Bug #215

BAD Signature error when upgrading package from edge

Added by Anonymous over 2 years ago. Updated almost 2 years ago.

Status:Closed Start date:12/09/2009
Priority:Normal Due date:
Assignee:Timo Teräs % Done:

90%

Category:Base libraries
Target version:Alpine 1.10.3

Description

After installing alpine 1.9.1 091021 on a USB stick, I changed my repository to an 'edge' mirror. When installing a package (asterisk) from that mirror, the following error appears:
ERROR: ncurses-5.7-r1: BAD signature
I've used all mirrors listed on the wiki, and the results are the same.

ncurses-http - strace apk add ncurses (using http://dl-3...v1.10/main) (75.2 kB) Nathan Angelacos, 05/29/2010 01:23 am

ncurses-local - strace apk add packages/main/ncurses-5.7-r1.apk (local responsitory) (56.4 kB) Nathan Angelacos, 05/29/2010 01:23 am

Associated revisions

Revision c6da7228
Added by Timo Teräs almost 2 years ago

main/openssl: update padlock sha patches

New version of padlock sha patches that do not use the seg.fault handler trick.
It requires application to properly use oneshot mode context flag, or the high
level full operation methods to take use of VIA C7 SHA acceleration. VIA Nano
support is included in this patch and supports the partial transforms, so it
gets accelerated always.

Fixes #215.

History

Updated by Natanael Copa over 2 years ago

which hardware do you use? We have seen issues similar this with padlock hardware. We have not been able to reproduce it thought.

Updated by Natanael Copa over 2 years ago

err.... did you do apk update before trying install anything?

Updated by Anonymous over 2 years ago

I'm using the via 1.2ghz processor. I did do an apk update before installing anything.

Updated by David Williams over 2 years ago

I just tried this on another (identical) server (with the same usb key) and the problem did not happen. So it must be a hardware problem.

Updated by David Williams over 2 years ago

We pulled the hardware out and noticed that it is the VIA 1.6 Ghz processor that was failing, not the 1.2 Ghz. So it may very well be a problem with the VIA 1.6 chip. It is the only one we have with that processor.

Updated by Natanael Copa over 2 years ago

  • Status changed from New to Rejected

i close this since its most likely something with your hardware. Please reopen if you get same problem on other boxes.

Updated by Nathan Angelacos almost 2 years ago

  • Status changed from Rejected to Feedback
  • Assignee set to Natanael Copa
  • Target version changed from Alpine 1.9 to Alpine 1.10.3

I can confirm this problem happens with a Via NANO 1.6GHz only when using remote repositories; I got the BAD signature error, and could not install a package even with --force --allow-untrusted.

However, doing a "fetch" then an apk add works fine

Since this particular machine has 4GB RAM, copied the entire 1.10 repository to /tmpfs, and can install packages without problems.

It appears that this is some kind of problem with the interaction between network activity and apk.

The processor info:

  1. cat /proc/cpuinfo
    processor : 0
    vendor_id : CentaurHauls
    cpu family : 6
    model : 15
    model name : VIA Nano processor L2200@1600MHz
    stepping : 2
    cpu MHz : 1615.998
    cache size : 1024 KB
    fdiv_bug : no
    hlt_bug : no
    f00f_bug : no
    coma_bug : no
    fpu : yes
    fpu_exception : yes
    cpuid level : 10
    wp : yes
    flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush acpi mmx fxsr sse sse2 ss tm syscall nx fxsr_opt rdtscp lm constant_tsc up rep_good pni monitor vmx est tm2 ssse3 cx16 xtpr rng rng_en ace ace_en ace2 phe phe_en lahf_lm
    bogomips : 3233.66
    clflush size : 64
    cache_alignment : 128
    address sizes : 40 bits physical, 48 bits virtual
    power management:

lsmod:

lsmod
Module Size Used by Tainted: G
ext3 86276 0
mbcache 3176 1 ext3
jbd 25612 1 ext3
af_packet 11768 0
serio_raw 2400 0
pcspkr 924 0
shpchp 19368 0
pci_hotplug 16544 1 shpchp
i2c_viapro 3976 0
viafb 56960 0
fb 28468 1 viafb
cfbfillrect 2148 1 viafb
cfbimgblt 1548 1 viafb
cfbcopyarea 2432 1 viafb
i2c_algo_bit 3192 1 viafb
i2c_core 9980 3 i2c_viapro,viafb,i2c_algo_bit
snd_hda_codec_via 17200 1
snd_hda_intel 14788 0
snd_hda_codec 33600 2 snd_hda_codec_via,snd_hda_intel
snd_pcm 34964 2 snd_hda_intel,snd_hda_codec
snd_page_alloc 4304 2 snd_hda_intel,snd_pcm
snd_timer 10608 1 snd_pcm
snd_hwdep 3244 1 snd_hda_codec
snd 28052 5 snd_hda_intel,snd_hda_codec,snd_pcm,snd_timer,snd_hwdep
soundcore 2784 1 snd
via_velocity 17492 0
bitrev 736 1 via_velocity
crc_ccitt 916 1 via_velocity
via_agp 4056 1
agpgart 16476 1 via_agp
evdev 5004 0
rtc_cmos 5496 0
rtc_core 8112 1 rtc_cmos
rtc_lib 1072 1 rtc_core
thermal 7780 0
button 2960 0
processor 23600 0
vfat 5652 0
fat 32884 1 vfat
nls_iso8859_1 2760 1
isofs 22400 1
sata_via 4948 0
ehci_hcd 24104 0
uhci_hcd 14516 0
ata_generic 1796 0
pata_via 5252 0
pata_acpi 1548 0
libata 102820 4 sata_via,ata_generic,pata_via,pata_acpi
sr_mod 8764 1
cdrom 22788 1 sr_mod
usb_storage 21712 1
usb_libusual 7504 1 usb_storage
usbcore 79740 5 ehci_hcd,uhci_hcd,usb_storage,usb_libusual
nls_base 4112 5 vfat,fat,nls_iso8859_1,isofs,usbcore
sd_mod 17944 0
scsi_mod 54908 4 libata,sr_mod,usb_storage,sd_mod
cramfs 36788 1
loop 8292 2

will try to do strace on apk later if possible.

Updated by Nathan Angelacos almost 2 years ago

Updating with strace reports.

/etc/apk/repositories: http://dl-3.alpinelinux.org/alpine/v1.10/packages/main

  • strace apk add ncurses 2>ncurses-http
  • strace apk add packages/main/ncurses-5.7-r1.apk 2>ncurses-local

The one pulling the config from http has 3 segfaults. That's probably not a good thing

Updated by Natanael Copa almost 2 years ago

  • Assignee changed from Natanael Copa to Timo Teräs

Updated by Timo Teräs almost 2 years ago

Nathan Angelacos wrote:

  • strace apk add ncurses 2>ncurses-http
  • strace apk add packages/main/ncurses-5.7-r1.apk 2>ncurses-local

The one pulling the config from http has 3 segfaults. That's probably not a good thing

It's the way the padlock sha1/sha256 acceleration works. Nano has better stuff in it though. I should update the openssl padlock patch to support nano in better way. Local repositories work because then everything is mmaped and the sha stuff works differently.

Updated by Timo Teräs almost 2 years ago

  • Status changed from Feedback to Assigned
  • % Done changed from 0 to 90

I've been able to reproduce and fix the issue. I just need to polish the openssl patches until pushing them. I expect this to happen tomorrow.

Updated by Nathan Angelacos almost 2 years ago

  • Status changed from Assigned to Closed

The patches fix the problem. Thanks!

Also available in: Atom PDF