Re: [PATCH] Remove the tcm_fc driver

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 1/11/21 1:47 AM, Bart Van Assche wrote:
Since its introduction in 2011, the tcm_fc driver processes SCSI read
commands as follows if the number of remaining bytes is a multiple of four:
- Queue data by calling fc_seq_send(). The FCoE driver translates that
   call into a dev_queue_xmit(skb) call which sends the data asynchronously.
- After having queued the data for sending, free the data buffer
   synchronously from inside ft_queue_status().

This race condition can be triggered by running fio --verify against the
FCoE initiator driver. Since this bug causes data corruption and since
nobody has reported this bug since the tcm_fc driver went upstream, this
is a strong indication that the tcm_fc driver is not being used. Hence
remove this driver from the kernel tree.

Cc: Kiran Patil <kiran.patil@xxxxxxxxx>
Cc: Hannes Reinecke <hare@xxxxxxxx>
Cc: Mike Christie <michael.christie@xxxxxxxxxx>
Cc: Christoph Hellwig <hch@xxxxxx>
Cc: Yi Zou <yi.zou@xxxxxxxxx>
Signed-off-by: Bart Van Assche <bvanassche@xxxxxxx>
---
  drivers/target/Kconfig           |   1 -
  drivers/target/Makefile          |   1 -
  drivers/target/tcm_fc/Kconfig    |   6 -
  drivers/target/tcm_fc/Makefile   |   7 -
  drivers/target/tcm_fc/tcm_fc.h   | 169 ----------
  drivers/target/tcm_fc/tfc_cmd.c  | 561 -------------------------------
  drivers/target/tcm_fc/tfc_conf.c | 491 ---------------------------
  drivers/target/tcm_fc/tfc_io.c   | 359 --------------------
  drivers/target/tcm_fc/tfc_sess.c | 503 ---------------------------
  9 files changed, 2098 deletions(-)
  delete mode 100644 drivers/target/tcm_fc/Kconfig
  delete mode 100644 drivers/target/tcm_fc/Makefile
  delete mode 100644 drivers/target/tcm_fc/tcm_fc.h
  delete mode 100644 drivers/target/tcm_fc/tfc_cmd.c
  delete mode 100644 drivers/target/tcm_fc/tfc_conf.c
  delete mode 100644 drivers/target/tcm_fc/tfc_io.c
  delete mode 100644 drivers/target/tcm_fc/tfc_sess.c

Please, no. I'd rather keep it; having the tcm_fc driver in allows one to have a simple virtualized FCoE testbed. We're using it quite heavily for QA and validation.

I'd be happy to fix the race condition instead.

Cheers,

Hannes
--
Dr. Hannes Reinecke                Kernel Storage Architect
hare@xxxxxxx                              +49 911 74053 688
SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), Geschäftsführer: Felix Imendörffer



[Index of Archives]     [Linux SCSI]     [Kernel Newbies]     [Linux SCSI Target Infrastructure]     [Share Photos]     [IDE]     [Security]     [Git]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux ATA RAID]     [Linux IIO]     [Device Mapper]

  Powered by Linux