Hello,
first off: Thank you for your hard work with this linux image!
Now to my question: Is it possible to get the eMMC module to work with linux-mainline kernel (4.18-rc8)? I encounter the following error:
I also tried to compile kernel for myself and tried the torvalds 4.19 sources. Both with similar results. 4.4 kernel works without problems.
Is usage of eMMC module currently impossible with mainline kernel?
Edit: There appears to be another kernel error (null pointer reference) in relation to the zram module, when using kernel 4.18. This one seems to appear more or less randomly, but I was able to reliable reproduce it by running 'mkswap' a bunch of times on the /dev/zram* devices. Looks something like this
As long as I remove the eMMC module and disable zram the mainline kernel seems to run very stable for a headless server and I'm quite happy with the USB3.0 performance.
first off: Thank you for your hard work with this linux image!
Now to my question: Is it possible to get the eMMC module to work with linux-mainline kernel (4.18-rc8)? I encounter the following error:
Code:
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: timeout for tag 7
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: ============ CQHCI REGISTER DUMP ===========
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: Caps: 0x00000000 | Version: 0x00000510
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: Config: 0x00000001 | Control: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: Int stat: 0x00000000 | Int enab: 0x00000006
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: Int sig: 0x00000006 | Int Coal: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: TDL base: 0x00000000 | TDL up32: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: Doorbell: 0x00000000 | TCN: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: Dev queue: 0x00000000 | Dev Pend: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: Task clr: 0x00000000 | SSC1: 0x00011000
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: SSC2: 0x00000000 | DCMD rsp: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: RED mask: 0xfdf9a080 | TERRI: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: cqhci: Resp idx: 0x00000000 | Resp arg: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Sys addr: 0x00000000 | Version: 0x00001002
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Argument: 0x00010000 | Trn mode: 0x00000010
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Present: 0x1fff0000 | Host ctl: 0x00000034
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x00000007
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Timeout: 0x0000000e | Int stat: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Int enab: 0x02ff4000 | Sig enab: 0x02ff4000
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: AC12 err: 0x00000000 | Slot int: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x800020f7
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Cmd: 0x00000d1a | Max curr: 0x00000000
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Resp[0]: 0x00000900 | Resp[1]: 0x642017c8
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Resp[2]: 0x4e436172 | Resp[3]: 0x00880103
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: Host ctl2: 0x00000083
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xf711c208
Aug 30 17:56:26 rockpro64 kernel: mmc1: sdhci: ============================================
Aug 30 17:56:26 rockpro64 kernel: mmc1: running CQE recovery
Aug 30 17:56:26 rockpro64 kernel: print_req_error: I/O error, dev mmcblk1boot0, sector 8064I also tried to compile kernel for myself and tried the torvalds 4.19 sources. Both with similar results. 4.4 kernel works without problems.
Is usage of eMMC module currently impossible with mainline kernel?
Edit: There appears to be another kernel error (null pointer reference) in relation to the zram module, when using kernel 4.18. This one seems to appear more or less randomly, but I was able to reliable reproduce it by running 'mkswap' a bunch of times on the /dev/zram* devices. Looks something like this
Code:
Sep 7 16:05:31 rockpro64 kernel: Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
Sep 7 16:05:31 rockpro64 kernel: Mem abort info:
Sep 7 16:05:31 rockpro64 kernel: ESR = 0x96000005
