Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
AMD Zen/Ryzen thread (part 2)
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3, 4, 5, 6, 7  
Reply to topic    Gentoo Forums Forum Index Gentoo Chat
View previous topic :: View next topic  
Author Message
krinn
Watchman
Watchman


Joined: 02 May 2003
Posts: 5972

PostPosted: Fri Nov 24, 2017 12:30 pm    Post subject: Reply with quote

Hu wrote:
To be fair to the vendor, I wouldn't assume incompetence if they can't do this. It could be as simple as that their fulfillment process doesn't support passing the customer's instructions from the point of order all the way to the person who will pull the chip from storage and ship it. It could also be that they have made a business decision that they don't want to go to the extra trouble of passing that request and respecting it in fulfillment. However, regardless of why they cannot / will not handle that type of filtering request, if they don't handle it, then the customer risks getting a bad product and shouldering the burden of correcting it. Hence, if you strongly prefer not to deal with getting a bad CPU, you shouldn't use a non-filtering vendor.

I'm never fair with vendor, they earn their life picking your money :)
If vendor doesn't want extra trouble of passing test (which for them should be looking at die numbering no? or it might even be wrote on the box), then they risk the trouble of having extra trouble of having to handle it thru rma.
So for vendor it's only a good bet to do so, if they refuse to rma the product after.

But if the vendor refuse to rma the product for that problem, it mean the vendor is aware of the problem.
You still see fairness from vendor doing "i'm selling product that might be bad, but if it was, i refuse to handle the rma myself".
Because this create an exception to rma, for a known problem, i wouldn't define this vendor as "fair".
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 39299
Location: 56N 3W

PostPosted: Fri Nov 24, 2017 2:00 pm    Post subject: Reply with quote

krinn,

Vendors (of anything) are only in business to make money.
They don't need to be fair ... only operate within the law.

If vendors refusing RMAs is legal, that's a bad place to be a consumer.

I have had a vendor refer me to the manufacturer in one case in the UK, on the grounds of reducing the turn round time.
It did too. The manufacturer shipped a replacement direct to me, next day.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3105
Location: Gainesville, Florida

PostPosted: Sat Nov 25, 2017 7:59 pm    Post subject: Reply with quote

FWIW, I just got an answer to my question I placed on the Newegg user reviews for R5 1600 Ryzens, and a customer said in their review
they received a UA 1733SUS R5 1600, which was ordered on Nov.20, 2017. That's good news for me- hope mine is the same,
or even more recent. My R7 1700 1733SUS RMA, I got last month is great. I've seen several more reports that the 1733SUS chips are
indeed good. If my ordered R5 1600 also turns out to be 1733SUS, that might indicate at least Newegg is now into more recent stock on the Ryzens. :D

UPDATE: Another Newegg customer says about his new R7 1700:
Quote:
Fixed version 11/8/2017 11:01:03 AM

Pros: For those wondering, I just received mine today (8 November 2017) from the New Jersey warehouse. It has the code "UA 1733 PGS", so it was made during week 33 of 2017 and is the fixed version.

Cons: Newegg should state the dates of the ones that are shipping.
Hopefully, we'll be seeing a lot more reviews on the Ryzen's Newegg is shipping. :D
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.26-r3, gcc-7.2.0- kernel-4.14.5 gentoo USE=experimental
Back to top
View user's profile Send private message
Hu
Moderator
Moderator


Joined: 06 Mar 2007
Posts: 11434

PostPosted: Sun Nov 26, 2017 12:13 am    Post subject: Reply with quote

When the vendor is not the manufacturer, and as in this case the item cannot reasonably be repaired, only replaced or scrapped, there is little value to the customer and none to the vendor in supporting RMA-to-vendor rather than requiring that the user RMA to the manufacturer.

Supposing for the sake of argument that the vendor does not screen for the Ryzen errata (and ignoring whether this is because screening is impossible, too expensive, or just not popular with management), what are the options for the vendor if the customer RMAs a bad CPU to the vendor?
  • Turn around and RMA the chip to the manufacturer. The customer could do this and cut out the middleman.
  • Put the chip back in the warehouse and sell it again, hoping that the next customer doesn't RMA it.
  • Put the chip in the trash bin.
All these options cost the vendor more money than refusing the RMA.

