- Linux-stable-mirror - lists.linaro.org

Re: [Linux-stable-mirror] [kaiser-discuss] KPTI backport to 3.16

by Dave Hansen

On 01/08/2018 07:28 AM, Yves-Alexis Perez wrote: > On Mon, 2018-01-08 at 15:19 +0000, Ben Hutchings wrote: >>> I tried this on a ThinkPad X230 (Ivy Bridge CPU, i7 3520M, so pcid but not >>> invpcid) on full UEFI mode and it doesn't boot at all. grub-efi loads the >>> kernel and initrd, but I don't have any message after “Loading initial >>> ramdisk” (even with debug earlyprintk=vga). >> If you boot with EFI you need to use earlyprintk=efi. > Thanks! With that, I can see the last few log lines before it hangs: > > Kernel/User page tables isolation: enabled > […] > bootconsole [earlyefi0] disabled Does "earlyprintk=efi,keep" help get more out to the console? > Also I must have done something wrong earlier today, because when booting with > 'nopcid' it does work (which is consistent with other people tests, afaict). If I had to guess, I'd say we're probably putting a PCID value into the low bits of CR3 before we've _enabled_ PCID support in the CPU.

7 years, 1 month

1
0
0 0

Re: [Linux-stable-mirror] 4.14 LTS build not listed in kernel releases feed

by Greg KH

On Mon, Jan 08, 2018 at 04:19:17PM +0100, Pavlos Parissis wrote: > Hi, > > The 4.14 kernel is listed as a 'Stable' but not as 'Longterm' in the kernel releases feed[1], is > this expected? Yes. A kernel can not become "LTS" until it has a chance to be that, right now it is still just a "Stable" release as 4.15 is not released. thanks, greg k-h

7 years, 1 month

1
0
0 0

Re: [Linux-stable-mirror] [PATCH 4.4 00/37] 4.4.110-stable review

by Guillaume Tucker

On 05/01/18 00:06, Kevin Hilman wrote: > kernelci.org bot <bot(a)kernelci.org> writes: > >> stable-rc/linux-4.4.y boot: 100 boots: 4 failed, 93 passed with 1 offline, 2 conflicts (v4.4.109-38-g99abd6cdd65e) >> >> Full Boot Summary: https://kernelci.org/boot/all/job/stable-rc/branch/linux-4.4.y/kernel/v4.4.… >> Full Build Summary: https://kernelci.org/build/stable-rc/branch/linux-4.4.y/kernel/v4.4.109-38-… >> >> Tree: stable-rc >> Branch: linux-4.4.y >> Git Describe: v4.4.109-38-g99abd6cdd65e >> Git Commit: 99abd6cdd65e984d89c8565508a7a96ea0fce179 >> Git URL: http://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git >> Tested: 53 unique boards, 19 SoC families, 16 builds out of 178 > > TL;DR; All is well. > >> Boot Regressions Detected: >> >> arm: >> >> exynos_defconfig: >> exynos5422-odroidxu3: >> lab-collabora: failing since 58 days (last pass: v4.4.95-21-g32458fcb7bd6 - first fail: v4.4.96-41-g336421367b9c) > > Long standing issue in lab-collabora (passing in other labs) Guillaume? This should be fixed now, with a tweak to the device config to enable relocating the ramdisk and dtb: https://review.linaro.org/#/c/23238/ >> multi_v7_defconfig: >> armada-xp-linksys-mamba: >> lab-free-electrons: new failure (last pass: v4.4.109-36-g8b381424010c) > > Not a kerel issue, bootROM fails to start bootloader. I pinged lab > owners (Free Electrons) > >> tegra124-nyan-big: >> lab-collabora: failing since 1 day (last pass: v4.4.109 - first fail: v4.4.109-36-g8b381424010c) >> >> tegra_defconfig: >> tegra124-nyan-big: >> lab-collabora: failing since 1 day (last pass: v4.4.108-65-g57856049c0f8 - first fail: v4.4.109) > > This one is booting fine, but the command to power-off the board is > timing out, resulting in a failure report. Indeed, this was due to a crash of the lavapdu daemon - it's back on track now. (On a side note, the tegra124-nyan-big is still failing to boot in mainline due to a genuine kernel driver issue.) Guillaume