Sep 7 16:05:31 rockpro64 kernel: Exception class = DABT (current EL), IL = 32 bits
Sep 7 16:05:31 rockpro64 kernel: SET = 0, FnV = 0
Sep 7 16:05:31 rockpro64 kernel: EA = 0, S1PTW = 0
Sep 7 16:05:31 rockpro64 kernel: Data abort info:
Sep 7 16:05:31 rockpro64 kernel: ISV = 0, ISS = 0x00000005
Sep 7 16:05:31 rockpro64 kernel: CM = 0, WnR = 0
Sep 7 16:05:31 rockpro64 kernel: user pgtable: 4k pages, 39-bit VAs, pgdp = 00000000604aa84e
Sep 7 16:05:31 rockpro64 kernel: [0000000000000000] pgd=0000000000000000, pud=0000000000000000
Sep 7 16:05:31 rockpro64 kernel: Internal error: Oops: 96000005 [#1] SMP
Sep 7 16:05:31 rockpro64 kernel: Modules linked in: zram xt_iprange iptable_filter xt_owner iptable_mangle ipt_MASQUERADE xt_mark iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack ip_tables x_tables rockchip_rga videobuf2_dma_sg v4l2_mem2mem rockchip_saradc crct10dif_ce dw_hdmi_i2s_audio phy_rockchip_pcie [last unloaded: zram]
Sep 7 16:05:31 rockpro64 kernel: CPU: 0 PID: 3115 Comm: mkswap Tainted: G W 4.18.6-pf5-g45828c4d-dirty #1
Sep 7 16:05:31 rockpro64 kernel: Hardware name: Pine64 RockPro64 (DT)
Sep 7 16:05:31 rockpro64 kernel: pstate: 80000005 (Nzcv daif -PAN -UAO)
Sep 7 16:05:31 rockpro64 kernel: pc : zcomp_compress+0x34/0x60 [zram]
Sep 7 16:05:31 rockpro64 kernel: lr : zram_bvec_rw.isra.26+0x1f8/0x720 [zram]
Sep 7 16:05:31 rockpro64 kernel: sp : ffffffc0e9dbf830
Sep 7 16:05:31 rockpro64 kernel: x29: ffffffc0e9dbf830 x28: 0000000000000000
Sep 7 16:05:31 rockpro64 kernel: x27: 0000000000000000 x26: ffffff8000ff0000
Sep 7 16:05:31 rockpro64 kernel: x25: 000000000040020a x24: ffffffc0e9c25280
Sep 7 16:05:31 rockpro64 kernel: x23: 000000010003864e x22: ffffffc0e67c13c0
Sep 7 16:05:31 rockpro64 kernel: x21: ffffffc0e95f2000 x20: ffffffc0e9dbf8dc
Sep 7 16:05:31 rockpro64 kernel: x19: 0000000000000000 x18: 000000000000002f
Sep 7 16:05:31 rockpro64 kernel: x17: 00000055843c12a8 x16: ffffff80082ae470
Sep 7 16:05:31 rockpro64 kernel: x15: 000000000000013e x14: 0000007f99a4fdd8
Sep 7 16:05:31 rockpro64 kernel: x13: 000000000000000c x12: ffffff80098b2000
Sep 7 16:05:31 rockpro64 kernel: x11: 0000000000000000 x10: 0000000000000000
Sep 7 16:05:31 rockpro64 kernel: x9 : 0000000000000000 x8 : ffffffc0ef7eb870
Sep 7 16:05:31 rockpro64 kernel: x7 : 0000000000000000 x6 : 000000000000003f
Sep 7 16:05:31 rockpro64 kernel: x5 : 0000000000000008 x4 : ffffffc0e9dbf8dc
Sep 7 16:05:31 rockpro64 kernel: x3 : ffffffc0ee4c5bcc x2 : ffffffc0e9dbf8dc
Sep 7 16:05:31 rockpro64 kernel: x1 : ffffffc0e95f2000 x0 : 0000000000002000
Sep 7 16:05:31 rockpro64 kernel: Process mkswap (pid: 3115, stack limit = 0x000000001829a9a6)
Sep 7 16:05:31 rockpro64 kernel: Call trace:
Sep 7 16:05:31 rockpro64 kernel: zcomp_compress+0x34/0x60 [zram]
Sep 7 16:05:31 rockpro64 kernel: zram_bvec_rw.isra.26+0x1f8/0x720 [zram]
Sep 7 16:05:31 rockpro64 kernel: zram_make_request+0x120/0x2fc [zram]
Sep 7 16:05:31 rockpro64 kernel: generic_make_request+0xf4/0x320
Sep 7 16:05:31 rockpro64 kernel: submit_bio+0x44/0x150
Sep 7 16:05:31 rockpro64 kernel: submit_bh_wbc+0x178/0x1c0
Sep 7 16:05:31 rockpro64 kernel: __block_write_full_page+0x230/0x440
Sep 7 16:05:31 rockpro64 kernel: block_write_full_page+0x114/0x130
Sep 7 16:05:31 rockpro64 kernel: blkdev_writepage+0x30/0x40
Sep 7 16:05:31 rockpro64 kernel: __writepage+0x38/0x90
Sep 7 16:05:31 rockpro64 kernel: write_cache_pages+0x1d4/0x420
Sep 7 16:05:31 rockpro64 kernel: generic_writepages+0x48/0x80
Sep 7 16:05:31 rockpro64 kernel: blkdev_writepages+0x28/0x40
Sep 7 16:05:31 rockpro64 kernel: do_writepages+0x3c/0xe0
Sep 7 16:05:31 rockpro64 kernel: __filemap_fdatawrite_range+0xbc/0x110
Sep 7 16:05:31 rockpro64 kernel: file_write_and_wait_range+0x48/0xd0
Sep 7 16:05:31 rockpro64 kernel: blkdev_fsync+0x3c/0x70
Sep 7 16:05:31 rockpro64 kernel: vfs_fsync_range+0x4c/0x90
Sep 7 16:05:31 rockpro64 kernel: do_fsync+0x48/0x90
Sep 7 16:05:31 rockpro64 kernel: sys_fsync+0x24/0x40
Sep 7 16:05:31 rockpro64 kernel: el0_svc_naked+0x30/0x34
Sep 7 16:05:31 rockpro64 kernel: Code: 52840000 b9000280 aa1503e1 aa1403e4 (a9401663)
Sep 7 16:05:31 rockpro64 kernel: ---[ end trace 43b007dfa181aadd ]---As long as I remove the eMMC module and disable zram the mainline kernel seems to run very stable for a headless server and I'm quite happy with the USB3.0 performance.