What are the options for the vendor to satisfy the customer upon receipt of the RMA'd chip?
  • RMA the bad chip to the manufacturer. Get a good chip from the manufacturer and send it to the customer. The extra round of handling virtually guarantees this will cost more time than having the customer RMA-to-manufacturer.
  • Pull a chip from the warehouse and send it to the customer. Since the vendor is assumed not to be screening chips, such a chip might itself get RMA'd, costing the vendor more time, money, and goodwill.
  • Issue a full refund, let the customer go buy some other product, possibly from a different vendor.
Again, no upside and plenty of downside.

If RMA were not required by law, you'd only see them offered in cases where the vendor was also willing to offer a comprehensive warranty, because they had statistics showing that they practically never had to make good on it, so it would be a form of advertisement rather than a real consumer protection. I sympathize with the position that being required to RMA to the manufacturer rather than the vendor is more complicated, but in this case, it seems like the right way to handle the situation. If the vendor was screening its shipments and had reasonable confidence that it could find and issue a good chip from its own supplies, vendor RMA would make sense. Here, they don't screen and don't have confidence, so RMA-to-vendor is not a good approach.
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 2038
Location: Illinois, USA

PostPosted: Mon Nov 27, 2017 12:18 am    Post subject: Reply with quote

I talked to my local Fry's and they said I can look at the package but not open it before purchase. My understanding is that the serial number can be read without opening the package.

I just returned today from a long out of state trip on family business that involved spending an order of magnitude more than I plan to on my "dream machine". I'm in no mood for a hassle about returns. Maybe January is better since I don't really want to pull more money out of the IRA this year.

EDIT: Sorry for the typo's. Can't wait until I get something powerful enough to run Linux decently plus run XP in virtualbox. Then I won't need to access the internet from XP anymore.


Last edited by Tony0945 on Mon Nov 27, 2017 1:34 am; edited 3 times in total
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 39299
Location: 56N 3W

PostPosted: Mon Nov 27, 2017 1:07 am    Post subject: Reply with quote

Tony0945,

January sales ... :)
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
Chiitoo
Moderator
Moderator


Joined: 28 Feb 2010
Posts: 1402
Location: Here and Away Again

PostPosted: Tue Nov 28, 2017 12:28 am    Post subject: Reply with quote

wrc1944 wrote:
So my plan would be while the drives in this box are still using the AM3+ mobo is to build a new kernel with "k10" instead of native, and with video, ethernet, usb 3.0, 3.1, etc., and sound support for the new AM4 board, and adjust my make.conf flags to something like:

Code:
CFLAGS="-march=amdfam10 -mtune=generic -O2 -fomit-frame-pointer -pipe"
CXXFLAGS="${CFLAGS}"
CHOST="x86_64-pc-linux-gnu"
ACCEPT_KEYWORDS="~amd64"

USE="mmx sse sse2 sse3 sse4a -ipv6 semantic-desktop gnome icu nspluginwrapper -gnome -qt3support -qt4"

CPU_FLAGS_X86="3dnow 3dnowext mmx mmxext popcnt sse sse2 sse3 sse4a"

Before I switched, I used something like this:

Code:
CFLAGS="-O2 -fomit-frame-pointer -pipe"
CXXFLAGS="${CFLAGS}"
CHOST="x86_64-pc-linux-gnu"

# Phenom
#CPU_FLAGS_X86="3dnow 3dnowext mmx mmxext popcnt sse sse2 sse3 sse4a"
# Ryzen
#CPU_FLAGS_X86="aes avx avx2 f16c fma3 mmx mmxext popcnt sse sse2 sse3 sse4_1 \
#       sse4_2 sse4a ssse3"
# Both
CPU_FLAGS_X86="mmx mmxext popcnt sse sse2 sse3 sse4a"

ACCEPT_KEYWORDS="~amd64"

The kernel I built with the 'CONFIG_GENERIC_CPU' just to be sure it should work okay on the first try.