7 years, 1 month

1
0
0 0

Re: [Linux-stable-mirror] [PATCH 4.14 00/14] 4.14.12-stable review

by Guillaume Tucker

+Sjoerd On 05/01/18 00:12, Kevin Hilman wrote: > kernelci.org bot <bot(a)kernelci.org> writes: > >> stable-rc/linux-4.14.y boot: 118 boots: 4 failed, 113 passed with 1 offline (v4.14.11-15-g732141e47ee6) >> >> Full Boot Summary: https://kernelci.org/boot/all/job/stable-rc/branch/linux-4.14.y/kernel/v4.1… >> Full Build Summary: https://kernelci.org/build/stable-rc/branch/linux-4.14.y/kernel/v4.14.11-15… >> >> Tree: stable-rc >> Branch: linux-4.14.y >> Git Describe: v4.14.11-15-g732141e47ee6 >> Git Commit: 732141e47ee614d70aeb8ad828a977ad19447e87 >> Git URL: http://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git >> Tested: 68 unique boards, 23 SoC families, 16 builds out of 185 >> >> Boot Regressions Detected: > > TL;DR; All is well. > > Same issues as reported for stable-rc/linux-4.4.y, but... > >> arm: >> >> multi_v7_defconfig: >> armada-xp-linksys-mamba: >> lab-free-electrons: new failure (last pass: v4.14.11) >> tegra124-nyan-big: >> lab-collabora: failing since 1 day (last pass: v4.14.10-147-g61f3176f64e3 - first fail: v4.14.11) >> >> omap2plus_defconfig: >> am335x-boneblack: >> lab-collabora: new failure (last pass: v4.14.11) > > ... this one is unique to 4.14, but is still a lab issue: bootloader > fails to DHCP. Turns out this is a known hardware issue with the ethernet phy on this type of board. It's possible to work around it (retry...), we'll take a look. Guillaume

7 years, 1 month

1
0
0 0

[Linux-stable-mirror] [PATCH] media: usbtv: prevent double free in error case

by Oliver Neukum

Quoting the original report: It looks like there is a double-free vulnerability in Linux usbtv driver on an error path of usbtv_probe function. When audio registration fails, usbtv_video_free function ends up freeing usbtv data structure, which gets freed the second time under usbtv_video_fail label. usbtv_audio_fail: usbtv_video_free(usbtv); => v4l2_device_put(&usbtv->v4l2_dev); => v4l2_device_put => kref_put => v4l2_device_release => usbtv_release (CALLBACK) => kfree(usbtv) (1st time) usbtv_video_fail: usb_set_intfdata(intf, NULL); usb_put_dev(usbtv->udev); kfree(usbtv); (2nd time) So, as we have refcounting, use it Reported-by: Yavuz, Tuba <tuba(a)ece.ufl.edu> Signed-off-by: Oliver Neukum <oneukum(a)suse.com> CC: stable(a)vger.kernel.org --- drivers/media/usb/usbtv/usbtv-core.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/media/usb/usbtv/usbtv-core.c b/drivers/media/usb/usbtv/usbtv-core.c index 127f8a0c098b..0c2e628e8723 100644 --- a/drivers/media/usb/usbtv/usbtv-core.c +++ b/drivers/media/usb/usbtv/usbtv-core.c @@ -112,6 +112,8 @@ static int usbtv_probe(struct usb_interface *intf, return 0; usbtv_audio_fail: + /* we must not free at this point */ + usb_get_dev(usbtv->udev); usbtv_video_free(usbtv); usbtv_video_fail: -- 2.13.6

7 years, 1 month

1
0
0 0

[Linux-stable-mirror] patch "iio: chemical: ccs811: Fix output of IIO_CONCENTRATION channels" added to staging-testing

