How to measure the RAM read/write performance
Arun KS
getarunks at gmail.com
Mon Mar 4 01:45:04 EST 2013
Hi Sandeep,
On Sat, Mar 2, 2013 at 12:21 PM, sandeep kumar
<coolsandyforyou at gmail.com> wrote:
>>Another easy way to make memory(ie..pages) non cacheble is use the
>>below function,
>>dma_alloc_coherent(NULL, size, &p, GFP_KERNEL);
>
> I did like what you said. With the timings,i can see it is reading directly
> from RAM. I have some doubts..
> --> What exactly happend here? Here read/write are done by CPU or DMA
> controller?
If you have used the virtual address returned by dma_alloc_coherent,
then obviously it is CPU who has done the read/write.
dma_alloc_coherent allocates memory, sets the page table attributes
for this memory as non cacheable.
When ever CPU access this memory, it generates virtual address, MMU
converts this to physical address using page tables.
Page tables has memory attributes which says that this memory is non-cachable.
Thanks,
Arun
>
> I am new to DMA kind of operations, i have one doubt in how to do DMA?
> Can you give me some example driver code, for this following case....
>
> --> I want to do a DMA from a mem mapped I/O (lets say physical 0x40000000)
> to RAM location @ 0x80000000.
> How do you do this?
>
> Thanks
> Sandeep
>
>
> On Fri, Mar 1, 2013 at 10:45 AM, Arun KS <getarunks at gmail.com> wrote:
>>
>> On Thu, Feb 28, 2013 at 3:54 PM, sandeep kumar
>> <coolsandyforyou at gmail.com> wrote:
>> >>1. use early_param to get the physical start address and size of
>> >> test_region, or you can just ignore this step and hard code to 510M and
>> >> 2M
>> >> for test purpose only.
>> >
>> >>2. use ioremap_nocache() to map this region to a virtual region. note
>> >> that
>> >> this funtion may fail if you are asking a very large vitual memroy
>> >> region.
>> > I did the following things,
>> >
>> > Thank you so much...It worked. With this i am able to measure RAM
>> > performance.
>>
>> Another easy way to make memory(ie..pages) non cacheble is use the
>> below function,
>> dma_alloc_coherent(NULL, size, &p, GFP_KERNEL);
>>
>> This will set the page tables attributes for this page no-cachable.
>> All reads & writes will always happen from main memory because MMU see
>> the memory attribute for this page as non cacheble.
>>
>> Thanks,
>> Arun
>> >
>> >
>> > On Thu, Feb 28, 2013 at 11:39 AM, sandeep kumar
>> > <coolsandyforyou at gmail.com>
>> > wrote:
>> >>
>> >> >1. use early_param to get the physical start address and size of
>> >> > test_region, or you can just ignore this step and hard code to 510M
>> >> > and 2M
>> >> > for test purpose only.
>> >>
>> >> >2. use ioremap_nocache() to map this region to a virtual region. note
>> >> > that this funtion may fail if you are asking a very large vitual
>> >> > memroy
>> >> > region.