By the by, you could probably remove those CPU flags from the USE flags, since to my knowledge, they don't exist any longer, at least in any packages I know of (someone correct me if I'm wrong). Might of course be in overlays somewhere still...

Tony0945 wrote:
I talked to my local Fry's and they said I can look at the package but not open it before purchase. My understanding is that the serial number can be read without opening the package.

The supposed numbers for the manufacture date can indeed be read without opening the package, at least on the two different packages I've had (for a Ryzen 1700).

I've tested two from week 22 (if the numbers indeed mean that), and both were no good, so my vendor, or 'supplier' as I like to call them, suggested me to send it to AMD directly. The RMA application went through smoothly without any extra questions or testing, I just need to send the thing back via a service I've never used, so it's taking me a while to get to it...

It does feel a bit 'meh' when it has been over a month since I first ordered the thing, yet I still can't use it (in part probably due to my own fault). Not to mention the 50 EUR discount it has seen since then, but at least I knew what I was getting into. That is to say, I was fully aware that this kind of things had a high chance of occurring, and I simply decided I was done waiting. :]
_________________
Kind Regards,
~ The Noob Unlimited ~

Sore wa sore, kore wa kore.
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3105
Location: Gainesville, Florida

PostPosted: Tue Nov 28, 2017 8:10 am    Post subject: Reply with quote

Did my pre-ryzen hardware switch-over emerge -e @world rebuild on the AM3+ 8320fx using
Code:
CFLAGS="-march=x86-64 -mtune=generic -O2 -fomit-frame-pointer -pipe -mno-fma4 -mno-tbm -mno-xop -mno-lwp"
went without any problems,, rebooted OK, and I then built a new 4.14.2-gentoo kernel with appropriate AM4 hardware flags along with gcc opts X8_64-genric (as a backup kernel which I would hope to be more likely to boot if my other fresh built generics fail).

Newegg just sent me a R5 Ryzen 1600 UA 1737PGS, so that's one less worry about getting a pre-week 25. Of course it's not an RMA, and I'll have to confirm it's indeed a good one, or not. I haven't seen any 1737PGS chip listed/confirmed as bad or good on any relevant forum/website discussion, or in the seemingly abandoned (no recent entries) google-Docs spreadsheet on the subject, so I'll give it it try when my mobo & RAM get here.
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.26-r3, gcc-7.2.0- kernel-4.14.5 gentoo USE=experimental


Last edited by wrc1944 on Tue Nov 28, 2017 4:11 pm; edited 1 time in total
Back to top
View user's profile Send private message
krinn
Watchman
Watchman


Joined: 02 May 2003
Posts: 5972

PostPosted: Tue Nov 28, 2017 2:01 pm    Post subject: Reply with quote

Hu wrote:
Again, no upside and plenty of downside.

No, because you don't take "vendor's fame" in the list.
If the vendor is handling rma with manufacturer, it will sure take more time than customer->manufacturer, but it will keep its fame for customer, that indeed will wait more, but won't do anything except waiting as the vendor handle the thing for him.
While it will cost money to the vendor, it's a good bet because the customer could only be sad against manufacturer, and fully happy by how vendor has fix things for him.
One happy customer is a customer (he might speak about you in good term to his friends, and rebuy from you knowing you will do what it is need to help him. And one unhappy is losing x customers (he won't speak good about yourself to his friends, that will not speak themselves good to their friends...).

Hu: who would you buy from if a friend told you "i have brought ryzen from X, X told me they will fix the issue. I've wait and X has gave me a good ryzen" or "i have brought ryzen from Z and Z told me to i have to handle it myself with amd" (which might more be express in this form "That Z bastard vendor told me they don't gives a shit about what i have brought to them!").
When in real both Z and X have sold your friend a faulty product.
Note also that refund gave the same effect: "i have issue with ryzen and Z offer me a refund" (as customer will knows "if i have a problem with Z, Z will gave my money back". (but vendor->manufacturer handling, will keep customer's money in vendor's pocket).
Back to top
View user's profile Send private message
Hu
Moderator
Moderator


Joined: 06 Mar 2007
Posts: 11434

PostPosted: Wed Nov 29, 2017 3:21 am    Post subject: Reply with quote

That depends on if I expect the RMA process to be smooth and efficient when working directly with the manufacturer. If the manufacturer can complete the RMA at least as smoothly as X, I'd be fine working straight with the manufacturer.