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled iio: chemical: ccs811: Fix output of IIO_CONCENTRATION channels to my staging git tree which can be found at git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging.git in the staging-testing branch. The patch will show up in the next release of the linux-next tree (usually sometime within the next 24 hours during the week.) The patch will be merged to the staging-next branch sometime soon, after it passes testing, and the merge window is open. If you have any questions about this process, please let me know. >From 1a0f51d5aaf71da6090ab839ba507b7df27efcbd Mon Sep 17 00:00:00 2001 From: Narcisa Ana Maria Vasile <narcisaanamaria12(a)gmail.com> Date: Wed, 6 Dec 2017 18:57:58 +0200 Subject: iio: chemical: ccs811: Fix output of IIO_CONCENTRATION channels in_concentration_raw should report, according to sysfs-bus-iio documentation, a "Raw (unscaled no offset etc.) percentage reading of a substance." Modify scale to convert from ppm/ppb to percentage: 1 ppm = 0.0001% 1 ppb = 0.0000001% There is no offset needed to convert the ppm/ppb to percentage, so remove offset from IIO_CONCENTRATION (IIO_MOD_CO2) channel. Cc'd stable to reduce chance of userspace breakage in the long run as we fix this wrong bit of ABI usage. Signed-off-by: Narcisa Ana Maria Vasile <narcisaanamaria12(a)gmail.com> Cc: <Stable(a)vger.kernel.org> Reviewed-by: Matt Ranostay <matt.ranostay(a)konsulko.com> Signed-off-by: Jonathan Cameron <Jonathan.Cameron(a)huawei.com> --- drivers/iio/chemical/ccs811.c | 13 +++---------- 1 file changed, 3 insertions(+), 10 deletions(-) diff --git a/drivers/iio/chemical/ccs811.c b/drivers/iio/chemical/ccs811.c index 97bce8345c6a..fbe2431f5b81 100644 --- a/drivers/iio/chemical/ccs811.c +++ b/drivers/iio/chemical/ccs811.c @@ -96,7 +96,6 @@ static const struct iio_chan_spec ccs811_channels[] = { .channel2 = IIO_MOD_CO2, .modified = 1, .info_mask_separate = BIT(IIO_CHAN_INFO_RAW) | - BIT(IIO_CHAN_INFO_OFFSET) | BIT(IIO_CHAN_INFO_SCALE), .scan_index = 0, .scan_type = { @@ -255,24 +254,18 @@ static int ccs811_read_raw(struct iio_dev *indio_dev, switch (chan->channel2) { case IIO_MOD_CO2: *val = 0; - *val2 = 12834; + *val2 = 100; return IIO_VAL_INT_PLUS_MICRO; case IIO_MOD_VOC: *val = 0; - *val2 = 84246; - return IIO_VAL_INT_PLUS_MICRO; + *val2 = 100; + return IIO_VAL_INT_PLUS_NANO; default: return -EINVAL; } default: return -EINVAL; } - case IIO_CHAN_INFO_OFFSET: - if (!(chan->type == IIO_CONCENTRATION && - chan->channel2 == IIO_MOD_CO2)) - return -EINVAL; - *val = -400; - return IIO_VAL_INT; default: return -EINVAL; } -- 2.15.1

7 years, 1 month

1
0
0 0

[Linux-stable-mirror] patch "iio: adc: stm32: fix scan of multiple channels with DMA" added to staging-testing