>> >> I did the following things,
>> >> 1) Reserved 3MB memory through ATAGS
>> >> 2) Wrote small driver to ioremap that memory in following way,
>> >>
>> >> void *tcpm_base = ioremap_nocache(0x03B00000, SZ_3MB);
>> >> 27 if(tcpm_base!=NULL)
>> >> 28 {
>> >> 29 printk("Jiffies %x %ld\n\n\n\n", jiffies, jiffies);
>> >> 30 for(i=0;i<SZ_4KB;i++)
>> >> 31 src = readl(tcpm_base+i);
>> >> 32 printk("Jiffies %x %ld\n\n\n\n", jiffies, jiffies);
>> >> 33 }
>> >> 34 else
>> >> 35 printk("unable to map 3MB\n");
>> >>
>> >> 3) I am getting the following error,
>> >>
>> >> [ 1.876647] Unable to handle kernel paging request at virtual
>> >> address
>> >> ea82c000
>> >> [ 1.880950] pgd = c0004000
>> >> [ 1.883636] [ea82c000] *pgd=49818811, *pte=00000000, *ppte=00000000
>> >> [ 1.889892] Internal error: Oops: 7 [#1] PREEMPT
>> >> [ 1.894500] Modules linked in:
>> >> [ 1.897521] CPU: 0 Not tainted (3.0.31-g1080f34-dirty #106)
>> >> [ 1.903442] PC is at sand_misc_init+0x4c/0xac
>> >> [ 1.907775] LR is at sand_misc_init+0x3c/0xac
>> >> [ 1.912109] pc : [<c0022ee0>] lr : [<c0022ed0>] psr: 80000013
>> >> [ 1.912139] sp : e982bf98 ip : 00000000 fp : 00000000
>> >> [ 1.923553] r10: 00000000 r9 : 00000000 r8 : 00000000
>> >> [ 1.928771] r7 : 00000000 r6 : c00461b4 r5 : ea828000 r4 :
>> >> 00000000
>> >> [ 1.935272] r3 : 00003fff r2 : 00003ffd r1 : c07ea2cf r0 :
>> >> 00000063
>> >> [ 1.941802] Flags: Nzcv IRQs on FIQs on Mode SVC_32 ISA ARM
>> >> Segment kernel
>> >> [ 1.949096] Control: 10c57c7d Table: 00404059 DAC: 00000015
>> >> [ 1.954803]
>> >> [ 1.954833] PC: 0xc0022e60:
>> >> [ 1.959075] 2e60 c09105f0 c0366554 c07ea27e c07ea296 e3a01000
>> >> e92d4010
>> >> e1a02001 eb0a9acd
>> >> [ 1.967224] 2e80 e59f3008 e5830008 e3a00001 e8bd8010 c0dd8380
>> >> e92d4037
>> >> e3a04000 e3a0063b
>> >> [ 1.975402] 2ea0 e3a01901 e1a02004 e5cd4007 e5cd4006 eb00b9aa
>> >> e2505000
>> >> 0a000019 e59f3070
>> >> [ 1.983551] 2ec0 e59f0070 e5931000 e5932000 eb18a072 e58d4000
>> >> e3033fff
>> >> ea000007 e59d2000
>> >> [ 1.991699] 2ee0 e7952002 f57ff04f e6ef2072 e5cd2007 e59d2000
>> >> e2822001
>> >> e58d2000 e59d2000
>> >> [ 1.999877] 2f00 e1520003 dafffff4 e59f3024 e59f0024 e5931000
>> >> e5932000
>> >> e28dd00c e8bd4030
>> >> [ 2.008026] 2f20 ea18a05d e59f0010 e28dd00c e8bd4030 ea18a059
>> >> c08ea600
>> >> c07ea2cf c07ea2e2
>> >> [ 2.016174] 2f40 e59f3040 e3a01000 e92d4010 e59f0038 e5932000
>> >> eb0a845a
>> >> e59f3030 e3500000
>> >> [ 2.024353]
>> >> [ 2.024353] LR: 0xc0022e50:
>> >> [ 2.028594] 2e50 e3e00015 e8bd81fc c0dd8380 c07ea279 c09105f0
>> >> c0366554
>> >> c07ea27e c07ea296
>> >> [ 2.036773] 2e70 e3a01000 e92d4010 e1a02001 eb0a9acd e59f3008
>> >> e5830008
>> >> e3a00001 e8bd8010