Of course, I wouldn't buy a Ryzen from a non-screening vendor at all right now, because I don't want to waste my time screening for such a widely present and irritating CPU erratum. I expect to need to do some basic testing of any systems I personally assemble, but that's to shake out defective parts that slipped past the manufacturer's general quality controls, not because I plan to buy hardware from a line with known serious errata. :)
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 2038
Location: Illinois, USA

PostPosted: Wed Nov 29, 2017 1:45 pm    Post subject: Reply with quote

Hu wrote:

Of course, I wouldn't buy a Ryzen from a non-screening vendor at all right now, because I don't want to waste my time screening for such a widely present and irritating CPU erratum. I expect to need to do some basic testing of any systems I personally assemble, but that's to shake out defective parts that slipped past the manufacturer's general quality controls, not because I plan to buy hardware from a line with known serious errata. :)

NewEgg had a Black Friday/Cyber Monday deal going (I think it's still on) for the 1800X at $329. That's tempting.
Back to top
View user's profile Send private message
krinn
Watchman
Watchman


Joined: 02 May 2003
Posts: 5972

PostPosted: Wed Nov 29, 2017 2:05 pm    Post subject: Reply with quote

Tony0945 wrote:
NewEgg had a Black Friday/Cyber Monday deal going (I think it's still on) for the 1800X at $329. That's tempting.

That's still high prize for a non-working cpu, i might could do a 1$ for the black Friday deal with some old pentium mmx if you're looking for prize :)
Out of fun, i would really be even more scared by such a deal, what would be best way for neweeg to end their bad stock of cpus if not selling them thru a deal none shall refuse?
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 2038
Location: Illinois, USA

PostPosted: Wed Nov 29, 2017 2:15 pm    Post subject: Reply with quote

I thought of that too, but at least I would get paid $170 for the RMA work instead of nothing.

However, I have had such heavy expenses this year that I will wait until next year and either buy a Ryzen on sale then (better chance of a later one) or the Ryzen 2. However, this whole segfault thing and how it was made light of by AMD and swept under the rug has left a bad taste in my mouth for AMD. It's not the same company that made my k6-3 and Phenom II. Still thinking of kabylake or coffeelake.

EDIT:
My bad, it's $319 https://www.newegg.com/Product/Product.aspx?Item=N82E16819113430
$397 at Amazon https://www.amazon.com/AMD-Ryzen-1800X-Processor-YD180XBCAEWOF/dp/B06W9JXK4G/ref=sr_1_1?s=pc&ie=UTF8&qid=1511969986&sr=1-1&keywords=1800x
$449 at Fry's https://www.frys.com/search?sqxts=1&cat=&nearbyStoreName=false&search_type=regular&query_string=1800x&fq=m10099998%20AMD
Lending credence to the theory that Newegg is dumping old stock.
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3105
Location: Gainesville, Florida

PostPosted: Thu Nov 30, 2017 2:51 pm    Post subject: Reply with quote

FWIW, the R5 1600 1737 PGS Newegg just sent me (2 days ago) has just finished a 1255 package Ryzen/AM4 tuned emerge -e1 @world --keep-going rebuild perfectly.

The new ASrock B350 pro4 came with the P3.0 BIOS (AGESA 1.0.0.6a) so I immediately updated to 3.20 AGESA to 1.0.0.6b. Only problem was my own fault. I forgot to --exclude libreoffice on this box, as i always do it separately due to relatively low disk space. A simple eclean -d and rm -rf /var/tmp/portage/* returned enough space for an emerge --resume --keep-going to finish up the last 224 packages.

Got a kick out of the sort of ironic fact that all my Linux installs had no problems, but my windows 10 pro insider fast-ring disk couldn't even boot up completely, and keeps bailing- looks like a reinstall for that one.