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled iio: adc: stm32: fix scan of multiple channels with DMA to my staging git tree which can be found at git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging.git in the staging-testing branch. The patch will show up in the next release of the linux-next tree (usually sometime within the next 24 hours during the week.) The patch will be merged to the staging-next branch sometime soon, after it passes testing, and the merge window is open. If you have any questions about this process, please let me know. >From 69545dfbcaf7a346d87dc6fb6f847b51a1845852 Mon Sep 17 00:00:00 2001 From: Fabrice Gasnier <fabrice.gasnier(a)st.com> Date: Fri, 5 Jan 2018 15:34:54 +0100 Subject: iio: adc: stm32: fix scan of multiple channels with DMA By default, watermark is set to '1'. Watermark is used to fine tune cyclic dma buffer period. In case watermark is left untouched (e.g. 1) and several channels are being scanned, buffer period is wrongly set (e.g. to 1 sample). As a consequence, data is never pushed to upper layer. Fix buffer period size, by taking scan channels number into account. Fixes: 2763ea0585c9 ("iio: adc: stm32: add optional dma support") Signed-off-by: Fabrice Gasnier <fabrice.gasnier(a)st.com> Cc: <Stable(a)vger.kernel.org> Signed-off-by: Jonathan Cameron <Jonathan.Cameron(a)huawei.com> --- drivers/iio/adc/stm32-adc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/iio/adc/stm32-adc.c b/drivers/iio/adc/stm32-adc.c index 6dbf9549cdc9..7f5def465340 100644 --- a/drivers/iio/adc/stm32-adc.c +++ b/drivers/iio/adc/stm32-adc.c @@ -1289,6 +1289,7 @@ static int stm32_adc_set_watermark(struct iio_dev *indio_dev, unsigned int val) { struct stm32_adc *adc = iio_priv(indio_dev); unsigned int watermark = STM32_DMA_BUFFER_SIZE / 2; + unsigned int rx_buf_sz = STM32_DMA_BUFFER_SIZE; /* * dma cyclic transfers are used, buffer is split into two periods. @@ -1297,7 +1298,7 @@ static int stm32_adc_set_watermark(struct iio_dev *indio_dev, unsigned int val) * - one buffer (period) driver can push with iio_trigger_poll(). */ watermark = min(watermark, val * (unsigned)(sizeof(u16))); - adc->rx_buf_sz = watermark * 2; + adc->rx_buf_sz = min(rx_buf_sz, watermark * 2 * adc->num_conv); return 0; } -- 2.15.1

7 years, 1 month

1
0
0 0

[Linux-stable-mirror] [PATCH 11/20] clocksource/drivers/stm32: Fix kernel panic with multiple timers

by Daniel Lezcano

The current code hides a couple of bugs. - The global variable 'clock_event_ddata' is overwritten each time the init function is invoked. This is fixed with a kmemdup instead of assigning the global variable. That prevents a memory corruption when several timers are defined in the DT. - The clockevent's event_handler is NULL if the time framework does not select the clockevent when registering it, this is fine but the init code generates in any case an interrupt leading to dereference this NULL pointer. The stm32 timer works with shadow registers, a mechanism to cache the registers. When a change is done in one buffered register, we need to artificially generate an event to force the timer to copy the content of the register to the shadowed register. The auto-reload register (ARR) is one of the shadowed register as well as the prescaler register (PSC), so in order to force the copy, we issue an event which in turn leads to an interrupt and the NULL dereference. This is fixed by inverting two lines where we clear the status register before enabling the update event interrupt. As this kernel crash is resulting from the combination of these two bugs, the fixes are grouped into a single patch. Cc: stable(a)vger.kernel.org Signed-off-by: Daniel Lezcano <daniel.lezcano(a)linaro.org> Tested-by: Benjamin Gaignard <benjamin.gaignard(a)st.com> Acked-by: Benjamin Gaignard <benjamin.gaignard(a)st.com> --- drivers/clocksource/timer-stm32.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/clocksource/timer-stm32.c b/drivers/clocksource/timer-stm32.c index 8f24237..4bfeb99 100644 --- a/drivers/clocksource/timer-stm32.c +++ b/drivers/clocksource/timer-stm32.c @@ -106,6 +106,10 @@ static int __init stm32_clockevent_init(struct device_node *np) unsigned long rate, max_delta; int irq, ret, bits, prescaler = 1; + data = kmemdup(&clock_event_ddata, sizeof(*data), GFP_KERNEL); + if (!data) + return -ENOMEM; + clk = of_clk_get(np, 0); if (IS_ERR(clk)) { ret = PTR_ERR(clk); @@ -156,8 +160,8 @@ static int __init stm32_clockevent_init(struct device_node *np) writel_relaxed(prescaler - 1, data->base + TIM_PSC); writel_relaxed(TIM_EGR_UG, data->base + TIM_EGR); - writel_relaxed(TIM_DIER_UIE, data->base + TIM_DIER); writel_relaxed(0, data->base + TIM_SR); + writel_relaxed(TIM_DIER_UIE, data->base + TIM_DIER); data->periodic_top = DIV_ROUND_CLOSEST(rate, prescaler * HZ); @@ -184,6 +188,7 @@ static int __init stm32_clockevent_init(struct device_node *np) err_clk_enable: clk_put(clk); err_clk_get: + kfree(data); return ret; } -- 2.7.4