>> >> [ 2.044921] 2e90 c0dd8380 e92d4037 e3a04000 e3a0063b e3a01901
>> >> e1a02004
>> >> e5cd4007 e5cd4006
>> >> [ 2.053070] 2eb0 eb00b9aa e2505000 0a000019 e59f3070 e59f0070
>> >> e5931000
>> >> e5932000 eb18a072
>> >> [ 2.061248] 2ed0 e58d4000 e3033fff ea000007 e59d2000 e7952002
>> >> f57ff04f
>> >> e6ef2072 e5cd2007
>> >> [ 2.069396] 2ef0 e59d2000 e2822001 e58d2000 e59d2000 e1520003
>> >> dafffff4
>> >> e59f3024 e59f0024
>> >> [ 2.077575] 2f10 e5931000 e5932000 e28dd00c e8bd4030 ea18a05d
>> >> e59f0010
>> >> e28dd00c e8bd4030
>> >> [ 2.085723] 2f30 ea18a059 c08ea600 c07ea2cf c07ea2e2 e59f3040
>> >> e3a01000
>> >> e92d4010 e59f0038
>> >> [ 2.093872]
>> >> [ 2.093902] SP: 0xe982bf18:
>> >> [ 2.098144] bf18 382e3120 30393536 00205d35 000000d0 00004fff
>> >> 00003b00
>> >> 192d8000 ea82bfff
>> >> [ 2.106292] bf38 03b00000 ffffffff e982bf84 c00461b4 00000000
>> >> c0044dac
>> >> 00000063 c07ea2cf
>> >> [ 2.114440] bf58 00003ffd 00003fff 00000000 ea828000 c00461b4
>> >> 00000000
>> >> 00000000 00000000
>> >> [ 2.122619] bf78 00000000 00000000 00000000 e982bf98 c0022ed0
>> >> c0022ee0
>> >> 80000013 ffffffff
>> >> [ 2.130767] bf98 00003ffd 0000a27e 00000000 c0037b4c c0022e94
>> >> c003f3fc
>> >> e9814a80 00373231
>> >> [ 2.138946] bfb8 00000000 00000000 00000000 00000236 c0037b4c
>> >> c003818c
>> >> c00461b4 00000013
>> >> [ 2.147094] bfd8 00000000 00000000 00000000 c0008374 00000000
>> >> c0008300
>> >> c00461b4 c00461b4
>> >> [ 2.155242] bff8 00000000 00000000 00000000 00000001 00000000
>> >> e9817940
>> >> c08e8ef4 00000000
>> >> [ 2.163421]
>> >> [ 2.163421] R1: 0xc07ea24f:
>> >> [ 2.167663] a24c 65207265 726f7272 6f6c6220 25206b63 000a646c
>> >> 706f6f6c
>> >> 26006425 3e2d6f6c
>> >> [ 2.175842] a26c 635f6f6c 6d5f6c74 78657475 6f6f6c00 363c0070
>> >> 6f6f6c3e
>> >> 6d203a70 6c75646f
>> >> [ 2.183990] a28c 6f6c2065 64656461 363c000a 6f6f6c3e 6f203a70
>> >> 6f207475
>> >> 656d2066 79726f6d
>> >> [ 2.192138] a2ac 6162000a 6e696b63 69665f67 7300656c 6c657a69
>> >> 74696d69
>> >> 74756100 656c636f
>> >> [ 2.200317] a2cc 4a007261 69666669 25207365 6c252078 0a0a0a64
>> >> 6e75000a
>> >> 656c6261 206f7420
>> >> [ 2.208465] a2ec 2070616d 0a424d33 656d7000 65725f6d 6e6f6967
>> >> 333c0073
>> >> 656d703e 7325286d
>> >> [ 2.216613] a30c 736b3a29 635f7465 74616572 6e615f65 64615f64
>> >> 61662064
>> >> 000a6c69 703e343c
>> >> [ 2.224792] a32c 3a6d656d 6d6f7320 69687465 6920676e 65762073
>> >> 77207972
>> >> 676e6f72 6f79202c
>> >> [ 2.232940] a34c 72612075 6c632065 6e69736f 20612067 62206d76
>> >> 696b6361
>> >> 6120676e 6c61206e
>> >> [ 2.241119]
>> >> [ 2.241119] R5: 0xea827f80:
>> >> [ 2.245361] 7f80 ******** ******** ******** ******** ********
>> >> ********
