Is current release of sdfat library nice on SPI to allow other slaves?

I did a lot of search but still did not get explicit answer whether SDFAT library releases SPI bus after transaction (CLK and MISO, subject to a proper schematic) . Besides most posts on this issue are old. I am using SDFAT library downloaded last December .
Now I am posting here only regarding library functions. Proper microSD module I finally got by modifying Sparkfun's level shifting microSD board with addition of TI's TXS0104E chip. So level conversion and proper return to tri-state is not issue any more. Yet I could not find any places in the SD access functions where SDFAT library would re-establish SPI settings such as SPI.beginTransaction() or otherwise similar ...
If I add beginTransaction() and endTransaction() to my Linear LTC6804 SPI slave code, I still don;t know whether SD card accessing routines do the same "niceties"...
Meanwhile they conflict in CLK speed. Dirty fix I did - setting up both Linear and SD card to a slow 1mHz speed - did the trick, but the whole system is too slow ...
Please help...

Where do you all get all these weird cards?

If your concern is; will the SS line go HIGH after the transactions, either SD or SdFat will work.

If you need the SD to release the MISO line without that 3-state IC, use SdFat. If you want to use the SD library, the fix is easy to implement.

The problem is not electric. I am talking about SPI settings. In my current design it is mismatch of SPI speed clock, Later down the road it could be phase mode difference, etc. I expect library to re-establish SPI setting proper for a given slave and then, when access is finished, release it. Should another slave be accessed, another SPI settings recalled and then released... this is according to SPI standard and well described on arduino SPI page. SPI.beginTransaction(SPIsetting) accomplishes this well. That's what library is for... When I call something sdFile.write() - somewhere down the library calls are there SPI settings recalled for SD card? Or should I encapsulate each call to sdFile.write() with my code SPI.beginTransaction(SPIsetting) etc...
????????????????

While I browsed sdfat library code I found only chip select encapsulation (low-high) for each SD transaction. This does the trick with switching level shifters ON-OFF. Good, but the SPI settings (speed, mode) seems to be the same (changed previously for another SPI slave)... or am I missing this code in library?

I've been manipulating SPI settings manually for so long I haven't bothered with the automatic stuff. You can change them all on the fly.

edit: SdFat does handle SPI.beginTransaction. It is in SdSpiCard.cpp SdSpi.h.

  void beginTransaction(uint8_t divisor) {
#if ENABLE_SPI_TRANSACTIONS
    SPI.beginTransaction(SPISettings(4000000, MSBFIRST, SPI_MODE0));
#else  // ENABLE_SPI_TRANSACTIONS
    SPI.setBitOrder(MSBFIRST);
    SPI.setDataMode(SPI_MODE0);
#endif  // ENABLE_SPI_TRANSACTIONS
#ifndef SPI_CLOCK_DIV128
    SPI.setClockDivider(divisor);
#else  // SPI_CLOCK_DIV128
    int v;
    if (divisor <= 2) {
      v = SPI_CLOCK_DIV2;
    } else  if (divisor <= 4) {
      v = SPI_CLOCK_DIV4;
    } else  if (divisor <= 8) {
      v = SPI_CLOCK_DIV8;
    } else  if (divisor <= 16) {
      v = SPI_CLOCK_DIV16;
    } else  if (divisor <= 32) {
      v = SPI_CLOCK_DIV32;
    } else  if (divisor <= 64) {
      v = SPI_CLOCK_DIV64;
    } else {
      v = SPI_CLOCK_DIV128;
    }
    SPI.setClockDivider(v);
#endif  // SPI_CLOCK_DIV128

Every time the SdFat library sets the CS LOW, it calls that function. This is from SdSpiCard.cpp.

void SdSpiCard::chipSelectLow() {
#if WDT_YIELD_TIME_MICROS
  static uint32_t last;
  if ((micros() - last) > WDT_YIELD_TIME_MICROS) {
    SysCall::yield();
    last = micros();
  }
#endif  // WDT_YIELD_TIME_MICROS
  if (m_selected) {
    SD_CS_DBG("chipSelectLow error");
    return;
  }

// here
  spiBeginTransaction(m_sckDivisor);

  digitalWrite(m_chipSelectPin, LOW);
  m_selected = true;
}

That's exactly what I was looking for but had hard time to find this...
Thank you very much Tim
Vlad