7 years, 1 month

1
0
0 0

[Linux-stable-mirror] Patch "Map the vsyscall page with _PAGE_USER" has been added to the 4.9-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled Map the vsyscall page with _PAGE_USER to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: map-the-vsyscall-page-with-_page_user.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. From: Borislav Petkov <bp(a)suse.de> Date: Thu, 4 Jan 2018 17:42:45 +0100 Subject: Map the vsyscall page with _PAGE_USER From: Borislav Petkov <bp(a)suse.de> This needs to happen early in kaiser_pagetable_walk(), before the hierarchy is established so that _PAGE_USER permission can be really set. A proper fix would be to teach kaiser_pagetable_walk() to update those permissions but the vsyscall page is the only exception here so ... Signed-off-by: Borislav Petkov <bp(a)suse.de> Acked-by: Hugh Dickins <hughd(a)google.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- arch/x86/entry/vsyscall/vsyscall_64.c | 5 +++++ arch/x86/include/asm/vsyscall.h | 2 ++ arch/x86/mm/kaiser.c | 34 ++++++++++++++++++++++++++++++---- 3 files changed, 37 insertions(+), 4 deletions(-) --- a/arch/x86/entry/vsyscall/vsyscall_64.c +++ b/arch/x86/entry/vsyscall/vsyscall_64.c @@ -66,6 +66,11 @@ static int __init vsyscall_setup(char *s } early_param("vsyscall", vsyscall_setup); +bool vsyscall_enabled(void) +{ + return vsyscall_mode != NONE; +} + static void warn_bad_vsyscall(const char *level, struct pt_regs *regs, const char *message) { --- a/arch/x86/include/asm/vsyscall.h +++ b/arch/x86/include/asm/vsyscall.h @@ -12,12 +12,14 @@ extern void map_vsyscall(void); * Returns true if handled. */ extern bool emulate_vsyscall(struct pt_regs *regs, unsigned long address); +extern bool vsyscall_enabled(void); #else static inline void map_vsyscall(void) {} static inline bool emulate_vsyscall(struct pt_regs *regs, unsigned long address) { return false; } +bool vsyscall_enabled(void) { return false; } #endif #endif /* _ASM_X86_VSYSCALL_H */ --- a/arch/x86/mm/kaiser.c +++ b/arch/x86/mm/kaiser.c @@ -19,6 +19,7 @@ #include <asm/pgalloc.h> #include <asm/desc.h> #include <asm/cmdline.h> +#include <asm/vsyscall.h> int kaiser_enabled __read_mostly = 1; EXPORT_SYMBOL(kaiser_enabled); /* for inlined TLB flush functions */ @@ -110,12 +111,13 @@ static inline unsigned long get_pa_from_ * * Returns a pointer to a PTE on success, or NULL on failure. */ -static pte_t *kaiser_pagetable_walk(unsigned long