>> >> ******** ********
>> >> [ 2.253509] 7fa0 ******** ******** ******** ******** ********
>> >> ********
>> >> ******** ********
>> >> [ 2.261688] 7fc0 ******** ******** ******** ******** ********
>> >> ********
>> >> ******** ********
>> >> [ 2.269836] 7fe0 ******** ******** ******** ******** ********
>> >> ********
>> >> ******** ********
>> >> [ 2.278015] 8000 00000000 00000000 00000000 00000000 00000000
>> >> 00000000
>> >> 00000000 00000000
>> >> [ 2.286163] 8020 00000000 00000000 00000000 00000000 00000000
>> >> 00000000
>> >> 00000000 00000000
>> >> [ 2.294311] 8040 00000000 00000000 00000000 00000000 00000000
>> >> 00000000
>> >> 00000000 00000000
>> >> [ 2.302490] 8060 00000000 00000000 00000000 00000000 00000000
>> >> 00000000
>> >> 00000000 00000000
>> >> [ 2.310638]
>> >> [ 2.310638] R6: 0xc0046134:
>> >> [ 2.314880] 6134 eb038dff eb02a1f9 e1a03007 e1a00005 e1a02006
>> >> eb025566
>> >> e59f3008 e5834000
>> >> [ 2.323059] 6154 e8bd41f0 ea0259b3 c09059d4 c08fbcd4 c064e808
>> >> c0791c0d
>> >> c0958220 e59fc020
>> >> [ 2.331207] 6174 e92d4007 e59f301c e59cc000 e1a02001 e59f1014
>> >> e58dc000
>> >> eb055aab e3a00000
>> >> [ 2.339385] 6194 e8bd800e c0958220 c0791c27 c0791c1b e121f007
>> >> e1a00004
>> >> e1a0e006 e1a0f005
>> >> [ 2.347534] 61b4 eb02959a e320f000 e59f300c e5932000 e2822001
>> >> e5832000
>> >> e12fff1e c0958224
>> >> [ 2.355682] 61d4 e59f300c e5932000 e2422001 e5832000 e12fff1e
>> >> c0958224
>> >> e12fff1e e12fff1e
>> >> [ 2.363861] 61f4 e12fff1e e12fff1e eafffffe e59f3014 e92d4010
>> >> e5933004
>> >> e3530000 08bd8010
>> >> [ 2.372009] 6214 e12fff33 e8bd8010 c0958224 e59f3014 e1a01000
>> >> e92d4010
>> >> e5d30000 e1a0e00f
>> >> [ 2.380187] Process swapper (pid: 1, stack limit = 0xe982a2e8)
>> >> [ 2.385986] Stack: (0xe982bf98 to 0xe982c000)
>> >> [ 2.390319] bf80:
>> >> 00003ffd 0000a27e
>> >> [ 2.398498] bfa0: 00000000 c0037b4c c0022e94 c003f3fc e9814a80
>> >> 00373231
>> >> 00000000 00000000
>> >> [ 2.406646] bfc0: 00000000 00000236 c0037b4c c003818c c00461b4
>> >> 00000013
>> >> 00000000 00000000
>> >> [ 2.414825] bfe0: 00000000 c0008374 00000000 c0008300 c00461b4
>> >> c00461b4
>> >> 00000000 00000000
>> >> [ 2.423004] [<c0022ee0>] (sand_misc_init+0x4c/0xac) from
>> >> [<c003f3fc>]
>> >> (do_one_initcall+0xd0/0x1a4)
>> >> [ 2.431915] [<c003f3fc>] (do_one_initcall+0xd0/0x1a4) from
>> >> [<c0008374>]
>> >> (kernel_init+0x74/0x118)
>> >> [ 2.440673] [<c0008374>] (kernel_init+0x74/0x118) from [<c00461b4>]
>> >> (kernel_thread_exit+0x0/0x8)
>> >> [ 2.449462] Code: e58d4000 e3033fff ea000007 e59d2000 (e7952002)
>> >> [ 2.455566] ---[ end trace f76f3c76dcb9b9ef ]---
>> >> [ 2.460144] Kernel panic - not syncing: Attempted to kill init!