I keep another win10 HOME retail just installed last week which would boot OK, but no LAN driver worked, until I used the AMD driver cd.

However, following the changed hardware routine for new motherboard failed to re-activate. I had installed with an "offline account", and had it activated on the old AM3+ system, so since I have a retail version of win10, guess I'll eventually get it activated. Keeps wanting me to have a microsoft account in order to finish the reactivate for the new mobo/cpu routine Still..... :?
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.26-r3, gcc-7.2.0- kernel-4.14.5 gentoo USE=experimental


Last edited by wrc1944 on Thu Nov 30, 2017 6:37 pm; edited 2 times in total
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 2038
Location: Illinois, USA

PostPosted: Thu Nov 30, 2017 5:18 pm    Post subject: Reply with quote

Win 8 thru 10 - gag me with a spoon! Win 7 barely tolerable. I only have it to run SageTV on an ancient machine in the basement - Athlon64 X2 boots only from IDE and has a SATA1 interface. I'm a pack rat so I have a variety of old IDE drives and even had an early SATA-1 1TB WD drive that holds the recordings. Bought the Win 7 license from a Canadian dealer on ebay for $11 US. That's the total cost for a headless system built out of old odds and ends. It was sad to wipe the old Gentoo installation (with Gnome 2!) for Windows.
Back to top
View user's profile Send private message
gcyoung
Tux's lil' helper
Tux's lil' helper


Joined: 04 Jul 2007
Posts: 116
Location: England

PostPosted: Fri Dec 01, 2017 4:24 pm    Post subject: Ryzen CPU Reply with quote

Hope I'm not intruding into a ryzen 7 gamers' club, but I read a positive review of the ryzen 3 in a Linux magazine, so; wishing to upgrade from my AMD64 dual core processor I bought one and looked around to get some info on settings for it. This forum was the only place I could find with reference to ryzen processors. Elsewhere were just technical performance details.

Anyway, for anyone thinking of getting a ryzen3 I can state that my experience has been positive. I installed a stage 3 setup with conservative settings of '-j4' and 'native' and everything went smoothly:- no problems. I have now just completed an "emerge -e world" on 883 ebuilds with the following settings:

[MAKEOPTS=-J6
CFLAGS="-march=znver1 -O2 -pipe"
CXXFLAGS="${CFLAGS}"
CPU_FLAGS_X86="3dnow 3dnowext 3dnowprefetch mmx mmxext popcnt sse sse2 sse3 sse4a sse4_1 sse4_2 "
]
gcc-4.6.0
kernel-4.12.12

I bought the cheaper ryzen 3 1200 processor, which operates at 3.1 GHz. I believe the 3 1300 is faster at around 4GHz and it can be clocked.
Again, no problems. My old processor also ran at 3.1 GHz but I'm pretty sure the ryzen 'emerged' the system at a lot less than half the time of any previous complete installation.
Back to top
View user's profile Send private message
tekni5
n00b
n00b


Joined: 05 Dec 2017
Posts: 1

PostPosted: Tue Dec 05, 2017 8:15 am    Post subject: Reply with quote

EDIT:

Further analysis and testing indicates that default VDDCR_SOC for 2133 might be too low and may cause false segfaults to arise via kill-ryzen test, when bumping to higher frequency automatic changes in bios increase VDDCR and may alleviate the problem. If you are testing a post 25 week batch, make sure that your RAM settings are correct as you might fail and start doing an RMA when there is no need to do it! Testing on low voltage SOC is not a good idea!

2133 VDDCR_SOC = 0.880V
2133 DRAM Voltage = 1.200V

2933 VDDCR_SOC = 1.096V
2933 DRAM Voltage = 1.368V

------

Been reading this thread and similar for the past few days, I have something interesting to report.

My Ryzen 1700 UA1733PGS segfaults in a minute or two running kill-ryzen under default BIOS x370 Taichi settings with RAM defaulting to 2133MHz, when I select XMP Profile 2933MHz I don't seem to experience a segfault or at least it's significantly delayed, been testing for many hours.

Is it possible that some of the segfaults are caused by improper memory frequency? Especially if you have a CPU which is after week 25.