address) +static pte_t *kaiser_pagetable_walk(unsigned long address, bool user) { pmd_t *pmd; pud_t *pud; pgd_t *pgd = native_get_shadow_pgd(pgd_offset_k(address)); gfp_t gfp = (GFP_KERNEL | __GFP_NOTRACK | __GFP_ZERO); + unsigned long prot = _KERNPG_TABLE; if (pgd_none(*pgd)) { WARN_ONCE(1, "All shadow pgds should have been populated"); @@ -123,6 +125,17 @@ static pte_t *kaiser_pagetable_walk(unsi } BUILD_BUG_ON(pgd_large(*pgd) != 0); + if (user) { + /* + * The vsyscall page is the only page that will have + * _PAGE_USER set. Catch everything else. + */ + BUG_ON(address != VSYSCALL_ADDR); + + set_pgd(pgd, __pgd(pgd_val(*pgd) | _PAGE_USER)); + prot = _PAGE_TABLE; + } + pud = pud_offset(pgd, address); /* The shadow page tables do not use large mappings: */ if (pud_large(*pud)) { @@ -135,7 +148,7 @@ static pte_t *kaiser_pagetable_walk(unsi return NULL; spin_lock(&shadow_table_allocation_lock); if (pud_none(*pud)) { - set_pud(pud, __pud(_KERNPG_TABLE | __pa(new_pmd_page))); + set_pud(pud, __pud(prot | __pa(new_pmd_page))); __inc_zone_page_state(virt_to_page((void *) new_pmd_page), NR_KAISERTABLE); } else @@ -155,7 +168,7 @@ static pte_t *kaiser_pagetable_walk(unsi return NULL; spin_lock(&shadow_table_allocation_lock); if (pmd_none(*pmd)) { - set_pmd(pmd, __pmd(_KERNPG_TABLE | __pa(new_pte_page))); + set_pmd(pmd, __pmd(prot | __pa(new_pte_page))); __inc_zone_page_state(virt_to_page((void *) new_pte_page), NR_KAISERTABLE); } else @@ -191,7 +204,7 @@ static int kaiser_add_user_map(const voi ret = -EIO; break; } - pte = kaiser_pagetable_walk(address); + pte = kaiser_pagetable_walk(address, flags & _PAGE_USER); if (!pte) { ret = -ENOMEM; break; @@ -318,6 +331,19 @@ void __init kaiser_init(void) kaiser_init_all_pgds(); + /* + * Note that this sets _PAGE_USER and it needs to happen when the + * pagetable hierarchy gets created, i.e., early. Otherwise + * kaiser_pagetable_walk() will encounter initialized PTEs in the + * hierarchy and not set the proper permissions, leading to the + * pagefaults with page-protection violations when trying to read the + * vsyscall page. For example. + */ + if (vsyscall_enabled()) + kaiser_add_user_map_early((void *)VSYSCALL_ADDR, + PAGE_SIZE, + __PAGE_KERNEL_VSYSCALL); + for_each_possible_cpu(cpu) { void *percpu_vaddr = __per_cpu_user_mapped_start + per_cpu_offset(cpu); Patches currently in stable-queue which might be from bp(a)suse.de are queue-4.9/map-the-vsyscall-page-with-_page_user.patch