>> >> [ 2.466064] [<c004ad10>] (unwind_backtrace+0x0/0x12c) from
>> >> [<c064af70>]
>> >> (panic+0x90/0x1bc)
>> >> [ 2.474304] [<c064af70>] (panic+0x90/0x1bc) from [<c00eb8dc>]
>> >> (do_exit+0xb8/0x734)
>> >> [ 2.481842] [<c00eb8dc>] (do_exit+0xb8/0x734) from [<c0048f8c>]
>> >> (die+0x208/0x23c)
>> >> [ 2.489318] [<c0048f8c>] (die+0x208/0x23c) from [<c004e158>]
>> >> (__do_kernel_fault+0x64/0x84)
>> >> [ 2.497558] [<c004e158>] (__do_kernel_fault+0x64/0x84) from
>> >> [<c004e3e0>] (do_page_fault+0x268/0x288)
>> >> [ 2.506683] [<c004e3e0>] (do_page_fault+0x268/0x288) from
>> >> [<c003f270>]
>> >> (do_DataAbort+0x34/0x94)
>> >> [ 2.515350] [<c003f270>] (do_DataAbort+0x34/0x94) from [<c0044dac>]
>> >> (__dabt_svc+0x4c/0x60)
>> >> [ 2.523590] Exception stack(0xe982bf50 to 0xe982bf98)
>> >> [ 2.528625] bf40: 00000063
>> >> c07ea2cf
>> >> 00003ffd 00003fff
>> >> [ 2.536804] bf60: 00000000 ea828000 c00461b4 00000000 00000000
>> >> 00000000
>> >> 00000000 00000000
>> >> [ 2.544952] bf80: 00000000 e982bf98 c0022ed0 c0022ee0 80000013
>> >> ffffffff
>> >> [ 2.551544] [<c0044dac>] (__dabt_svc+0x4c/0x60) from [<c0022ee0>]
>> >> (sand_misc_init+0x4c/0xac)
>> >> [ 2.559967] [<c0022ee0>] (sand_misc_init+0x4c/0xac) from
>> >> [<c003f3fc>]
>> >> (do_one_initcall+0xd0/0x1a4)
>> >> [ 2.568908] [<c003f3fc>] (do_one_initcall+0xd0/0x1a4) from
>> >> [<c0008374>]
>> >> (kernel_init+0x74/0x118)
>> >> [ 2.577667] [<c0008374>] (kernel_init+0x74/0x118) from [<c00461b4>]
>> >> (kernel_thread_exit+0x0/0x8)
>> >>
>> >>
>> >> Any idea what went wrong.
>> >> I am sure about the ioremap() start address, thats what i reserved in
>> >> ATAGS.
>> >>
>> >> Thanks
>> >> Sandeep
>> >>
>> >>
>> >> On Thu, Feb 28, 2013 at 10:30 AM, sandeep kumar
>> >> <coolsandyforyou at gmail.com> wrote:
>> >>>
>> >>> >1. use early_param to get the physical start address and size of
>> >>> > test_region, or you can just ignore this step and hard code to 510M
>> >>> > and 2M
>> >>> > for test purpose only.
>> >>>
>> >>> >2. use ioremap_nocache() to map this region to a virtual region. note
>> >>> > that this funtion may fail if you are asking a very large vitual
>> >>> > memroy
>> >>> > region.