I highly recommend everyone who still has a Ryzen chip that segfaults, retest their CPUs with different RAM settings and report back, especially higher frequency or frequency your ram is rated at, instead of doing the low frequency defaults that happen on most motherboards.

I posted more info here: https://www.reddit.com/r/Amd/comments/7ho4uv/is_the_culprit_of_linux_segfault_on_ryzen_cpu/
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3105
Location: Gainesville, Florida

PostPosted: Tue Dec 05, 2017 4:57 pm    Post subject: Reply with quote

tekni5,
Thanks very much for posting here on this topic, and the reddit link to your post over there, entitled:
Quote:
Discussion: Is the culprit of Linux segfault on Ryzen CPU low/incorrect memory frequency? Weird results with 1700 UA1733PGS

which already has 25 informed detailed responses.

Most regulars on this thread (and part 1) should find your detailed testing results and commentary very interesting. 8)

My own RMA chip UA1733SUS has been running normally for many days now, mostly on two Gentoo installs with major updates/rebuilds giving it a pretty good testing, plus several other binary distros that needed updating, and a fresh win10 pro installation on an EVO 850 SSD.

XMP, SMT, Opcache, and 16 threads enabled. Really nice seeing 16 threads chewing through a Gentoo emerge -e1 @world rebuild with no problems! :)

Stock BIOS settings, except I did disable C-state and ASLR for the first round of updates/rebuilding.

Addendum: Also just built another Ryzen box (Asrock B350 Pro4 and Ryzen R5 1600 (UA1737PGS), and no segfault,etc. problems running Gentoo ~amd64 on it either.
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.26-r3, gcc-7.2.0- kernel-4.14.5 gentoo USE=experimental
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3105
Location: Gainesville, Florida

PostPosted: Mon Dec 11, 2017 12:20 am    Post subject: Reply with quote

Those Ryzen users of this thread might wish to keep apprised of the PIE/17.0 profile discussion here, https://forums.gentoo.org/viewtopic.php?p=8154394#8154394 as it may become relevant to the problems with pre week 25 cpus or even RMA'ed chips that are having the segfaults and other issues. IMO PIE or -PIE and disabling/enabling ASLR needs some more understanding as to what the implications are going to be when we have to use the new profile and the system must be recompiled with PIE enabled.

If indeed some Ryzens are having problems due to randomization and ASLR, wouldn't PIE be likely to perhaps produce similar problems as ASLR apparently has done? My old R7 1707SUT chip (which I RMA'd for a 1733PGS) definitely was drastically more segfault prone when ASLR was enabled, to the point of being unusable. When I disabled it, segfaults were still ocurring, but not so much as to prevent all updating.

Maybe the RMA'd chips will not be likely to have issues, but I'd think all those thousands of pre week 25's out there might really have trouble with a full-blown PIE compiled system. Hope I'm wrong. Personally, I need a lot more knowledge before I can make an informed conclusion as to go with 17.0 PIE, or hang on to profile 13.0 as long as I can.

Or, maybe go with a profile 17.0 -pie? Hope this is not going to be a problem for Ryzen systems running Gentoo. :roll:
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.26-r3, gcc-7.2.0- kernel-4.14.5 gentoo USE=experimental
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 39299
Location: 56N 3W

PostPosted: Mon Dec 11, 2017 9:46 am    Post subject: Reply with quote

wrc1944,

Building with +pie provides the capability to use ASLR with that code.
If the kernel does not support ASLR, it can't be used.

However, the changed addressing mechanisms in the code to implement PIE will be there and the CPU will need to do the extra work, even though the Randomisation is missing.

Its not clear yet if the PIE mechanism is enough to trigger the early Ryzen problem, in which case, turning off ASLR in the kernel will make no difference - It will still happen.
My money is on that some address bit patterns trigger the problem more than others, so turning off ASLR in the kernel reduces but does not eliminate the problem.

If you have an early faulty Ryzen, RMA it now. Once its a year old AMD will likely be pointing at the warranty ...
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo Chat All times are GMT
Goto page Previous  1, 2, 3, 4, 5, 6, 7
Page 7 of 7

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum