From 4f423ddf56e5ecb1fb2eac83b8e228e3d0aae0f6 Mon Sep 17 00:00:00 2001 From: Frithiof Jensen Date: Mon, 12 Feb 2007 00:53:07 -0800 Subject: [PATCH] [PATCH] EDAC: Add memory scrubbing controls API to core This is an attempt of providing an interface for memory scrubbing control in EDAC. This patch modifies the EDAC Core to provide the Interface for memory controller modules to implment. The following things are still outstanding: - K8 is the first implemenation, The patch provide a method of configuring the K8 hardware memory scrubber via the 'mcX' sysfs directory. There should be some fallback to a generic scrubber implemented in software if the hardware does not support scrubbing. Or .. the scrubbing sysfs entry should not be visible at all. - Only works with SDRAM, not cache, The K8 can scrub cache and l2cache also - but I think this is not so useful as the cache is busy all the time (one hopes). One would also expect that cache scrubbing requires hardware support. - Error Handling, I would like that errors are returned to the user in "terms of file system". - Presentation, I chose Bandwidth in Bytes/Second as a representation of the scrubbing rate for the following reasons: I like that the sysfs entries are sort-of textual, related to something that makes sense instead of magical values that must be looked up. "My People" wants "% main memory scrubbed per hour" others prefer "% memory bandwidth used" as representation, "bandwith used" makes it easy to calculate both versions in one-liner scripts. If one later wants to scrub cache, the scaling becomes wierd for K8 changing from "blocks of 64 byte memory" to "blocks of 64 cache lines" to "blocks of 64 bit". Using "bandwidth used" makes sense in all three cases, (I.M.O. anyway ;-). - Discovery, There is no way to discover the possible settings and what they do without reading the code and the documentation. *I* do not know how to make that work in a practical way. - Bugs(??), other tools can set invalid values in the memory scrub control register, those will read back as '-1', requiring the user to reset the scrub rate. This is how *I* think it should be. - Afflicting other areas of code, I made changes to edac_mc.c and edac_mc.h which will show up globally - this is not nice, it would be better that the memory scrubbing fuctionality and interface could be entirely contained within the memory controller it applies to. Frithiof Jensen edac_mc.c and its .h file is a CORE helper module for EDAC driver modules. This provides the abstraction for device specific drivers. It is fine to modify this CORE to provide help for new features of the the drivers doug thompson Signed-off-by: Frithiof Jensen Signed-off-by: doug thompson Acked-by: Alan Cox Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- Documentation/drivers/edac/edac.txt | 16 ++++++++- drivers/edac/edac_mc.c | 55 +++++++++++++++++++++++++++++ drivers/edac/edac_mc.h | 12 +++++++ 3 files changed, 82 insertions(+), 1 deletion(-) diff --git a/Documentation/drivers/edac/edac.txt b/Documentation/drivers/edac/edac.txt index 7b3d969d2964..3c5a9e4297b4 100644 --- a/Documentation/drivers/edac/edac.txt +++ b/Documentation/drivers/edac/edac.txt @@ -339,7 +339,21 @@ Device Symlink: 'device' - Symlink to the memory controller device + Symlink to the memory controller device. + +Sdram memory scrubbing rate: + + 'sdram_scrub_rate' + + Read/Write attribute file that controls memory scrubbing. The scrubbing + rate is set by writing a minimum bandwith in bytes/sec