksource. This causes the symbol lookup to fail in libc and it will then always use the standard syscall path. This patch includes a workaround for a bug in QEMU which results in RDHWR on the CP0 count register always returning a constant (incorrect) value. A fix for this has been submitted, and the workaround can be removed after the fix has been in stable releases for a reasonable amount of time. A simple performance test which calls gettimeofday() 1000 times in a loop and calculates the average execution time gives the following results on a Malta + I6400 (running at 20MHz): - Syscall: ~31000 ns - VDSO (GIC): ~15000 ns - VDSO (CP0): ~9500 ns [markos.chandras@imgtec.com: - Minor code re-arrangements in order for mappings to be made in the order they appear to the process' address space. - Move do_{monotonic, realtime} outside of the MIPS_CLOCK_VSYSCALL ifdef - Use gic_get_usm_range so we can do the GIC mapping in the arch/mips/kernel/vdso instead of the GIC irqchip driver] Signed-off-by: Alex Smith <alex.smith@imgtec.com> Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-kernel@vger.kernel.org Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/11338/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Diffstat (limited to 'arch/mips/kernel')
-rw-r--r--arch/mips/kernel/csrc-r4k.c44
-rw-r--r--arch/mips/kernel/vdso.c71
2 files changed, 109 insertions, 6 deletions
diff --git a/arch/mips/kernel/csrc-r4k.c b/arch/mips/kernel/csrc-r4k.c
index e5ed7ada1433..1f910563fdf6 100644
--- a/arch/mips/kernel/csrc-r4k.c
+++ b/arch/mips/kernel/csrc-r4k.c
@@ -28,6 +28,43 @@ static u64 notrace r4k_read_sched_clock(void)
return read_c0_count();
}
+static inline unsigned int rdhwr_count(void)
+{
+ unsigned int count;
+
+ __asm__ __volatile__(
+ " .set push\n"
+ " .set mips32r2\n"
+ " rdhwr %0, $2\n"
+ " .set pop\n"
+ : "=r" (count));
+
+ return count;
+}
+
+static bool rdhwr_count_usable(void)
+{
+ unsigned int prev, curr, i;
+
+ /*
+ * Older QEMUs have a broken implementation of RDHWR for the CP0 count
+ * which always returns a constant value. Try to identify this and don't
+ * use it in the VDSO if it is broken. This workaround can be removed
+ * once the fix has been in QEMU stable for a reasonable amount of time.
+ */
+ for (i = 0, prev = rdhwr_count(); i < 100; i++) {
+ curr = rdhwr_count();
+
+ if (curr != prev)
+ return true;
+
+ prev = curr;
+ }
+
+ pr_warn("Not using R4K clocksource in VDSO due to broken RDHWR\n");
+ return false;
+}
+
int __init init_r4k_clocksource(void)
{
if (!cpu_has_counter || !mips_hpt_frequency)
@@ -36,6 +73,13 @@ int __init init_r4k_clocksource(void)
/* Calculate a somewhat reasonable rating value */
clocksource_mips.rating = 200 + mips_hpt_frequency / 10000000;
+ /*
+ * R2 onwards makes the count accessible to user mode so it can be used
+ * by the VDSO (HWREna is configured by configure_hwrena()).
+ */
+ if (cpu_has_mips_r2_r6 && rdhwr_count_usable())
+ clocksource_mips.archdata.vdso_clock_mode = VDSO_CLOCK_R4K;
+
clocksource_register_hz(&clocksource_mips, mips_hpt_frequency);
sched_clock_register(r4k_read_sched_clock, 32, mips_hpt_frequency);
diff --git a/arch/mips/kernel/vdso.c b/arch/mips/kernel/vdso.c
index 56cc3c4377fb..975e99759bab 100644
--- a/arch/mips/kernel/vdso.c
+++ b/arch/mips/kernel/vdso.c
@@ -12,9 +12,12 @@
#include <linux/elf.h>
#include <linux/err.h>
#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/irqchip/mips-gic.h>
#include <linux/mm.h>
#include <linux/sched.h>
#include <linux/slab.h>
+#include <linux/timekeeper_internal.h>
#include <asm/abi.h>
#include <asm/vdso.h>
@@ -23,7 +26,7 @@
static union mips_vdso_data vdso_data __page_aligned_data;
/*
- * Mapping for the VDSO data pages. The real pages are mapped manually, as
+ * Mapping for the VDSO data/GIC pages. The real pages are mapped manually, as
* what we map and where within the area they are mapped is determined at
* runtime.
