Nov 04, 2015 - 10:16pm
According to the URL ( https://01.org/zh/linuxgraphics/downloads/intel-graphics-installer-linux... ) support OS releases ....it doesn't support CentOS/RHEL 7.0 .
why ? How can I do on our platform now ?
Due to the OS and Platform is fixed ,and what's your suggestion related Graphic driver (BDW-E i915 part) ?
Platfrom : CPU Broadwell-E + PCH Lynx Point SSKU
OS : CentOS 7.0 with Kernel 3.10.0 Text mode.
Becasue we meet some dirver issue on this platfrom as below list, it seemed related I915 driver stability on CentOS 7.0 (kernel 3.10.0) .
Do you have the install guide/stuff of BDW Grapchic (i915) on CentOS/RHEL 7.0 ?
From Jacky Chen
BDW-i915 booting and running issue list -
Issue 1.) Booting stage : Intel I915 driver initial fail message .
[ 3.492960] WARNING: at drivers/gpu/drm/i915/i915_drv.c:452 intel_detect_pch+0x372/0x3e0 [i915]()
[ 3.501995] Modules linked in: i915(+) i2c_algo_bit drm_kms_helper ahci drm libahci libata i2c_core video
Issue 2.) After we booting the OS(CentOS 7.0)and log-in system , then we try to plug and play the DVI or VGA cable .... it will pops some strange message (as below red message) and system sometime will halt .
[ 54.104905] irq 16: nobody cared (try booting with the "irqpoll" option)
[ 54.111654] CPU: 4 PID: 0 Comm: swapper/4 Tainted: GF O-------------- 3.10.0-123.el7.x86_64 #1
[ 54.121257] Hardware name: American Megatrends Inc. MIC-6313/MIC-6313, BIOS 6313V012 09/22/2015
[ 54.129996] ffff88024291dc00 e3c9165ee7ff0646 ffff88024ed03e40 ffffffff815e19ba
[ 54.137660] ffff88024ed03e68 ffffffff810f96e2 ffff88024291dc00 0000000000000010
[ 54.145324] 0000000000000000 ffff88024ed03ea8 ffffffff810f9b02 e3c9165ee7ff0646
[ 54.152990] Call Trace:
[ 54.155475] <IRQ> [<ffffffff815e19ba>] dump_stack+0x19/0x1b
[ 54.161367] [<ffffffff810f96e2>] __report_bad_irq+0x32/0xd0
[ 54.167066] [<ffffffff810f9b02>] note_interrupt+0x132/0x1f0
[ 54.172764] [<ffffffff810f7221>] handle_irq_event_percpu+0xe1/0x1e0
[ 54.179154] [<ffffffff810f735d>] handle_irq_event+0x3d/0x60
[ 54.184852] [<ffffffff810fa7ba>] handle_fasteoi_irq+0x5a/0x100
[ 54.190813] [<ffffffff81014c3f>] handle_irq+0xbf/0x150
[ 54.196078] [<ffffffff815ed78a>] ? atomic_notifier_call_chain+0x1a/0x20
[ 54.202867] [<ffffffff815f434f>] do_IRQ+0x4f/0xf0
[ 54.207752] [<ffffffff815e94ad>] common_interrupt+0x6d/0x6d
[ 54.213502] <EOI> [<ffffffff814834e2>] ? cpuidle_enter_state+0x52/0xc0
[ 54.220397] [<ffffffff81483615>] cpuidle_idle_call+0xc5/0x200
[ 54.226323] [<ffffffff8101bc7e>] arch_cpu_idle+0xe/0x30
[ 54.231725] [<ffffffff810b4725>] cpu_startup_entry+0xf5/0x290
[ 54.237650] [<ffffffff815cfee1>] start_secondary+0x265/0x27b
[ 54.243486] handlers:
[ 54.245852] [<ffffffff813f8060>] usb_hcd_irq
[ 54.250276] Disabling IRQ #16
<<< System HALT ..... >>>
Hi - To answer your question about distribution support - we've got a set budget allocated to us and we've got to decide which distributions to support - and we are also porting some fairly fast-moving components, so there's a limit to how far back we can provide support before the version gap between the 01.org stack and what's in the distribution becomes difficult to bridge. So we target distributions that are a) popular desktop choices (and there's plenty of dicussion and disagreement about that) and b) fast-moving (tend to have very recent kernels and the like).
Given those criteria (which are at least partly subjective), CentOS didn't make the cut.
However that's not to say there's no hope - the 01.org stack contains no proprietary or secret components, it's all out in the open - so if you can upgrade to a more recent kernel you'll automatically get a newer i915 driver - and from the looks of the trace above that may well solve your problem.
I'm not terribly familiar with CentOS but I imagine there's some at least semi-official way to get newer kernels? I did a bit of searching and people seemed to point to http://elrepo.org/