7 years, 1 month

1
0
0 0

[Linux-stable-mirror] Patch "proc: much faster /proc/vmstat" has been added to the 4.4-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled proc: much faster /proc/vmstat to the 4.4-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: proc-much-faster-proc-vmstat.patch and it can be found in the queue-4.4 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From 68ba0326b4e14988f9e0c24a6e12a85cf2acd1ca Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan <adobriyan(a)gmail.com> Date: Fri, 7 Oct 2016 17:02:14 -0700 Subject: proc: much faster /proc/vmstat From: Alexey Dobriyan <adobriyan(a)gmail.com> commit 68ba0326b4e14988f9e0c24a6e12a85cf2acd1ca upstream. Every current KDE system has process named ksysguardd polling files below once in several seconds: $ strace -e trace=open -p $(pidof ksysguardd) Process 1812 attached open("/etc/mtab", O_RDONLY|O_CLOEXEC) = 8 open("/etc/mtab", O_RDONLY|O_CLOEXEC) = 8 open("/proc/net/dev", O_RDONLY) = 8 open("/proc/net/wireless", O_RDONLY) = -1 ENOENT (No such file or directory) open("/proc/stat", O_RDONLY) = 8 open("/proc/vmstat", O_RDONLY) = 8 Hell knows what it is doing but speed up reading /proc/vmstat by 33%! Benchmark is open+read+close 1.000.000 times. BEFORE $ perf stat -r 10 taskset -c 3 ./proc-vmstat Performance counter stats for 'taskset -c 3 ./proc-vmstat' (10 runs): 13146.768464 task-clock (msec) # 0.960 CPUs utilized ( +- 0.60% ) 15 context-switches # 0.001 K/sec ( +- 1.41% ) 1 cpu-migrations # 0.000 K/sec ( +- 11.11% ) 104 page-faults # 0.008 K/sec ( +- 0.57% ) 45,489,799,349 cycles # 3.460 GHz ( +- 0.03% ) 9,970,175,743 stalled-cycles-frontend # 21.92% frontend cycles idle ( +- 0.10% ) 2,800,298,015 stalled-cycles-backend # 6.16% backend cycles idle ( +- 0.32% ) 79,241,190,850 instructions # 1.74 insn per cycle # 0.13 stalled cycles per insn ( +- 0.00% ) 17,616,096,146 branches # 1339.956 M/sec ( +- 0.00% ) 176,106,232 branch-misses # 1.00% of all branches ( +- 0.18% ) 13.691078109 seconds time elapsed ( +- 0.03% ) ^^^^^^^^^^^^ AFTER $ perf stat -r 10 taskset -c 3 ./proc-vmstat Performance counter stats for 'taskset -c 3 ./proc-vmstat' (10 runs): 8688.353749 task-clock (msec) # 0.950 CPUs utilized ( +- 1.25% ) 10 context-switches # 0.001 K/sec ( +- 2.13% ) 1 cpu-migrations # 0.000 K/sec 104 page-faults # 0.012 K/sec ( +- 0.56% ) 30,384,010,730 cycles # 3.497 GHz ( +- 0.07% ) 12,296,259,407 stalled-cycles-frontend # 40.47% frontend cycles idle ( +- 0.13% ) 3,370,668,651 stalled-cycles-backend # 11.09% backend cycles idle ( +- 0.69% ) 28,969,052,879 instructions # 0.95 insn per cycle # 0.42 stalled cycles per insn ( +- 0.01% ) 6,308,245,891 branches # 726.058 M/sec ( +- 0.00% ) 214,685,502 branch-misses # 3.40% of all branches ( +- 0.26% ) 9.146081052 seconds time elapsed ( +- 0.07% ) ^^^^^^^^^^^ vsnprintf() is slow because: 1. format_decode() is busy looking for format specifier: 2 branches per character (not in this case, but in others) 2. approximately million branches while parsing format mini language and everywhere 3. just look at what string() does /proc/vmstat is good case because most of its content are strings Link: http://lkml.kernel.org/r/20160806125455.GA1187@p183.telecom.by Signed-off-by: Alexey Dobriyan <adobriyan(a)gmail.com> Cc: Joe Perches <joe(a)perches.com> Cc: Andi Kleen <andi(a)firstfloor.org> Cc: Al Viro <viro(a)zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds(a)linux-foundation.org> Signed-off-by: Mel Gorman <mgorman(a)suse.de> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- mm/vmstat.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) --- a/mm/vmstat.c +++ b/mm/vmstat.c @@ -1351,7 +1351,9 @@ static int vmstat_show(struct seq_file * unsigned long *l = arg; unsigned long off = l - (unsigned long *)m->private; - seq_printf(m, "%s %lu\n", vmstat_text[off], *l); + seq_puts(m, vmstat_text[off]); + seq_put_decimal_ull(m, ' ', *l); + seq_putc(m, '\n'); return 0; } Patches currently in stable-queue which might be from adobriyan(a)gmail.com are queue-4.4/proc-much-faster-proc-vmstat.patch

7 years, 1 month

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror