summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2017-10-26libc: wunlock() part 2cinap_lenrek
the initial issue was that wunlock() would wakeup readers while holding the spinlock causing deadlock in libthread programs where rendezvous() would do a thread switch within the same process which then can acquire the RWLock again. the first fix tried to prevent holding the spinlock, waking up one reader at a time with releasing an re-acquiering the spinlock. this violates the invariant that readers can only wakup writers in runlock() when multiple readers where queued at the time of wunlock(). at the first wakeup, q->head != nil so runlock() would find a reader queued on runlock() when it expected a writer. this (hopefully last) fix unlinks *all* the reader QLp's atomically and in order while holding the spinlock and then traverses the dequeued chain of QLp structures again to call rendezvous() so the invariant described above holds.
2017-10-23upas/smtpd: don't call syslog() from the note handler, this can deadlockcinap_lenrek
when the alarm hits while the process is currently in syslog(), holding the sl lock, then calling syslog again will deadlock: /proc/1729193/text:386 plan 9 executable /sys/lib/acid/port /sys/lib/acid/386 acid: lstk() sleep()+0x7 /sys/src/libc/9syscall/sleep.s:5 lock(lk=0x394d8)+0xb7 /sys/src/libc/port/lock.c:25 i=0x3e8 syslog(logname=0x41c64,cons=0x0,fmt=0x41c6a)+0x2d /sys/src/libc/9sys/syslog.c:60 err=0x79732f27 d=0x0 ctim=0x0 buf=0x0 p=0x0 arg=0x0 n=0x0 catchalarm(msg=0xdfffc854)+0x7a /sys/src/cmd/upas/smtp/smtpd.c:71 notifier+0x30 /sys/src/libc/port/atnotify.c:15
2017-10-20libc: cleanup atexit and put exits() in its own compilation unitcinap_lenrek
this avoids having to pull in atexit() and its dependencies (lock(), unlock()) into every program. (as exits() is called by _main() from main9.s).
2017-10-20vt: block when sending input to host (fixes truncated paste)cinap_lenrek
2017-10-17libsec: make sectorNumber argument for aes_xts routines uvlongcinap_lenrek
2017-10-17libsec: add AES CFB and AES OFB stream cipherscinap_lenrek
2017-10-17aux/wpa: prevent PTK re-installation attack by replaying AP retransmitscinap_lenrek
this implements the mitigation suggested in section "6.5 Countermeasures" of "Key Reinstallation Attacks: Forcing Nonce Resuse in WPA2" [1]. [1] https://papers.mathyvanhoef.com/ccs2017.pdf
2017-10-16ape/libsec: fix the build, bring ape libsec.h in sync with plan9 versioncinap_lenrek
2017-10-06ssh: remove extern declarations for pkcs1padbuf() and asn1encodedigest() ↵cinap_lenrek
(now in libsec.h)
2017-10-06rsa: add auth/rsa2asn1, check write error in auth/rsa2x509 and auth/rsa2pub, ↵cinap_lenrek
document in rsa(8)
2017-10-06libsec: export asn1encodedigest(), asn1encodeRSApub(), asn1toRSApub(), ↵cinap_lenrek
pkcs1padbuf() and pkcs1unpadbuf()
2017-10-05libsec: allow \r\n terminated lines in decodePEM()cinap_lenrek
2017-10-04check for fs= in netauditaiju
2017-10-04hgwebfs: simplify retry loop constructioncinap_lenrek
2017-10-04ndb/cs: icmp only supports version 4 addresses, icmpv6 only version 6 addressescinap_lenrek
2017-10-04libauthsrv: preserve readcons() error message from read() errorcinap_lenrek
2017-10-04hg: disable tag caching, allows accessing hg repo from dumpcinap_lenrek
2017-09-299boot: limit read size to 4K for efi simple file system protocolcinap_lenrek
copying files from the uefi shell works, reading plan9.ini works, loading the kernel by calling Read to read in the DATA section of the kernel *FAILS*. my guess is that uefi filesystem driver or nvme driver tries to allocate a temporary buffer and hasnt got the space. limiting the read size fixes it.
2017-09-27pc64: add ether82598 driver to configurationcinap_lenrek
2017-09-27ether82598: support for T540-T1, use physical addresses for isaconf portcinap_lenrek
reading mac doesnt work yet, requires ea= option in isaconf.
2017-09-27sdnvme: identify namespace list fails on intel ssd, just assume nsid=[1]cinap_lenrek
2017-09-27kernel: don't tokenize inplace in isaconfig() to make /dev/reboot workcinap_lenrek
2017-09-27kernel: get rid of 36 bit Paerange mask in mtrr (supporting machines with ↵cinap_lenrek
more than 64GB of memory)
2017-09-23upas/fs: fix putcache(), sub-part messages should never go into the lrucinap_lenrek
we accidentally added non-top messages (attatchments) to the lru, resulting in attachments to be freed. this is wrong.
2017-09-22xhci: do bounds checking in capability walking, check if controller vanished ↵cinap_lenrek
on init (thunderbolt unplug)
2017-09-22sshfs: use mtime for qid.vers, fix wstat without name change, fix wstat ↵cinap_lenrek
memory leak
2017-09-13upas/fs: replace fixed cache table with lru linked listcinap_lenrek
the cachetab just keeps track of recent messages that have not been called cachefree() on. under some conditions, the fixed table could overflow (all messages having refs > 0). with a linked list, overflow becomes non fatal and the algorithm is simpler to implement.
2017-09-11winwatch: show windows with empty labels (thanks jpm)cinap_lenrek
2017-09-10rename pcf kernel to pc, remove pcf, pccpuf, pccpu64 kernels, update ↵cinap_lenrek
documentation there isnt much of a point in keep maintaining separate kernel configurations for terminal and cpu kernels as the role can be switched with service=cpu boot parameter. to make stuff cosistent, we will just have one "pc" kernel and one "pc64" kernel configuration now.
2017-09-10audiohda: add pci id for nvidia GM204BurnZeZ
2017-09-10audiohda: add pci id for Intel 9 SeriesBurnZeZ
2017-09-10vmx(3): document changes to devvmx interfacecinap_lenrek
2017-09-10audiohda: add pci id for ICH10 (thanks echoline)cinap_lenrek
2017-09-02devvmx: call vmxshutdown from reboot() function manuallyaiju
2017-08-29sshfs: fork ssh in its own namespace so it wont keep the mountpoint opencinap_lenrek
2017-08-29sshfs: start sendproc and recvproc in the sane notegroup as the fs process ↵cinap_lenrek
so theadexitsall() works on sshfs: ending.
2017-08-29ether82563: make the ethernet of thinkpad p50 work (thanks sam-d)cinap_lenrek
2017-08-29audiohda: Intel Sunrise Point-H support (thanks sam-d)cinap_lenrek
2017-08-29sdnvme: pass 0 instead of 0xffffffff as NSID for identify controller and ↵cinap_lenrek
create completion/submission queue commands (thanks Ori_B) Samsung SSD 960 EVO fails with invalid namespace error otherwise...
2017-08-28devdup: remove useless OCEXEC check, handled by namec()cinap_lenrek
2017-08-28mergecinap_lenrek
2017-08-28devsegment: handle ORCLOSE on segment directory correctly, fix wrong qid, ↵cinap_lenrek
missing COPEN flag for segmentcreate()
2017-08-28devvmx, vmx: lilu dallas multivmaiju
2017-08-28vmx(1): don't realloc virtio queues -- breaks pointersaiju
2017-08-27sdiahci: Intel 200 Series Chipset Family PCH support (thanks aiju)cinap_lenrek
we used to tweak arround in the ICH registers for all intel controllers, which is wrong, as the 200 series has different magic registes. see the datasheet: https://www.intel.com/content/www/us/en/chipsets/200-series-chipset-pch-datasheet-vol-2.html this caused the clocks to be disabled for the 6th port causing a full machine lockup touching the 6th port registers. the next problem was that aiju's bios disabled the unused ports somehow but didnt clear ther PI bits, so that they would stay in Sbist status even after a port reset. so the port would get stuck in the Dportreset state forever. the fix for this was to use a one second timeout for the port reset procedure.
2017-08-24libmach: support disassembling from memoryaiju
2017-08-24vmx(1): VGA framebuffer should be normal memoryaiju
2017-08-24vmx(1): fixed code that assumed uintptr==uvlongaiju
2017-08-24vmx(1): memory map improvements, x86 simulator for MMIOaiju
2017-08-24devvmx: more efficient data structure for memory map; simplified (more ↵aiju
reliable) step function
generated by cgit v1.2.3 (git 2.39.1) at 2026-01-03 17:45:07 +0000