>> >>>
>> >>> Sounds good, i am gonna try this and let you know.. :)
>> >>>
>> >>>
>> >>> On Wed, Feb 27, 2013 at 8:19 PM, buyitian <buyit at live.cn> wrote:
>> >>>>
>> >>>> ----------------------------------------
>> >>>> > From: buyit at live.cn
>> >>>> > To: coolsandyforyou at gmail.com; kernelnewbies at kernelnewbies.org
>> >>>> > Subject: RE: How to measure the RAM read/write performance
>> >>>> > Date: Wed, 27 Feb 2013 22:33:15 +0800
>> >>>> > CC: dhylands at gmail.com
>> >>>> >
>> >>>> > ________________________________
>> >>>> > > From: coolsandyforyou at gmail.com
>> >>>> > > Date: Tue, 26 Feb 2013 17:01:54 +0530
>> >>>> > > Subject: How to measure the RAM read/write performance
>> >>>> > > To: kernelnewbies at kernelnewbies.org
>> >>>> > > CC: dhylands at gmail.com
>> >>>> > >
>> >>>> > > Hi All
>> >>>> > > In performance benchmark tools, When we profile read/write
>> >>>> > > timings
>> >>>> > > mostly, those read/writes are done to cache only.
>> >>>> > >
>> >>>> > > I want to measure my DDR(RAM chip) performance.
>> >>>> > > So i want to make sure, every read/write should happen to DDR RAM
>> >>>> > > chip only.
>> >>>> > >
>> >>>> > > How can i achieve this...Any ideas/suggestions...?
>> >>>> >
>> >>>> > try to reserve a large region from bootloader(L4 in Qualcomm
>> >>>> > platform), let's say it is 10MB continuous physical memory.
>> >>>>
>> >>>> sorry, to be accurate, reserve physical memory is done by kernel
>> >>>> cmdline, this cmdline parameter can be passed from L4 to kernel, or
>> >>>> configed
>> >>>> by kernel itself.
>> >>>> the cmdline will be like below:
>> >>>> mem=510M at 0 test_region=2M at 510M
>> >>>>
>> >>>> above example tells kernel you have totally 512MB physical memory,
>> >>>> but
>> >>>> kernel will only use the first 510MB, the latter 2MB memory is used
>> >>>> by you.
>> >>>> how to map and use this region depends on you.
>> >>>>
>> >>>> > in kernel, map this region to an continuous virtual region, note
>> >>>> > that
>> >>>> > the pgprot should be uncachable since you want to test without
>> >>>> > cache.
>> >>>>
>> >>>> 1. use early_param to get the physical start address and size of
>> >>>> test_region, or you can just ignore this step and hard code to 510M
>> >>>> and 2M
>> >>>> for test purpose only.
>> >>>>
>> >>>> 2. use ioremap_nocache() to map this region to a virtual region. note
>> >>>> that this funtion may fail if you are asking a very large vitual
>> >>>> memroy
>> >>>> region.
>> >>>>
>> >>>> > once you configed like this, you can read/write to this vitual
>> >>>> > region
>> >>>> > without data cache invovled.
>> >>>> >
>> >>>> > >
>> >>>> > > --
>> >>>> > > With regards,
>> >>>> > > Sandeep Kumar Anantapalli,
>> >>>> > >
>> >>>> > > _______________________________________________ Kernelnewbies
>> >>>> > > mailing
>> >>>> > > list Kernelnewbies at kernelnewbies.org
>> >>>> > > http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>> >>>> > _______________________________________________
>> >>>> > Kernelnewbies mailing list
>> >>>> > Kernelnewbies at kernelnewbies.org
>> >>>> > http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>> >>>
>> >>>
>> >>>
>> >>>
>> >>> --
>> >>> With regards,
>> >>> Sandeep Kumar Anantapalli,
>> >>
>> >>
>> >>
>> >>
>> >> --
>> >> With regards,
>> >> Sandeep Kumar Anantapalli,
>> >
>> >
>> >
>> >
>> > --
>> > With regards,
>> > Sandeep Kumar Anantapalli,
>> >
>> > _______________________________________________
>> > Kernelnewbies mailing list
>> > Kernelnewbies at kernelnewbies.org
>> > http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>> >
>
>
>
>
> --
> With regards,
> Sandeep Kumar Anantapalli,
More information about the Kernelnewbies
mailing list