*/
@@ -64,25 +67,67 @@ static int __init init_vdso(void)
}
subsys_initcall(init_vdso);
+void update_vsyscall(struct timekeeper *tk)
+{
+ vdso_data_write_begin(&vdso_data);
+
+ vdso_data.xtime_sec = tk->xtime_sec;
+ vdso_data.xtime_nsec = tk->tkr_mono.xtime_nsec;
+ vdso_data.wall_to_mono_sec = tk->wall_to_monotonic.tv_sec;
+ vdso_data.wall_to_mono_nsec = tk->wall_to_monotonic.tv_nsec;
+ vdso_data.cs_shift = tk->tkr_mono.shift;
+
+ vdso_data.clock_mode = tk->tkr_mono.clock->archdata.vdso_clock_mode;
+ if (vdso_data.clock_mode != VDSO_CLOCK_NONE) {
+ vdso_data.cs_mult = tk->tkr_mono.mult;
+ vdso_data.cs_cycle_last = tk->tkr_mono.cycle_last;
+ vdso_data.cs_mask = tk->tkr_mono.mask;
+ }
+
+ vdso_data_write_end(&vdso_data);
+}
+
+void update_vsyscall_tz(void)
+{
+ if (vdso_data.clock_mode != VDSO_CLOCK_NONE) {
+ vdso_data.tz_minuteswest = sys_tz.tz_minuteswest;
+ vdso_data.tz_dsttime = sys_tz.tz_dsttime;
+ }
+}
+
int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
{
struct mips_vdso_image *image = current->thread.abi->vdso;
struct mm_struct *mm = current->mm;
- unsigned long base, vdso_addr;
+ unsigned long gic_size, vvar_size, size, base, data_addr, vdso_addr;
struct vm_area_struct *vma;
+ struct resource gic_res;
int ret;
down_write(&mm->mmap_sem);
- base = get_unmapped_area(NULL, 0, PAGE_SIZE + image->size, 0, 0);
+ /*
+ * Determine total area size. This includes the VDSO data itself, the
+ * data page, and the GIC user page if present. Always create a mapping
+ * for the GIC user area if the GIC is present regardless of whether it
+ * is the current clocksource, in case it comes into use later on. We
+ * only map a page even though the total area is 64K, as we only need
+ * the counter registers at the start.
+ */
+ gic_size = gic_present ? PAGE_SIZE : 0;
+ vvar_size = gic_size + PAGE_SIZE;
+ size = vvar_size + image->size;
+
+ base = get_unmapped_area(NULL, 0, size, 0, 0);
if (IS_ERR_VALUE(base)) {
ret = base;
goto out;
}
- vdso_addr = base + PAGE_SIZE;
+ data_addr = base + gic_size;
+ vdso_addr = data_addr + PAGE_SIZE;
- vma = _install_special_mapping(mm, base, PAGE_SIZE,
+ vma = _install_special_mapping(mm, base, vvar_size,
VM_READ | VM_MAYREAD,
&vdso_vvar_mapping);
if (IS_ERR(vma)) {
@@ -90,8 +135,22 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
goto out;
}
+ /* Map GIC user page. */
+ if (gic_size) {
+ ret = gic_get_usm_range(&gic_res);
+ if (ret)
+ goto out;
+
+ ret = io_remap_pfn_range(vma, base,
+ gic_res.start >> PAGE_SHIFT,
+ gic_size,
+ pgprot_noncached(PAGE_READONLY));
+ if (ret)
+ goto out;
+ }
+
/* Map data page. */
- ret = remap_pfn_range(vma, base,
+ ret = remap_pfn_range(vma, data_addr,
virt_to_phys(&vdso_data) >> PAGE_SHIFT,
PAGE_SIZE, PAGE_READONLY);
if (ret)