to the attribute + file. The rate will be translated to an internal value that gives at + least the specified rate. + + Reading the file will return the actual scrubbing rate employed. + + If configuration fails or memory scrubbing is not implemented, the value + of the attribute file will be -1. diff --git a/drivers/edac/edac_mc.c b/drivers/edac/edac_mc.c index 1b4fc9221803..1df8c0c5d40a 100644 --- a/drivers/edac/edac_mc.c +++ b/drivers/edac/edac_mc.c @@ -927,6 +927,57 @@ static ssize_t mci_reset_counters_store(struct mem_ctl_info *mci, return count; } +/* memory scrubbing */ +static ssize_t mci_sdram_scrub_rate_store(struct mem_ctl_info *mci, + const char *data, size_t count) +{ + u32 bandwidth = -1; + + if (mci->set_sdram_scrub_rate) { + + memctrl_int_store(&bandwidth, data, count); + + if (!(*mci->set_sdram_scrub_rate)(mci, &bandwidth)) { + edac_printk(KERN_DEBUG, EDAC_MC, + "Scrub rate set successfully, applied: %d\n", + bandwidth); + } else { + /* FIXME: error codes maybe? */ + edac_printk(KERN_DEBUG, EDAC_MC, + "Scrub rate set FAILED, could not apply: %d\n", + bandwidth); + } + } else { + /* FIXME: produce "not implemented" ERROR for user-side. */ + edac_printk(KERN_WARNING, EDAC_MC, + "Memory scrubbing 'set'control is not implemented!\n"); + } + return count; +} + +static ssize_t mci_sdram_scrub_rate_show(struct mem_ctl_info *mci, char *data) +{ + u32 bandwidth = -1; + + if (mci->get_sdram_scrub_rate) { + if (!(*mci->get_sdram_scrub_rate)(mci, &bandwidth)) { + edac_printk(KERN_DEBUG, EDAC_MC, + "Scrub rate successfully, fetched: %d\n", + bandwidth); + } else { + /* FIXME: error codes maybe? */ + edac_printk(KERN_DEBUG, EDAC_MC, + "Scrub rate fetch FAILED, got: %d\n", + bandwidth); + } + } else { + /* FIXME: produce "not implemented" ERROR for user-side. */ + edac_printk(KERN_WARNING, EDAC_MC, + "Memory scrubbing 'get' control is not implemented!\n"); + } + return sprintf(data, "%d\n", bandwidth); +} + /* default attribute files for the MCI object */ static ssize_t mci_ue_count_show(struct mem_ctl_info *mci, char *data) { @@ -1033,6 +1084,9 @@ MCIDEV_ATTR(ce_noinfo_count,S_IRUGO,mci_ce_noinfo_show,NULL); MCIDEV_ATTR(ue_count,S_IRUGO,mci_ue_count_show,NULL); MCIDEV_ATTR(ce_count,S_IRUGO,mci_ce_count_show,NULL); +/* memory scrubber attribute file */ +MCIDEV_ATTR(sdram_scrub_rate,S_IRUGO|S_IWUSR,mci_sdram_scrub_rate_show,mci_sdram_scrub_rate_store); + static struct mcidev_attribute *mci_attr[] = { &mci_attr_reset_counters, &mci_attr_mc_name, @@ -1042,6 +1096,7 @@ static struct mcidev_attribute *mci_attr[] = { &mci_attr_ce_noinfo_count, &mci_attr_ue_count, &mci_attr_ce_count, + &mci_attr_sdram_scrub_rate, NULL }; diff --git a/drivers/edac/edac_mc.h b/drivers/edac/edac_mc.h index a1cfd4e3c97d..c41986a9ed5b 100644 --- a/drivers/edac/edac_mc.h +++ b/drivers/edac/edac_mc.h @@ -315,6 +315,18 @@ struct mem_ctl_info { unsigned long scrub_cap; /* chipset scrub capabilities */ enum scrub_type scrub_mode; /* current scrub mode */ + /* Translates sdram memory scrub rate given in bytes/sec to the + internal representation and configures whatever else needs + to be configured. + */ + int (*set_sdram_scrub_rate) (struct mem_ctl_info *mci, u32 *bw); + + /* Get the current sdram memory scrub rate from the internal + representation and converts it to the closest matching + bandwith in bytes/sec. + */ + int (*get_sdram_scrub_rate) (struct mem_ctl_info *mci, u32 *bw); + /* pointer to edac checking routine */ void (*edac_check) (struct mem_ctl_info * mci); /* -- 2.30.2